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

Diff of /code/trunk/ChangeLog

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

revision 27 by nigel, Sat Feb 24 21:38:49 2007 UTC revision 1022 by ph10, Tue Aug 28 12:28:15 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    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
29        grep:
30    
31        (a) There is now no limit to the number of patterns to be matched.
32    
33        (b) An error is given if a pattern is too long.
34    
35        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36            are now supported.
37    
38        (d) --exclude-from and --include-from (multiple use) have been added.
39    
40        (e) Exclusions and inclusions now apply to all files and directories, not
41            just to those obtained from scanning a directory recursively.
42    
43        (f) Multiple uses of -f and --file-list are now supported.
44    
45        (g) In a Windows environment, the default for -d has been changed from
46            "read" (the GNU grep default) to "skip", because otherwise the presence
47            of a directory in the file list provokes an error.
48    
49        (h) The documentation has been revised and clarified in places.
50    
51    9.  Improve the matching speed of capturing brackets.
52    
53    10. Changed the meaning of \X so that it now matches a Unicode extended
54        grapheme cluster.
55    
56    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57        out POSIX threads when JIT support is configured.
58    
59    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
60    
61    
62    
63    Version 8.31 06-July-2012
64    -------------------------
65    
66    1.  Fixing a wrong JIT test case and some compiler warnings.
67    
68    2.  Removed a bashism from the RunTest script.
69    
70    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
71        to unsigned type, result still unsigned" that was given by an MS compiler
72        on encountering the code "-sizeof(xxx)".
73    
74    4.  Partial matching support is added to the JIT compiler.
75    
76    5.  Fixed several bugs concerned with partial matching of items that consist
77        of more than one character:
78    
79        (a) /^(..)\1/ did not partially match "aba" because checking references was
80            done on an "all or nothing" basis. This also applied to repeated
81            references.
82    
83        (b) \R did not give a hard partial match if \r was found at the end of the
84            subject.
85    
86        (c) \X did not give a hard partial match after matching one or more
87            characters at the end of the subject.
88    
89        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
90            a partial match for the string "\r".
91    
92        (e) When newline was set to CRLF, the metacharacter "." did not recognize
93            a partial match for a CR character at the end of the subject string.
94    
95    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
96        running pcretest, the text "(JIT)" added to the output whenever JIT is
97        actually used to run the match.
98    
99    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
100        or /S+[+] with a digit between 1 and 7.
101    
102    8.  OP_NOT now supports any UTF character not just single-byte ones.
103    
104    9.  (*MARK) control verb is now supported by the JIT compiler.
105    
106    10. The command "./RunTest list" lists the available tests without actually
107        running any of them. (Because I keep forgetting what they all are.)
108    
109    11. Add PCRE_INFO_MAXLOOKBEHIND.
110    
111    12. Applied a (slightly modified) user-supplied patch that improves performance
112        when the heap is used for recursion (compiled with --disable-stack-for-
113        recursion). Instead of malloc and free for each heap frame each time a
114        logical recursion happens, frames are retained on a chain and re-used where
115        possible. This sometimes gives as much as 30% improvement.
116    
117    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
118        call.
119    
120    14. As documented, (*COMMIT) is now confined to within a positive assertion.
121    
122    15. It is now possible to link pcretest with libedit as an alternative to
123        libreadline.
124    
125    16. (*COMMIT) control verb is now supported by the JIT compiler.
126    
127    17. The Unicode data tables have been updated to Unicode 6.1.0.
128    
129    18. Added --file-list option to pcregrep.
130    
131    19. Added binary file support to pcregrep, including the -a, --binary-files,
132        -I, and --text options.
133    
134    20. The madvise function is renamed for posix_madvise for QNX compatibility
135        reasons. Fixed by Giuseppe D'Angelo.
136    
137    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
138    
139    22. Changed the option for creating symbolic links for 16-bit man pages from
140        -s to -sf so that re-installing does not cause issues.
141    
142    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
143    
144    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
145        matches in certain environments (the workspace was not being correctly
146        retained). Also added to pcre_dfa_exec() a simple plausibility check on
147        some of the workspace data at the beginning of a restart.
148    
149    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
150        was not doing so when it should - probably a typo introduced by SVN 528
151        (change 8.10/14).
152    
153    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
154        \w+ when the character tables indicated that \x{c4} was a word character.
155        There were several related cases, all because the tests for doing a table
156        lookup were testing for characters less than 127 instead of 255.
157    
158    27. If a pattern contains capturing parentheses that are not used in a match,
159        their slots in the ovector are set to -1. For those that are higher than
160        any matched groups, this happens at the end of processing. In the case when
161        there were back references that the ovector was too small to contain
162        (causing temporary malloc'd memory to be used during matching), and the
163        highest capturing number was not used, memory off the end of the ovector
164        was incorrectly being set to -1. (It was using the size of the temporary
165        memory instead of the true size.)
166    
167    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
168        ovector size, it uses memory at the end of the block that it has got.
169    
170    29. Check for an overlong MARK name and give an error at compile time. The
171        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
172    
173    30. JIT compiler update.
174    
175    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
176        Rill for the patch.
177    
178    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
179    
180    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
181    
182    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
183        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
184        was enabled.
185    
186    35. Improve JIT code generation for greedy plus quantifier.
187    
188    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
189        1 to "aa" instead of to an empty string. The bug affected repeated groups
190        that could potentially match an empty string.
191    
192    37. Optimizing single character iterators in JIT.
193    
194    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
195        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
196        codepoints that are too big for the mode are faulted, and in a UTF mode,
197        disallowed codepoints are also faulted.
198    
199    39. If PCRE was compiled with UTF support, in three places in the DFA
200        matcher there was code that should only have been obeyed in UTF mode, but
201        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
202        processing when bytes with values greater than 127 were present. In 16-bit
203        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
204        both cases the values are those that cannot be the first data item in a UTF
205        character. The three items that might have provoked this were recursions,
206        possessively repeated groups, and atomic groups.
207    
208    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
209        and pcregrep, because some OS require shared objects to be explicitly
210        passed to ld, causing the link step to fail if they are not.
211    
212    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
213        mode, patterns that started with \h* or \R* might be incorrectly matched.
214    
215    
216    Version 8.30 04-February-2012
217    -----------------------------
218    
219    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
220        name is defined in ctype.h.
221    
222    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
223        only in quite long subpatterns.
224    
225    3.  Removed the function pcre_info(), which has been obsolete and deprecated
226        since it was replaced by pcre_fullinfo() in February 2000.
227    
228    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
229        match a (*MARK), and the match failed at the start of the subject, a
230        reference to memory before the start of the subject could occur. This bug
231        was introduced by fix 17 of release 8.21.
232    
233    5.  A reference to an unset group with zero minimum repetition was giving
234        totally wrong answers (in non-JavaScript-compatibility mode). For example,
235        /(another)?(\1?)test/ matched against "hello world test". This bug was
236        introduced in release 8.13.
237    
238    6.  Add support for 16-bit character strings (a large amount of work involving
239        many changes and refactorings).
240    
241    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
242        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
243        from a file.
244    
245    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
246        rounding is not applied in this particular case).
247    
248    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
249        if they appear, or are escaped, in patterns.
250    
251    10. Get rid of a number of -Wunused-but-set-variable warnings.
252    
253    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
254        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
255        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
256        also returns the mark "x". This bug applied to capturing parentheses,
257        non-capturing parentheses, and atomic parentheses. It also applied to some
258        assertions.
259    
260    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
261        information out of configure.ac instead of relying on pcre.h.generic, which
262        is not stored in the repository.
263    
264    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
265        -lreadline.
266    
267    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
268    
269    15. Applied Graycode's patch to put the top-level frame on the stack rather
270        than the heap when not using the stack for recursion. This gives a
271        performance improvement in many cases when recursion is not deep.
272    
273    16. Experimental code added to "pcretest -C" to output the stack frame size.
274    
275    
276    Version 8.21 12-Dec-2011
277    ------------------------
278    
279    1.  Updating the JIT compiler.
280    
281    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
282        are added as well.
283    
284    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
285        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
286        calling _pcre_jit_exec. Some extra comments are added.
287    
288    4.  (*MARK) settings inside atomic groups that do not contain any capturing
289        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
290        was introduced by change 18 for 8.20.
291    
292    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
293        ECMA-262 standard.
294    
295    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
296        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
297        This bug was probably introduced by change 9 of 8.13.
298    
299    7.  While fixing 6 above, I noticed that a number of other items were being
300        incorrectly rejected as "not fixed length". This arose partly because newer
301        opcodes had not been added to the fixed-length checking code. I have (a)
302        corrected the bug and added tests for these items, and (b) arranged for an
303        error to occur if an unknown opcode is encountered while checking for fixed
304        length instead of just assuming "not fixed length". The items that were
305        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
306        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
307        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
308    
309    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
310        being incorrectly compiled and would have given unpredicatble results.
311    
312    9.  A possessively repeated subpattern with minimum repeat count greater than
313        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
314        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
315        the first (A) could occur when it should not.
316    
317    10. Add a cast and remove a redundant test from the code.
318    
319    11. JIT should use pcre_malloc/pcre_free for allocation.
320    
321    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
322        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
323        is anything other than /usr, -L is still shown).
324    
325    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
326    
327    14. Perl does not support \N without a following name in a [] class; PCRE now
328        also gives an error.
329    
330    15. If a forward reference was repeated with an upper limit of around 2000,
331        it caused the error "internal error: overran compiling workspace". The
332        maximum number of forward references (including repeats) was limited by the
333        internal workspace, and dependent on the LINK_SIZE. The code has been
334        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
335        the default depends on LINK_SIZE. There is a new upper limit (for safety)
336        of around 200,000 forward references. While doing this, I also speeded up
337        the filling in of repeated forward references.
338    
339    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
340        incorrectly expecting the subject to contain another "a" after the start.
341    
342    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
343        in the match, the SKIP should be ignored. This was not happening; instead
344        the SKIP was being treated as NOMATCH. For patterns such as
345        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
346        tested.
347    
348    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
349        now much more compatible with Perl, in particular in cases where the result
350        is a non-match for a non-anchored pattern. For example, if
351        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
352        "m", where previously it did not return a name. A side effect of this
353        change is that for partial matches, the last encountered mark name is
354        returned, as for non matches. A number of tests that were previously not
355        Perl-compatible have been moved into the Perl-compatible test files. The
356        refactoring has had the pleasing side effect of removing one argument from
357        the match() function, thus reducing its stack requirements.
358    
359    19. If the /S+ option was used in pcretest to study a pattern using JIT,
360        subsequent uses of /S (without +) incorrectly behaved like /S+.
361    
362    21. Retrieve executable code size support for the JIT compiler and fixing
363        some warnings.
364    
365    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
366        not work when the shorter character appeared right at the end of the
367        subject string.
368    
369    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
370        systems.
371    
372    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
373        output it when the /M option is used in pcretest.
374    
375    25. The CheckMan script was not being included in the distribution. Also, added
376        an explicit "perl" to run Perl scripts from the PrepareRelease script
377        because this is reportedly needed in Windows.
378    
379    26. If study data was being save in a file and studying had not found a set of
380        "starts with" bytes for the pattern, the data written to the file (though
381        never used) was taken from uninitialized memory and so caused valgrind to
382        complain.
383    
384    27. Updated RunTest.bat as provided by Sheri Pierce.
385    
386    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
387    
388    29. Computation of memory usage for the table of capturing group names was
389        giving an unnecessarily large value.
390    
391    
392    Version 8.20 21-Oct-2011
393    ------------------------
394    
395    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
396        a POSIX class. After further experiments with Perl, which convinced me that
397        Perl has bugs and confusions, a closing square bracket is no longer allowed
398        in a POSIX name. This bug also affected patterns with classes that started
399        with full stops.
400    
401    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
402        captured substring, but while checking the failing first alternative,
403        substring 1 is temporarily captured. If the output vector supplied to
404        pcre_exec() was not big enough for this capture, the yield of the function
405        was still zero ("insufficient space for captured substrings"). This cannot
406        be totally fixed without adding another stack variable, which seems a lot
407        of expense for a edge case. However, I have improved the situation in cases
408        such as /(a)(b)x|abc/ matched against "abc", where the return code
409        indicates that fewer than the maximum number of slots in the ovector have
410        been set.
411    
412    3.  Related to (2) above: when there are more back references in a pattern than
413        slots in the output vector, pcre_exec() uses temporary memory during
414        matching, and copies in the captures as far as possible afterwards. It was
415        using the entire output vector, but this conflicts with the specification
416        that only 2/3 is used for passing back captured substrings. Now it uses
417        only the first 2/3, for compatibility. This is, of course, another edge
418        case.
419    
420    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
421        main code base, and can be used by building with --enable-jit. When this is
422        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
423        runtime --no-jit option is given.
424    
425    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
426        ovector, the return from the function was zero, implying that there were
427        other matches that did not fit. The correct "exactly full" value is now
428        returned.
429    
430    6.  If a subpattern that was called recursively or as a subroutine contained
431        (*PRUNE) or any other control that caused it to give a non-standard return,
432        invalid errors such as "Error -26 (nested recursion at the same subject
433        position)" or even infinite loops could occur.
434    
435    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
436        computing the minimum length on reaching *ACCEPT, and so ended up with the
437        wrong value of 1 rather than 0. Further investigation indicates that
438        computing a minimum subject length in the presence of *ACCEPT is difficult
439        (think back references, subroutine calls), and so I have changed the code
440        so that no minimum is registered for a pattern that contains *ACCEPT.
441    
442    8.  If (*THEN) was present in the first (true) branch of a conditional group,
443        it was not handled as intended. [But see 16 below.]
444    
445    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
446        Sheri Pierce.
447    
448    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
449        the first byte in a match must be "a".
450    
451    11. Change 17 for 8.13 increased the recursion depth for patterns like
452        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
453        pattern contains any instances of (*THEN). If it does not, the old
454        optimizations are restored. It would be nice to do this on a per-group
455        basis, but at the moment that is not feasible.
456    
457    12. In some environments, the output of pcretest -C is CRLF terminated. This
458        broke RunTest's code that checks for the link size. A single white space
459        character after the value is now allowed for.
460    
461    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
462        For "fr", it uses the Windows-specific input and output files.
463    
464    14. If (*THEN) appeared in a group that was called recursively or as a
465        subroutine, it did not work as intended. [But see next item.]
466    
467    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
468        pattern fragments (but not containing any | characters). If A and B are
469        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
470        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
471        D. In other words, Perl considers parentheses that do not contain any |
472        characters to be part of a surrounding alternative, whereas PCRE was
473        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
474        differently. PCRE now behaves in the same way as Perl, except in the case
475        of subroutine/recursion calls such as (?1) which have in any case always
476        been different (but PCRE had them first :-).
477    
478    16. Related to 15 above: Perl does not treat the | in a conditional group as
479        creating alternatives. Such a group is treated in the same way as an
480        ordinary group without any | characters when processing (*THEN). PCRE has
481        been changed to match Perl's behaviour.
482    
483    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
484        RunGrepTest script failed.
485    
486    18. Change 22 for version 13 caused atomic groups to use more stack. This is
487        inevitable for groups that contain captures, but it can lead to a lot of
488        stack use in large patterns. The old behaviour has been restored for atomic
489        groups that do not contain any capturing parentheses.
490    
491    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
492        suppress the check for a minimum subject length at run time. (If it was
493        given to pcre_exec() or pcre_dfa_exec() it did work.)
494    
495    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
496        fail to work when decoding hex characters in data strings in EBCDIC
497        environments.
498    
499    21. It appears that in at least one Mac OS environment, the isxdigit() function
500        is implemented as a macro that evaluates to its argument more than once,
501        contravening the C 90 Standard (I haven't checked a later standard). There
502        was an instance in pcretest which caused it to go wrong when processing
503        \x{...} escapes in subject strings. The has been rewritten to avoid using
504        things like p++ in the argument of isxdigit().
505    
506    
507    Version 8.13 16-Aug-2011
508    ------------------------
509    
510    1.  The Unicode data tables have been updated to Unicode 6.0.0.
511    
512    2.  Two minor typos in pcre_internal.h have been fixed.
513    
514    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
515        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
516        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
517    
518    4.  There were a number of related bugs in the code for matching backrefences
519        caselessly in UTF-8 mode when codes for the characters concerned were
520        different numbers of bytes. For example, U+023A and U+2C65 are an upper
521        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
522        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
523        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
524        2-byte code at the end of the subject (it thought there wasn't enough data
525        left).
526    
527    5.  Comprehensive information about what went wrong is now returned by
528        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
529        as the output vector has at least 2 elements. The offset of the start of
530        the failing character and a reason code are placed in the vector.
531    
532    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
533        now returned is for the first byte of the failing character, instead of the
534        last byte inspected. This is an incompatible change, but I hope it is small
535        enough not to be a problem. It makes the returned offset consistent with
536        pcre_exec() and pcre_dfa_exec().
537    
538    7.  pcretest now gives a text phrase as well as the error number when
539        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
540        failure, the offset and reason code are output.
541    
542    8.  When \R was used with a maximizing quantifier it failed to skip backwards
543        over a \r\n pair if the subsequent match failed. Instead, it just skipped
544        back over a single character (\n). This seems wrong (because it treated the
545        two characters as a single entity when going forwards), conflicts with the
546        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
547        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
548        has been changed.
549    
550    9.  Some internal refactoring has changed the processing so that the handling
551        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
552        time (the PCRE_DOTALL option was changed this way some time ago: version
553        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
554        which was always a bit of a fudge. It also means that there is one less
555        argument for the match() function, which reduces its stack requirements
556        slightly. This change also fixes an incompatibility with Perl: the pattern
557        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
558    
559    10. More internal refactoring has drastically reduced the number of recursive
560        calls to match() for possessively repeated groups such as (abc)++ when
561        using pcre_exec().
562    
563    11. While implementing 10, a number of bugs in the handling of groups were
564        discovered and fixed:
565    
566        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
567        (a|)*(?1) gave a compile-time internal error.
568        ((a|)+)+  did not notice that the outer group could match an empty string.
569        (^a|^)+   was not marked as anchored.
570        (.*a|.*)+ was not marked as matching at start or after a newline.
571    
572    12. Yet more internal refactoring has removed another argument from the match()
573        function. Special calls to this function are now indicated by setting a
574        value in a variable in the "match data" data block.
575    
576    13. Be more explicit in pcre_study() instead of relying on "default" for
577        opcodes that mean there is no starting character; this means that when new
578        ones are added and accidentally left out of pcre_study(), testing should
579        pick them up.
580    
581    14. The -s option of pcretest has been documented for ages as being an old
582        synonym of -m (show memory usage). I have changed it to mean "force study
583        for every regex", that is, assume /S for every regex. This is similar to -i
584        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
585        using it. It makes it easier to run collections of tests with and without
586        study enabled, and thereby test pcre_study() more easily. All the standard
587        tests are now run with and without -s (but some patterns can be marked as
588        "never study" - see 20 below).
589    
590    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
591        restoration of the capturing data to the outer values was not happening
592        correctly.
593    
594    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
595        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
596        pattern had matched an empty string, and so incorrectly returned a no
597        match.
598    
599    17. There was optimizing code for the last branch of non-capturing parentheses,
600        and also for the obeyed branch of a conditional subexpression, which used
601        tail recursion to cut down on stack usage. Unfortunately, now that there is
602        the possibility of (*THEN) occurring in these branches, tail recursion is
603        no longer possible because the return has to be checked for (*THEN). These
604        two optimizations have therefore been removed. [But see 8.20/11 above.]
605    
606    18. If a pattern containing \R was studied, it was assumed that \R always
607        matched two bytes, thus causing the minimum subject length to be
608        incorrectly computed because \R can also match just one byte.
609    
610    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
611        was incorrectly computed.
612    
613    20. If /S is present twice on a test pattern in pcretest input, it now
614        *disables* studying, thereby overriding the use of -s on the command line
615        (see 14 above). This is necessary for one or two tests to keep the output
616        identical in both cases.
617    
618    21. When (*ACCEPT) was used in an assertion that matched an empty string and
619        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
620    
621    22. When an atomic group that contained a capturing parenthesis was
622        successfully matched, but the branch in which it appeared failed, the
623        capturing was not being forgotten if a higher numbered group was later
624        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
625        group 1 to "a", when in fact it should be unset. This applied to multi-
626        branched capturing and non-capturing groups, repeated or not, and also to
627        positive assertions (capturing in negative assertions does not happen
628        in PCRE) and also to nested atomic groups.
629    
630    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
631        subject after a captured substring, to make it easier to tell which of a
632        number of identical substrings has been captured.
633    
634    24. The way atomic groups are processed by pcre_exec() has been changed so that
635        if they are repeated, backtracking one repetition now resets captured
636        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
637        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
638        "aaa". Previously, it would have been "a". As part of this code
639        refactoring, the way recursive calls are handled has also been changed.
640    
641    25. If an assertion condition captured any substrings, they were not passed
642        back unless some other capturing happened later. For example, if
643        (?(?=(a))a) was matched against "a", no capturing was returned.
644    
645    26. When studying a pattern that contained subroutine calls or assertions,
646        the code for finding the minimum length of a possible match was handling
647        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
648        group 1 called group 2 while simultaneously a separate group 2 called group
649        1). A stack overflow occurred in this case. I have fixed this by limiting
650        the recursion depth to 10.
651    
652    27. Updated RunTest.bat in the distribution to the version supplied by Tom
653        Fortmann. This supports explicit test numbers on the command line, and has
654        argument validation and error reporting.
655    
656    28. An instance of \X with an unlimited repeat could fail if at any point the
657        first character it looked at was a mark character.
658    
659    29. Some minor code refactoring concerning Unicode properties and scripts
660        should reduce the stack requirement of match() slightly.
661    
662    30. Added the '=' option to pcretest to check the setting of unused capturing
663        slots at the end of the pattern, which are documented as being -1, but are
664        not included in the return count.
665    
666    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
667        compiled something random. Now it gives a compile-time error (as does
668        Perl).
669    
670    32. A *MARK encountered during the processing of a positive assertion is now
671        recorded and passed back (compatible with Perl).
672    
673    33. If --only-matching or --colour was set on a pcregrep call whose pattern
674        had alternative anchored branches, the search for a second match in a line
675        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
676        matched the line "0102" twice. The same bug affected patterns that started
677        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
678        twice.
679    
680    34. Previously, PCRE did not allow quantification of assertions. However, Perl
681        does, and because of capturing effects, quantifying parenthesized
682        assertions may at times be useful. Quantifiers are now allowed for
683        parenthesized assertions.
684    
685    35. A minor code tidy in pcre_compile() when checking options for \R usage.
686    
687    36. \g was being checked for fancy things in a character class, when it should
688        just be a literal "g".
689    
690    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
691        appearance of a nested POSIX class supersedes an apparent external class.
692        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
693        unescaped square brackets may also appear as part of class names. For
694        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
695        more like Perl. (But see 8.20/1 above.)
696    
697    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
698        was because it thought it was \N{name}, which is not supported).
699    
700    39. Add minix to OS list not supporting the -S option in pcretest.
701    
702    40. PCRE tries to detect cases of infinite recursion at compile time, but it
703        cannot analyze patterns in sufficient detail to catch mutual recursions
704        such as ((?1))((?2)). There is now a runtime test that gives an error if a
705        subgroup is called recursively as a subpattern for a second time at the
706        same position in the subject string. In previous releases this might have
707        been caught by the recursion limit, or it might have run out of stack.
708    
709    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
710        happen only once. PCRE was, however incorrectly giving a compile time error
711        "recursive call could loop indefinitely" because it cannot analyze the
712        pattern in sufficient detail. The compile time test no longer happens when
713        PCRE is compiling a conditional subpattern, but actual runaway loops are
714        now caught at runtime (see 40 above).
715    
716    42. It seems that Perl allows any characters other than a closing parenthesis
717        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
718        has been changed to be the same.
719    
720    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
721        as not to get warnings when autogen.sh is called. Also changed
722        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
723    
724    44. To help people who use pcregrep to scan files containing exceedingly long
725        lines, the following changes have been made:
726    
727        (a) The default value of the buffer size parameter has been increased from
728            8K to 20K. (The actual buffer used is three times this size.)
729    
730        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
731            PCRE is built.
732    
733        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
734            to be set at run time.
735    
736        (d) Numerical values in pcregrep options can be followed by K or M, for
737            example --buffer-size=50K.
738    
739        (e) If a line being scanned overflows pcregrep's buffer, an error is now
740            given and the return code is set to 2.
741    
742    45. Add a pointer to the latest mark to the callout data block.
743    
744    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
745        partial match of an empty string instead of no match. This was specific to
746        the use of ".".
747    
748    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
749        complete match instead of a partial match. This bug was dependent on both
750        the PCRE_UTF8 and PCRE_DOTALL options being set.
751    
752    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
753        starting byte set, because \b was not being ignored.
754    
755    
756    Version 8.12 15-Jan-2011
757    ------------------------
758    
759    1.  Fixed some typos in the markup of the man pages, and wrote a script that
760        checks for such things as part of the documentation building process.
761    
762    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
763        --match-limit and --recursion-limit options (added for 8.11). In
764        particular, this made one of the standard tests fail. (The integer value
765        went into the wrong half of a long int.)
766    
767    3.  If the --colour option was given to pcregrep with -v (invert match), it
768        did strange things, either producing crazy output, or crashing. It should,
769        of course, ignore a request for colour when reporting lines that do not
770        match.
771    
772    4.  Another pcregrep bug caused similar problems if --colour was specified with
773        -M (multiline) and the pattern match finished with a line ending.
774    
775    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
776        matched in multiline mode, the following line was shown as part of the
777        match. This seems wrong, so I have changed it.
778    
779    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
780        the check for further matches in the same line (so they could be coloured)
781        to overrun the end of the current line. If another match was found, it was
782        incorrectly shown (and then shown again when found in the next line).
783    
784    7.  If pcregrep was compiled under Windows, there was a reference to the
785        function pcregrep_exit() before it was defined. I am assuming this was
786        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
787        reported by a user. I've moved the definition above the reference.
788    
789    
790    Version 8.11 10-Dec-2010
791    ------------------------
792    
793    1.  (*THEN) was not working properly if there were untried alternatives prior
794        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
795        backtracked to try for "b" instead of moving to the next alternative branch
796        at the same level (in this case, to look for "c"). The Perl documentation
797        is clear that when (*THEN) is backtracked onto, it goes to the "next
798        alternative in the innermost enclosing group".
799    
800    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
801        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
802        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
803        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
804        (*THEN).
805    
806    3.  If \s appeared in a character class, it removed the VT character from
807        the class, even if it had been included by some previous item, for example
808        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
809        of \s, but is part of the POSIX "space" class.)
810    
811    4.  A partial match never returns an empty string (because you can always
812        match an empty string at the end of the subject); however the checking for
813        an empty string was starting at the "start of match" point. This has been
814        changed to the "earliest inspected character" point, because the returned
815        data for a partial match starts at this character. This means that, for
816        example, /(?<=abc)def/ gives a partial match for the subject "abc"
817        (previously it gave "no match").
818    
819    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
820        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
821        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
822        has an implication that the given string is incomplete (because a partial
823        match is preferred over a full match). For this reason, these items now
824        give a partial match in this situation. [Aside: previously, the one case
825        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
826        match rather than a full match, which was wrong by the old rules, but is
827        now correct.]
828    
829    6.  There was a bug in the handling of #-introduced comments, recognized when
830        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
831        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
832        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
833        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
834        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
835        places in pcre_compile().
836    
837    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
838        comments when looking ahead for named forward references to subpatterns,
839        the only newline sequence it recognized was NL. It now handles newlines
840        according to the set newline convention.
841    
842    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
843        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
844        cater for a lack of strerror(). These oversights have been fixed.
845    
846    9.  Added --match-limit and --recursion-limit to pcregrep.
847    
848    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
849    
850    11. When the -o option was used, pcregrep was setting a return code of 1, even
851        when matches were found, and --line-buffered was not being honoured.
852    
853    12. Added an optional parentheses number to the -o and --only-matching options
854        of pcregrep.
855    
856    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
857        can match an empty string. The code to do it in pcretest and pcredemo
858        needed fixing:
859    
860        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
861            only one byte after an empty string match just before CRLF (this case
862            just got forgotten; "any" and "anycrlf" were OK).
863    
864        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
865            mode when an empty string match preceded an ASCII character followed by
866            a non-ASCII character. (The code for advancing by one character rather
867            than one byte was nonsense.)
868    
869        (c) The pcredemo.c sample program did not have any code at all to handle
870            the cases when CRLF is a valid newline sequence.
871    
872    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
873        as a starting offset was within the subject string. There is now a new
874        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
875        negative or greater than the length of the string. In order to test this,
876        pcretest is extended to allow the setting of negative starting offsets.
877    
878    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
879        starting offset points to the beginning of a UTF-8 character was
880        unnecessarily clumsy. I tidied it up.
881    
882    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
883        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
884    
885    17. Nobody had reported that the --include_dir option, which was added in
886        release 7.7 should have been called --include-dir (hyphen, not underscore)
887        for compatibility with GNU grep. I have changed it to --include-dir, but
888        left --include_dir as an undocumented synonym, and the same for
889        --exclude-dir, though that is not available in GNU grep, at least as of
890        release 2.5.4.
891    
892    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
893        characters from a string of bytes) have been redefined so as not to use
894        loops, in order to improve performance in some environments. At the same
895        time, I abstracted some of the common code into auxiliary macros to save
896        repetition (this should not affect the compiled code).
897    
898    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
899        compile-time error is now given if \c is not followed by an ASCII
900        character, that is, a byte less than 128. (In EBCDIC mode, the code is
901        different, and any byte value is allowed.)
902    
903    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
904        START_OPTIMIZE option, which is now allowed at compile time - but just
905        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
906        to pcregrep and other applications that have no direct access to PCRE
907        options. The new /Y option in pcretest sets this option when calling
908        pcre_compile().
909    
910    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
911        back references. Groups containing recursive back references were forced to
912        be atomic by that change, but in the case of named groups, the amount of
913        memory required was incorrectly computed, leading to "Failed: internal
914        error: code overflow". This has been fixed.
915    
916    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
917        pcretest.c, to avoid build problems in some Borland environments.
918    
919    
920    Version 8.10 25-Jun-2010
921    ------------------------
922    
923    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
924        THEN.
925    
926    2.  (*ACCEPT) was not working when inside an atomic group.
927    
928    3.  Inside a character class, \B is treated as a literal by default, but
929        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
930        causes the error). The code is unchanged, but I tidied the documentation.
931    
932    4.  Inside a character class, PCRE always treated \R and \X as literals,
933        whereas Perl faults them if its -w option is set. I have changed PCRE so
934        that it faults them when PCRE_EXTRA is set.
935    
936    5.  Added support for \N, which always matches any character other than
937        newline. (It is the same as "." when PCRE_DOTALL is not set.)
938    
939    6.  When compiling pcregrep with newer versions of gcc which may have
940        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
941        declared with attribute warn_unused_result" were given. Just casting the
942        result to (void) does not stop the warnings; a more elaborate fudge is
943        needed. I've used a macro to implement this.
944    
945    7.  Minor change to pcretest.c to avoid a compiler warning.
946    
947    8.  Added four artifical Unicode properties to help with an option to make
948        \s etc use properties (see next item). The new properties are: Xan
949        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
950    
951    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
952        use Unicode properties. (*UCP) at the start of a pattern can be used to set
953        this option. Modified pcretest to add /W to test this facility. Added
954        REG_UCP to make it available via the POSIX interface.
955    
956    10. Added --line-buffered to pcregrep.
957    
958    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
959        studied, and the match started with a letter with a code point greater than
960        127 whose first byte was different to the first byte of the other case of
961        the letter, the other case of this starting letter was not recognized
962        (#976).
963    
964    12. If a pattern that was studied started with a repeated Unicode property
965        test, for example, \p{Nd}+, there was the theoretical possibility of
966        setting up an incorrect bitmap of starting bytes, but fortunately it could
967        not have actually happened in practice until change 8 above was made (it
968        added property types that matched character-matching opcodes).
969    
970    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
971        possible starting bytes for non-anchored patterns.
972    
973    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
974        \R, and also a number of cases that involve Unicode properties, both
975        explicit and implicit when PCRE_UCP is set.
976    
977    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
978        input, it could crash or give wrong results if characters with values
979        greater than 0xc0 were present in the subject string. (Detail: it assumed
980        UTF-8 input when processing these items.)
981    
982    16. Added a lot of (int) casts to avoid compiler warnings in systems where
983        size_t is 64-bit (#991).
984    
985    17. Added a check for running out of memory when PCRE is compiled with
986        --disable-stack-for-recursion (#990).
987    
988    18. If the last data line in a file for pcretest does not have a newline on
989        the end, a newline was missing in the output.
990    
991    19. The default pcre_chartables.c file recognizes only ASCII characters (values
992        less than 128) in its various bitmaps. However, there is a facility for
993        generating tables according to the current locale when PCRE is compiled. It
994        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
995        characters, are recognized by isspace() and therefore were getting set in
996        these tables, and indeed these tables seem to approximate to ISO 8859. This
997        caused a problem in UTF-8 mode when pcre_study() was used to create a list
998        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
999        which of course cannot start UTF-8 characters. I have changed the code so
1000        that only real ASCII characters (less than 128) and the correct starting
1001        bytes for UTF-8 encodings are set for characters greater than 127 when in
1002        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1003        altogether.)
1004    
1005    20. Added the /T option to pcretest so as to be able to run tests with non-
1006        standard character tables, thus making it possible to include the tests
1007        used for 19 above in the standard set of tests.
1008    
1009    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1010        reference to a subpattern the other side of a comment that contains an
1011        opening parenthesis caused either an internal compiling error, or a
1012        reference to the wrong subpattern.
1013    
1014    
1015    Version 8.02 19-Mar-2010
1016    ------------------------
1017    
1018    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1019    
1020    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1021        configured.
1022    
1023    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1024        original author of that file, following a query about its status.
1025    
1026    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1027        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1028    
1029    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1030        quantifier applied to a forward-referencing subroutine call, could compile
1031        incorrect code or give the error "internal error: previously-checked
1032        referenced subpattern not found".
1033    
1034    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1035        variables to point to external functions. For these systems, therefore,
1036        pcre_malloc etc. are now initialized to local functions that call the
1037        relevant global functions.
1038    
1039    7.  There were two entries missing in the vectors called coptable and poptable
1040        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1041        I've fixed the data, and added a kludgy way of testing at compile time that
1042        the lengths are correct (equal to the number of opcodes).
1043    
1044    8.  Following on from 7, I added a similar kludge to check the length of the
1045        eint vector in pcreposix.c.
1046    
1047    9.  Error texts for pcre_compile() are held as one long string to avoid too
1048        much relocation at load time. To find a text, the string is searched,
1049        counting zeros. There was no check for running off the end of the string,
1050        which could happen if a new error number was added without updating the
1051        string.
1052    
1053    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1054    
1055    11. \K was not working if it appeared in an atomic group or in a group that
1056        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1057        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1058        the assertion is positive, but not if it is negative.
1059    
1060    12. Change 11 fortuitously reduced the size of the stack frame used in the
1061        "match()" function of pcre_exec.c by one pointer. Forthcoming
1062        implementation of support for (*MARK) will need an extra pointer on the
1063        stack; I have reserved it now, so that the stack frame size does not
1064        decrease.
1065    
1066    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1067        item in branch that calls a recursion is a subroutine call - as in the
1068        second branch in the above example - was incorrectly given the compile-
1069        time error "recursive call could loop indefinitely" because pcre_compile()
1070        was not correctly checking the subroutine for matching a non-empty string.
1071    
1072    14. The checks for overrunning compiling workspace could trigger after an
1073        overrun had occurred. This is a "should never occur" error, but it can be
1074        triggered by pathological patterns such as hundreds of nested parentheses.
1075        The checks now trigger 100 bytes before the end of the workspace.
1076    
1077    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1078    
1079    
1080    Version 8.01 19-Jan-2010
1081    ------------------------
1082    
1083    1.  If a pattern contained a conditional subpattern with only one branch (in
1084        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1085        computed the wrong minimum data length (which is of course zero for such
1086        subpatterns). This could cause incorrect "no match" results.
1087    
1088    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1089        the pattern is reset in the first branch, pcre_compile() failed with
1090        "internal error: code overflow at offset...". This happened only when
1091        the reset was to the original external option setting. (An optimization
1092        abstracts leading options settings into an external setting, which was the
1093        cause of this.)
1094    
1095    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1096        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1097        assertion pattern did not match (meaning that the assertion was true), it
1098        was incorrectly treated as false if the SKIP had been reached during the
1099        matching. This also applied to assertions used as conditions.
1100    
1101    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1102        assertion subpattern, including such a pattern used as a condition,
1103        unpredictable results occurred, instead of the error return
1104        PCRE_ERROR_DFA_UITEM.
1105    
1106    5.  The C++ GlobalReplace function was not working like Perl for the special
1107        situation when an empty string is matched. It now does the fancy magic
1108        stuff that is necessary.
1109    
1110    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1111        removed. (These were left over from very, very early versions of PCRE.)
1112    
1113    7.  Some cosmetic changes to the code to make life easier when compiling it
1114        as part of something else:
1115    
1116        (a) Change DEBUG to PCRE_DEBUG.
1117    
1118        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1119            called "current" as "current_branch", to prevent a collision with the
1120            Linux macro when compiled as a kernel module.
1121    
1122        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1123            prevent a collision with the Linux macro when compiled as a kernel
1124            module.
1125    
1126    8.  In pcre_compile() there are some checks for integer overflows that used to
1127        cast potentially large values to (double). This has been changed to that
1128        when building, a check for int64_t is made, and if it is found, it is used
1129        instead, thus avoiding the use of floating point arithmetic. (There is no
1130        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1131        double.
1132    
1133    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1134        2005 (difference between two addresses compared to an unsigned value).
1135    
1136    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1137        custom one, because of the following reported problem in Windows:
1138    
1139          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1140              under Win32.
1141          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1142              therefore missing the function definition.
1143          - The compiler thus generates a "C" signature for the test function.
1144          - The linker fails to find the "C" function.
1145          - PCRE fails to configure if asked to do so against libbz2.
1146    
1147    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1148        messages were output:
1149    
1150          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1151          rerunning libtoolize, to keep the correct libtool macros in-tree.
1152          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1153    
1154        I have done both of these things.
1155    
1156    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1157        most of the time, it *can* run out if it is given a pattern that contains a
1158        runaway infinite recursion. I updated the discussion in the pcrestack man
1159        page.
1160    
1161    13. Now that we have gone to the x.xx style of version numbers, the minor
1162        version may start with zero. Using 08 or 09 is a bad idea because users
1163        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1164        interpreted as invalid octal numbers. I've updated the previous comment in
1165        configure.ac, and also added a check that gives an error if 08 or 09 are
1166        used.
1167    
1168    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1169        causing partial matching to fail when the end of the subject matched \W
1170        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1171    
1172    15. There were some discrepancies between the declarations in pcre_internal.h
1173        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1174        their definitions. The declarations used "const uschar *" and the
1175        definitions used USPTR. Even though USPTR is normally defined as "const
1176        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1177        reported that: "This difference in casting confuses some C++ compilers, for
1178        example, SunCC recognizes above declarations as different functions and
1179        generates broken code for hbpcre." I have changed the declarations to use
1180        USPTR.
1181    
1182    16. GNU libtool is named differently on some systems. The autogen.sh script now
1183        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1184        (FreeBSD).
1185    
1186    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1187        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1188        comment: "Figure out how to create a longlong from a string: strtoll and
1189        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1190        instance, but it only takes 2 args instead of 3!"
1191    
1192    18. A subtle bug concerned with back references has been fixed by a change of
1193        specification, with a corresponding code fix. A pattern such as
1194        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1195        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1196        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1197        same bug. Such groups have to be quantified to be useful, or contained
1198        inside another quantified group. (If there's no repetition, the reference
1199        can never match.) The problem arises because, having left the group and
1200        moved on to the rest of the pattern, a later failure that backtracks into
1201        the group uses the captured value from the final iteration of the group
1202        rather than the correct earlier one. I have fixed this in PCRE by forcing
1203        any group that contains a reference to itself to be an atomic group; that
1204        is, there cannot be any backtracking into it once it has completed. This is
1205        similar to recursive and subroutine calls.
1206    
1207    
1208    Version 8.00 19-Oct-09
1209    ----------------------
1210    
1211    1.  The table for translating pcre_compile() error codes into POSIX error codes
1212        was out-of-date, and there was no check on the pcre_compile() error code
1213        being within the table. This could lead to an OK return being given in
1214        error.
1215    
1216    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1217        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1218        in a Windows environment.
1219    
1220    3.  The pcregrep --count option prints the count for each file even when it is
1221        zero, as does GNU grep. However, pcregrep was also printing all files when
1222        --files-with-matches was added. Now, when both options are given, it prints
1223        counts only for those files that have at least one match. (GNU grep just
1224        prints the file name in this circumstance, but including the count seems
1225        more useful - otherwise, why use --count?) Also ensured that the
1226        combination -clh just lists non-zero counts, with no names.
1227    
1228    4.  The long form of the pcregrep -F option was incorrectly implemented as
1229        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1230        but it seems right to fix it, and I didn't think it was worth preserving
1231        the old behaviour.
1232    
1233    5.  The command line items --regex=pattern and --regexp=pattern were not
1234        recognized by pcregrep, which required --regex pattern or --regexp pattern
1235        (with a space rather than an '='). The man page documented the '=' forms,
1236        which are compatible with GNU grep; these now work.
1237    
1238    6.  No libpcreposix.pc file was created for pkg-config; there was just
1239        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1240    
1241    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1242        when UCP support is not needed, by modifying the Python script that
1243        generates it from Unicode data files. This should not matter if the module
1244        is correctly used as a library, but I received one complaint about 50K of
1245        unwanted data. My guess is that the person linked everything into his
1246        program rather than using a library. Anyway, it does no harm.
1247    
1248    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1249        was a minimum greater than 1 for a wide character in a possessive
1250        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1251        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1252        character. Chaos in the form of incorrect output or a compiling loop could
1253        result.
1254    
1255    9.  The restrictions on what a pattern can contain when partial matching is
1256        requested for pcre_exec() have been removed. All patterns can now be
1257        partially matched by this function. In addition, if there are at least two
1258        slots in the offset vector, the offset of the earliest inspected character
1259        for the match and the offset of the end of the subject are set in them when
1260        PCRE_ERROR_PARTIAL is returned.
1261    
1262    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1263        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1264        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1265        and may be more useful for multi-segment matching.
1266    
1267    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1268        used to be given if ever the end of the subject was reached; now it is
1269        given only if matching could not proceed because another character was
1270        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1271        string "Z", which now yields "no match" instead of "partial match". In the
1272        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1273        final character ended with (*FAIL).
1274    
1275    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1276        if the pattern had a "must contain" character that was already found in the
1277        earlier partial match, unless partial matching was again requested. For
1278        example, with the pattern /dog.(body)?/, the "must contain" character is
1279        "g". If the first part-match was for the string "dog", restarting with
1280        "sbody" failed. This bug has been fixed.
1281    
1282    13. The string returned by pcre_dfa_exec() after a partial match has been
1283        changed so that it starts at the first inspected character rather than the
1284        first character of the match. This makes a difference only if the pattern
1285        starts with a lookbehind assertion or \b or \B (\K is not supported by
1286        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1287        matching functions compatible, and I think it's the right thing to do.
1288    
1289    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1290        so that the demonstration program is easily available in environments where
1291        PCRE has not been installed from source.
1292    
1293    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1294        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1295        library.
1296    
1297    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1298        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1299        is not the first non-POSIX option to be added. Clearly some people find
1300        these options useful.
1301    
1302    17. If a caller to the POSIX matching function regexec() passes a non-zero
1303        value for nmatch with a NULL value for pmatch, the value of
1304        nmatch is forced to zero.
1305    
1306    18. RunGrepTest did not have a test for the availability of the -u option of
1307        the diff command, as RunTest does. It now checks in the same way as
1308        RunTest, and also checks for the -b option.
1309    
1310    19. If an odd number of negated classes containing just a single character
1311        interposed, within parentheses, between a forward reference to a named
1312        subpattern and the definition of the subpattern, compilation crashed with
1313        an internal error, complaining that it could not find the referenced
1314        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1315        [The bug was that it was starting one character too far in when skipping
1316        over the character class, thus treating the ] as data rather than
1317        terminating the class. This meant it could skip too much.]
1318    
1319    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1320        /g option in pcretest when the pattern contains \K, which makes it possible
1321        to have an empty string match not at the start, even when the pattern is
1322        anchored. Updated pcretest and pcredemo to use this option.
1323    
1324    21. If the maximum number of capturing subpatterns in a recursion was greater
1325        than the maximum at the outer level, the higher number was returned, but
1326        with unset values at the outer level. The correct (outer level) value is
1327        now given.
1328    
1329    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1330        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1331        make it do so. The string so far is captured, making this feature
1332        compatible with Perl.
1333    
1334    23. The tests have been re-organized, adding tests 11 and 12, to make it
1335        possible to check the Perl 5.10 features against Perl 5.10.
1336    
1337    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1338        pattern matches a fixed length string. PCRE did not allow this; now it
1339        does. Neither allows recursion.
1340    
1341    25. I finally figured out how to implement a request to provide the minimum
1342        length of subject string that was needed in order to match a given pattern.
1343        (It was back references and recursion that I had previously got hung up
1344        on.) This code has now been added to pcre_study(); it finds a lower bound
1345        to the length of subject needed. It is not necessarily the greatest lower
1346        bound, but using it to avoid searching strings that are too short does give
1347        some useful speed-ups. The value is available to calling programs via
1348        pcre_fullinfo().
1349    
1350    26. While implementing 25, I discovered to my embarrassment that pcretest had
1351        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1352        study optimizations had never been tested with that matching function.
1353        Oops. What is worse, even when it was passed study data, there was a bug in
1354        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1355        were also very few tests of studied patterns with pcre_dfa_exec().
1356    
1357    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1358        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1359        on the other side of the coin, they are no longer allowed to have different
1360        names, because these cannot be distinguished in PCRE, and this has caused
1361        confusion. (This is a difference from Perl.)
1362    
1363    28. When duplicate subpattern names are present (necessarily with different
1364        numbers, as required by 27 above), and a test is made by name in a
1365        conditional pattern, either for a subpattern having been matched, or for
1366        recursion in such a pattern, all the associated numbered subpatterns are
1367        tested, and the overall condition is true if the condition is true for any
1368        one of them. This is the way Perl works, and is also more like the way
1369        testing by number works.
1370    
1371    
1372    Version 7.9 11-Apr-09
1373    ---------------------
1374    
1375    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1376        (pcretest), all targets were linked against these libraries. This included
1377        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1378        libraries. This caused unwanted dependencies to be created. This problem
1379        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1380        pcretest is linked with readline.
1381    
1382    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1383        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1384        moved outside it again, because FALSE and TRUE are already defined in AIX,
1385        but BOOL is not.
1386    
1387    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1388        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1389    
1390    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1391        hyphens) following file names and line numbers when outputting matching
1392        lines. This is not true; no spaces are inserted. I have also clarified the
1393        wording for the --colour (or --color) option.
1394    
1395    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1396        was not coloured; this is different to GNU grep, so I have changed it to be
1397        the same.
1398    
1399    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1400        each matching line was coloured. Now it goes on to look for further matches
1401        of any of the test patterns, which is the same behaviour as GNU grep.
1402    
1403    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1404        doesn't make sense to accept an empty string match in pcregrep, so I have
1405        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1406        seems to be how GNU grep behaves.
1407    
1408    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1409        start or after a newline", because the conditional assertion was not being
1410        correctly handled. The rule now is that both the assertion and what follows
1411        in the first alternative must satisfy the test.
1412    
1413    9.  If auto-callout was enabled in a pattern with a conditional group whose
1414        condition was an assertion, PCRE could crash during matching, both with
1415        pcre_exec() and pcre_dfa_exec().
1416    
1417    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1418        used for matching.
1419    
1420    11. Unicode property support in character classes was not working for
1421        characters (bytes) greater than 127 when not in UTF-8 mode.
1422    
1423    12. Added the -M command line option to pcretest.
1424    
1425    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1426    
1427    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1428    
1429    16. Added comments and documentation about mis-use of no_arg in the C++
1430        wrapper.
1431    
1432    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1433        from Martin Jerabek that uses macro names for all relevant character and
1434        string constants.
1435    
1436    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1437        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1438        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1439        these, but not everybody uses configure.
1440    
1441    19. A conditional group that had only one branch was not being correctly
1442        recognized as an item that could match an empty string. This meant that an
1443        enclosing group might also not be so recognized, causing infinite looping
1444        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1445        with the subject "ab", where knowledge that the repeated group can match
1446        nothing is needed in order to break the loop.
1447    
1448    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1449        exec(), but without supplying a callout function, matching went wrong.
1450    
1451    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1452        leak if the size of the offset vector was greater than 30. When the vector
1453        is smaller, the saved offsets during recursion go onto a local stack
1454        vector, but for larger vectors malloc() is used. It was failing to free
1455        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1456        error, in fact).
1457    
1458    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1459        heapframe that is used only when UTF-8 support is enabled. This caused no
1460        problem, but was untidy.
1461    
1462    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1463        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1464        included within another project.
1465    
1466    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1467        slightly modified by me:
1468    
1469          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1470              not building pcregrep.
1471    
1472          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1473              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1474    
1475    25. Forward references, both numeric and by name, in patterns that made use of
1476        duplicate group numbers, could behave incorrectly or give incorrect errors,
1477        because when scanning forward to find the reference group, PCRE was not
1478        taking into account the duplicate group numbers. A pattern such as
1479        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1480    
1481    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1482        the feature of a custom pointer more persuasive (as requested by a user).
1483    
1484    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1485        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1486        pre-defined.
1487    
1488    28. Added support for (*UTF8) at the start of a pattern.
1489    
1490    29. Arrange for flags added by the "release type" setting in CMake to be shown
1491        in the configuration summary.
1492    
1493    
1494    Version 7.8 05-Sep-08
1495    ---------------------
1496    
1497    1.  Replaced UCP searching code with optimized version as implemented for Ad
1498        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1499        stage table and inline lookup instead of a function, giving speed ups of 2
1500        to 5 times on some simple patterns that I tested. Permission was given to
1501        distribute the MultiStage2.py script that generates the tables (it's not in
1502        the tarball, but is in the Subversion repository).
1503    
1504    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1505        scripts.
1506    
1507    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1508        a group with a zero qualifier. The result of the study could be incorrect,
1509        or the function might crash, depending on the pattern.
1510    
1511    4.  Caseless matching was not working for non-ASCII characters in back
1512        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1513        It now works when Unicode Property Support is available.
1514    
1515    5.  In pcretest, an escape such as \x{de} in the data was always generating
1516        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1517        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1518        truncation.
1519    
1520    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1521    
1522    7.  Added two (int) casts to pcregrep when printing the difference of two
1523        pointers, in case they are 64-bit values.
1524    
1525    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1526        test 2 if it fails.
1527    
1528    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1529        and a #define of that name to empty if it is not externally set. This is to
1530        allow users of MSVC to set it if necessary.
1531    
1532    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1533        the convenience functions in the pcre_get.c source file.
1534    
1535    11. An option change at the start of a pattern that had top-level alternatives
1536        could cause overwriting and/or a crash. This command provoked a crash in
1537        some environments:
1538    
1539          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1540    
1541        This potential security problem was recorded as CVE-2008-2371.
1542    
1543    12. For a pattern where the match had to start at the beginning or immediately
1544        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1545        pcre_dfa_exec() could read past the end of the passed subject if there was
1546        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1547        pcretest so that it places the subject at the end of its malloc-ed buffer.
1548    
1549    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1550        exec() might read past the end of the data buffer in UTF-8 mode.
1551    
1552    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1553        the data contained the byte 0x85 as part of a UTF-8 character within its
1554        first line. This applied both to normal and DFA matching.
1555    
1556    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1557        /^[^d]*?$/8 failed to match "abc".
1558    
1559    16. Added a missing copyright notice to pcrecpp_internal.h.
1560    
1561    17. Make it more clear in the documentation that values returned from
1562        pcre_exec() in ovector are byte offsets, not character counts.
1563    
1564    18. Tidied a few places to stop certain compilers from issuing warnings.
1565    
1566    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1567        supplied by Stefan Weber. I made a further small update for 7.8 because
1568        there is a change of source arrangements: the pcre_searchfuncs.c module is
1569        replaced by pcre_ucd.c.
1570    
1571    
1572    Version 7.7 07-May-08
1573    ---------------------
1574    
1575    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1576        a string to a long long, pretend we don't even have a long long." This is
1577        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1578    
1579    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1580        pre-7.6 versions, which defined a global no_arg variable instead of putting
1581        it in the RE class. (See also #8 below.)
1582    
1583    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1584        Lopes.
1585    
1586    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1587    
1588        (1) The include/exclude patterns were being applied to the whole pathnames
1589            of files, instead of just to the final components.
1590    
1591        (2) If there was more than one level of directory, the subdirectories were
1592            skipped unless they satisfied the include/exclude conditions. This is
1593            inconsistent with GNU grep (and could even be seen as contrary to the
1594            pcregrep specification - which I improved to make it absolutely clear).
1595            The action now is always to scan all levels of directory, and just
1596            apply the include/exclude patterns to regular files.
1597    
1598    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1599        --exclude_dir in the tests to avoid scanning .svn directories.
1600    
1601    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1602        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1603        doesn't support NULs in patterns.
1604    
1605    7.  Added some missing "const"s to declarations of static tables in
1606        pcre_compile.c and pcre_dfa_exec.c.
1607    
1608    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1609        caused by fix #2  above. (Subsequently also a second patch to fix the
1610        first patch. And a third patch - this was a messy problem.)
1611    
1612    9.  Applied Craig's patch to remove the use of push_back().
1613    
1614    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1615        matching function regexec().
1616    
1617    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1618        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1619        references. PCRE supports relative numbers with this syntax (I don't think
1620        Oniguruma does).
1621    
1622    12. Previously, a group with a zero repeat such as (...){0} was completely
1623        omitted from the compiled regex. However, this means that if the group
1624        was called as a subroutine from elsewhere in the pattern, things went wrong
1625        (an internal error was given). Such groups are now left in the compiled
1626        pattern, with a new opcode that causes them to be skipped at execution
1627        time.
1628    
1629    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1630        to the way PCRE behaves:
1631    
1632        (a) A lone ] character is dis-allowed (Perl treats it as data).
1633    
1634        (b) A back reference to an unmatched subpattern matches an empty string
1635            (Perl fails the current match path).
1636    
1637        (c) A data ] in a character class must be notated as \] because if the
1638            first data character in a class is ], it defines an empty class. (In
1639            Perl it is not possible to have an empty class.) The empty class []
1640            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1641            The negative empty class [^] matches any one character, independently
1642            of the DOTALL setting.
1643    
1644    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1645        non-existent subpattern following a character class starting with ']' and
1646        containing () gave an internal compiling error instead of "reference to
1647        non-existent subpattern". Fortunately, when the pattern did exist, the
1648        compiled code was correct. (When scanning forwards to check for the
1649        existencd of the subpattern, it was treating the data ']' as terminating
1650        the class, so got the count wrong. When actually compiling, the reference
1651        was subsequently set up correctly.)
1652    
1653    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1654        it was being rejected as not supported by pcre_dfa_exec(), even though
1655        other assertions are supported. I have made pcre_dfa_exec() support
1656        (*FAIL).
1657    
1658    16. The implementation of 13c above involved the invention of a new opcode,
1659        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1660        cannot be changed at match time, I realized I could make a small
1661        improvement to matching performance by compiling OP_ALLANY instead of
1662        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1663        on the OP_ANY path.
1664    
1665    17. Compiling pcretest on Windows with readline support failed without the
1666        following two fixes: (1) Make the unistd.h include conditional on
1667        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1668    
1669    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1670        ncurses library to be included for pcretest when ReadLine support is
1671        requested, but also to allow for it to be overridden. This patch came from
1672        Daniel Bergström.
1673    
1674    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1675        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1676        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1677        spotting this.
1678    
1679    
1680    Version 7.6 28-Jan-08
1681    ---------------------
1682    
1683    1.  A character class containing a very large number of characters with
1684        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1685        overflow.
1686    
1687    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1688        HAVE_LONG_LONG is not defined.
1689    
1690    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1691        bring them up to date and include new features. This patch includes:
1692    
1693        - Fixed PH's badly added libz and libbz2 support.
1694        - Fixed a problem with static linking.
1695        - Added pcredemo. [But later removed - see 7 below.]
1696        - Fixed dftables problem and added an option.
1697        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1698            HAVE_LONG_LONG.
1699        - Added readline support for pcretest.
1700        - Added an listing of the option settings after cmake has run.
1701    
1702    4.  A user submitted a patch to Makefile that makes it easy to create
1703        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1704        Makefile.am that cause it to include this special target, without
1705        affecting anything else. Note that the same mingw target plus all
1706        the other distribution libraries and programs are now supported
1707        when configuring with CMake (see 6 below) instead of with
1708        Configure/Make.
1709    
1710    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1711        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1712        exported in the Windows port". It has not yet been confirmed that the patch
1713        solves the problem, but it does no harm.
1714    
1715    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1716        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1717        with CMake, and also correct the comment about stack recursion.
1718    
1719    7.  Remove the automatic building of pcredemo from the ./configure system and
1720        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1721        of a program that users should build themselves after PCRE is installed, so
1722        building it automatically is not really right. What is more, it gave
1723        trouble in some build environments.
1724    
1725    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1726    
1727    
1728    Version 7.5 10-Jan-08
1729    ---------------------
1730    
1731    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1732        values in parens when parsing an RE using the C++ wrapper."
1733    
1734    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1735        Characters greater than 255 were excluded from the class instead of being
1736        included.
1737    
1738    3.  The same bug as (2) above applied to negated POSIX classes such as
1739        [:^space:].
1740    
1741    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1742        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1743        I have changed it.
1744    
1745    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1746        first named subpattern) and a construct such as (?&a) would reference the
1747        first named subpattern whose name started with "a" (in other words, the
1748        length check was missing). Both these problems are fixed. "Subpattern name
1749        expected" is now given for (?&) (a zero-length name), and this patch also
1750        makes it give the same error for \k'' (previously it complained that that
1751        was a reference to a non-existent subpattern).
1752    
1753    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1754        this is right because (?- can be followed by option settings as well as by
1755        digits. I have, however, made the messages clearer.
1756    
1757    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1758        than the number used in the conditional) now cause a compile-time error.
1759        This is actually not compatible with Perl, which accepts such patterns, but
1760        treats the conditional as always being FALSE (as PCRE used to), but it
1761        seems to me that giving a diagnostic is better.
1762    
1763    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1764        and messages.
1765    
1766    9.  Fix two occurrences of "backslash" in comments that should have been
1767        "backspace".
1768    
1769    10. Remove two redundant lines of code that can never be obeyed (their function
1770        was moved elsewhere).
1771    
1772    11. The program that makes PCRE's Unicode character property table had a bug
1773        which caused it to generate incorrect table entries for sequences of
1774        characters that have the same character type, but are in different scripts.
1775        It amalgamated them into a single range, with the script of the first of
1776        them. In other words, some characters were in the wrong script. There were
1777        thirteen such cases, affecting characters in the following ranges:
1778    
1779          U+002b0 - U+002c1
1780          U+0060c - U+0060d
1781          U+0061e - U+00612
1782          U+0064b - U+0065e
1783          U+0074d - U+0076d
1784          U+01800 - U+01805
1785          U+01d00 - U+01d77
1786          U+01d9b - U+01dbf
1787          U+0200b - U+0200f
1788          U+030fc - U+030fe
1789          U+03260 - U+0327f
1790          U+0fb46 - U+0fbb1
1791          U+10450 - U+1049d
1792    
1793    12. The -o option (show only the matching part of a line) for pcregrep was not
1794        compatible with GNU grep in that, if there was more than one match in a
1795        line, it showed only the first of them. It now behaves in the same way as
1796        GNU grep.
1797    
1798    13. If the -o and -v options were combined for pcregrep, it printed a blank
1799        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1800        does the same. The return code can be used to tell if there were any
1801        non-matching lines.
1802    
1803    14. Added --file-offsets and --line-offsets to pcregrep.
1804    
1805    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1806        infinitely looping recursion. The bug was that positive lookaheads were not
1807        being skipped when checking for a possible empty match (negative lookaheads
1808        and both kinds of lookbehind were skipped).
1809    
1810    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1811        inclusion of <windows.h> to before rather than after the definition of
1812        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1813    
1814    17. Specifying a possessive quantifier with a specific limit for a Unicode
1815        character property caused pcre_compile() to compile bad code, which led at
1816        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1817        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1818        caused the error; without that there was no problem.
1819    
1820    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1821    
1822    19. Added --enable-pcretest-libreadline.
1823    
1824    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1825        RE::GlobalReplace(). As a result, the number of replacements returned was
1826        double what it should be. I removed one of the increments, but Craig sent a
1827        later patch that removed the other one (the right fix) and added unit tests
1828        that check the return values (which was not done before).
1829    
1830    21. Several CMake things:
1831    
1832        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1833            the names libpcre and libpcreposix, not just pcre and pcreposix.
1834    
1835        (2) The above change means that pcretest and pcregrep are now correctly
1836            linked with the newly-built libraries, not previously installed ones.
1837    
1838        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1839    
1840    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1841        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1842        UTF-8 newline character). The key issue is that the pattern starts .*;
1843        this means that the match must be either at the beginning, or after a
1844        newline. The bug was in the code for advancing after a failed match and
1845        checking that the new position followed a newline. It was not taking
1846        account of UTF-8 characters correctly.
1847    
1848    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1849        character classes. PCRE was not treating the sequence [:...:] as a
1850        character class unless the ... were all letters. Perl, however, seems to
1851        allow any characters between [: and :], though of course it rejects as
1852        unknown any "names" that contain non-letters, because all the known class
1853        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1854        for example, whereas PCRE did not - it did not recognize a POSIX character
1855        class. This seemed a bit dangerous, so the code has been changed to be
1856        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1857        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1858        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1859        Perl does, and where it didn't before.
1860    
1861    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1862        Windows environments %n is disabled by default.
1863    
1864    
1865    Version 7.4 21-Sep-07
1866    ---------------------
1867    
1868    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1869        means that a class such as [\s] counted as "explicit reference to CR or
1870        LF". That isn't really right - the whole point of the change was to try to
1871        help when there was an actual mention of one of the two characters. So now
1872        the change happens only if \r or \n (or a literal CR or LF) character is
1873        encountered.
1874    
1875    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1876        of both had grown to the point where there were only 3 bits left.
1877        Fortunately, there was spare space in the data structure, and so I have
1878        moved the internal flags into a new 16-bit field to free up more option
1879        bits.
1880    
1881    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1882        but did not set the internal JCHANGED flag - either of these is enough to
1883        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1884        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1885        start sets both bits.
1886    
1887    4.  Added options (at build time, compile time, exec time) to change \R from
1888        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1889    
1890    5.  doc/pcresyntax.html was missing from the distribution.
1891    
1892    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1893        compatibility, even though it is no longer used.
1894    
1895    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1896        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1897        windows.h file is present (where different names are used). [This was
1898        reversed later after testing - see 16 below.]
1899    
1900    8.  Changed all #include <config.h> to #include "config.h". There were also
1901        some further <pcre.h> cases that I changed to "pcre.h".
1902    
1903    9.  When pcregrep was used with the --colour option, it missed the line ending
1904        sequence off the lines that it output.
1905    
1906    10. It was pointed out to me that arrays of string pointers cause lots of
1907        relocations when a shared library is dynamically loaded. A technique of
1908        using a single long string with a table of offsets can drastically reduce
1909        these. I have refactored PCRE in four places to do this. The result is
1910        dramatic:
1911    
1912          Originally:                          290
1913          After changing UCP table:            187
1914          After changing error message table:   43
1915          After changing table of "verbs"       36
1916          After changing table of Posix names   22
1917    
1918        Thanks to the folks working on Gregex for glib for this insight.
1919    
1920    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1921        unicode-properties was also set.
1922    
1923    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1924    
1925    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1926        checked only for CRLF.
1927    
1928    14. Added casts to pcretest.c to avoid compiler warnings.
1929    
1930    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1931    
1932    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1933        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1934        entirely. This removes changes made in 7 above.
1935    
1936    17. The CMake files have been updated, and there is now more information about
1937        building with CMake in the NON-UNIX-USE document.
1938    
1939    
1940    Version 7.3 28-Aug-07
1941    ---------------------
1942    
1943     1. In the rejigging of the build system that eventually resulted in 7.1, the
1944        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1945        brackets there is not right, since it causes compilers to look for an
1946        installed pcre.h, not the version that is in the source that is being
1947        compiled (which of course may be different). I have changed it back to:
1948    
1949          #include "pcre.h"
1950    
1951        I have a vague recollection that the change was concerned with compiling in
1952        different directories, but in the new build system, that is taken care of
1953        by the VPATH setting the Makefile.
1954    
1955     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1956        when the subject happened to end in the byte 0x85 (e.g. if the last
1957        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1958        characters but of course it shouldn't be taken as a newline when it is part
1959        of another character. The bug was that, for an unlimited repeat of . in
1960        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1961        characters when looking for a newline.
1962    
1963     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1964    
1965     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1966        in debug output.
1967    
1968     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1969        long printing in the pcrecpp unittest when running under MinGW.
1970    
1971     6. ESC_K was left out of the EBCDIC table.
1972    
1973     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1974        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1975        limit also applies to "virtual nesting" when a pattern is recursive, and in
1976        this case 1000 isn't so big. I have been able to remove this limit at the
1977        expense of backing off one optimization in certain circumstances. Normally,
1978        when pcre_exec() would call its internal match() function recursively and
1979        immediately return the result unconditionally, it uses a "tail recursion"
1980        feature to save stack. However, when a subpattern that can match an empty
1981        string has an unlimited repetition quantifier, it no longer makes this
1982        optimization. That gives it a stack frame in which to save the data for
1983        checking that an empty string has been matched. Previously this was taken
1984        from the 1000-entry workspace that had been reserved. So now there is no
1985        explicit limit, but more stack is used.
1986    
1987     8. Applied Daniel's patches to solve problems with the import/export magic
1988        syntax that is required for Windows, and which was going wrong for the
1989        pcreposix and pcrecpp parts of the library. These were overlooked when this
1990        problem was solved for the main library.
1991    
1992     9. There were some crude static tests to avoid integer overflow when computing
1993        the size of patterns that contain repeated groups with explicit upper
1994        limits. As the maximum quantifier is 65535, the maximum group length was
1995        set at 30,000 so that the product of these two numbers did not overflow a
1996        32-bit integer. However, it turns out that people want to use groups that
1997        are longer than 30,000 bytes (though not repeat them that many times).
1998        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1999        made it possible to implement the integer overflow checks in a much more
2000        dynamic way, which I have now done. The artificial limitation on group
2001        length has been removed - we now have only the limit on the total length of
2002        the compiled pattern, which depends on the LINK_SIZE setting.
2003    
2004    10. Fixed a bug in the documentation for get/copy named substring when
2005        duplicate names are permitted. If none of the named substrings are set, the
2006        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2007        empty string.
2008    
2009    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2010        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2011        because the ] is interpreted as the first data character and the
2012        terminating ] is not found. PCRE has been made compatible with Perl in this
2013        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2014        cause memory overwriting.
2015    
2016    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2017        string has been matched (to stop an infinite loop). It was not recognizing
2018        a conditional subpattern that could match an empty string if that
2019        subpattern was within another subpattern. For example, it looped when
2020        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2021        condition was not nested. This bug has been fixed.
2022    
2023    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2024        past the start of the subject in the presence of bytes with the top bit
2025        set, for example "\x8aBCD".
2026    
2027    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2028        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2029    
2030    14. Optimized (?!) to (*FAIL).
2031    
2032    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2033        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2034        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2035        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2036        does: it's just the validity check that is more restrictive.
2037    
2038    16. Inserted checks for integer overflows during escape sequence (backslash)
2039        processing, and also fixed erroneous offset values for syntax errors during
2040        backslash processing.
2041    
2042    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2043        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2044    
2045    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2046        caused an overrun.
2047    
2048    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2049        something other than just ASCII characters) inside a group that had an
2050        unlimited repeat caused a loop at compile time (while checking to see
2051        whether the group could match an empty string).
2052    
2053    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2054        [\P{Any}] did so. (Error in the code for printing property names.)
2055    
2056    21. An orphan \E inside a character class could cause a crash.
2057    
2058    22. A repeated capturing bracket such as (A)? could cause a wild memory
2059        reference during compilation.
2060    
2061    23. There are several functions in pcre_compile() that scan along a compiled
2062        expression for various reasons (e.g. to see if it's fixed length for look
2063        behind). There were bugs in these functions when a repeated \p or \P was
2064        present in the pattern. These operators have additional parameters compared
2065        with \d, etc, and these were not being taken into account when moving along
2066        the compiled data. Specifically:
2067    
2068        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2069            length.
2070    
2071        (b) An item such as \pL+ within a repeated group could cause crashes or
2072            loops.
2073    
2074        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2075            "reference to non-existent subpattern" error.
2076    
2077        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2078    
2079    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2080        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2081    
2082    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2083    
2084    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2085        character were causing crashes (broken optimization).
2086    
2087    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2088        \p or \P) caused a compile-time loop.
2089    
2090    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2091        break. For example, the unstudied pattern [\r\n]A does not match the string
2092        "\r\nA" because change 7.0/46 below moves the current point on by two
2093        characters after failing to match at the start. However, the pattern \nA
2094        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2095        the same is true. There doesn't seem any very clean way out of this, but
2096        what I have chosen to do makes the common cases work: PCRE now takes note
2097        of whether there can be an explicit match for \r or \n anywhere in the
2098        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2099        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2100        pattern has explicit CR or LF references.
2101    
2102    29. Added (*CR) etc for changing newline setting at start of pattern.
2103    
2104    
2105    Version 7.2 19-Jun-07
2106    ---------------------
2107    
2108     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2109        which is apparently normally available under Windows.
2110    
2111     2. Re-jig the pcregrep tests with different newline settings in an attempt
2112        to make them independent of the local environment's newline setting.
2113    
2114     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2115    
2116     4. Some of the "internals" tests were previously cut out when the link size
2117        was not 2, because the output contained actual offsets. The recent new
2118        "Z" feature of pcretest means that these can be cut out, making the tests
2119        usable with all link sizes.
2120    
2121     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2122        stack recursion. This gives a massive performance boost under BSD, but just
2123        a small improvement under Linux. However, it saves one field in the frame
2124        in all cases.
2125    
2126     6. Added more features from the forthcoming Perl 5.10:
2127    
2128        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2129            recursion call. It refers to the nth most recently opened parentheses.
2130    
2131        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2132            to be opened parentheses.
2133    
2134        (c) Conditions that refer to capturing parentheses can be specified
2135            relatively, for example, (?(-2)... or (?(+3)...
2136    
2137        (d) \K resets the start of the current match so that everything before
2138            is not part of it.
2139    
2140        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2141    
2142        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2143            reference syntax.
2144    
2145        (g) (?| introduces a group in which the numbering of parentheses in each
2146            alternative starts with the same number.
2147    
2148        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2149    
2150     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2151        PCRE_INFO_JCHANGED.
2152    
2153     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2154        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2155        for detecting groups that can match an empty string.
2156    
2157     9. A pattern with a very large number of alternatives (more than several
2158        hundred) was running out of internal workspace during the pre-compile
2159        phase, where pcre_compile() figures out how much memory will be needed. A
2160        bit of new cunning has reduced the workspace needed for groups with
2161        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2162        workspace instead of running out of the 4096 that are available.
2163    
2164    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2165    
2166    11. Applied patch from Google to remove an optimization that didn't quite work.
2167        The report of the bug said:
2168    
2169          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2170          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2171          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2172    
2173    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2174        it matched the wrong number of bytes.
2175    
2176    
2177    Version 7.1 24-Apr-07
2178    ---------------------
2179    
2180     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2181        that is more "standard", making use of automake and other Autotools. There
2182        is some re-arrangement of the files and adjustment of comments consequent
2183        on this.
2184    
2185     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2186        for recursive directory scanning broke on some systems because the files
2187        are not scanned in any specific order and on different systems the order
2188        was different. A call to "sort" has been inserted into RunGrepTest for the
2189        approprate test as a short-term fix. In the longer term there may be an
2190        alternative.
2191    
2192     3. I had an email from Eric Raymond about problems translating some of PCRE's
2193        man pages to HTML (despite the fact that I distribute HTML pages, some
2194        people do their own conversions for various reasons). The problems
2195        concerned the use of low-level troff macros .br and .in. I have therefore
2196        removed all such uses from the man pages (some were redundant, some could
2197        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2198        HTML has been updated to handle .nf/.fi and to complain if it encounters
2199        .br or .in.
2200    
2201     4. Updated comments in configure.ac that get placed in config.h.in and also
2202        arranged for config.h to be included in the distribution, with the name
2203        config.h.generic, for the benefit of those who have to compile without
2204        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2205    
2206     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2207        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2208        makevp.bat for the current PCRE, using the additional files
2209        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2210    
2211     6. A Windows user reported a minor discrepancy with test 2, which turned out
2212        to be caused by a trailing space on an input line that had got lost in his
2213        copy. The trailing space was an accident, so I've just removed it.
2214    
2215     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2216        that is needed.
2217    
2218     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2219        as "const" (a) because they are and (b) because it helps the PHP
2220        maintainers who have recently made a script to detect big data structures
2221        in the php code that should be moved to the .rodata section. I remembered
2222        to update Builducptable as well, so it won't revert if ucptable.h is ever
2223        re-created.
2224    
2225     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2226        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2227        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2228        support is not required. This saves 1.5-2K of code, which is important in
2229        some applications.
2230    
2231        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2232        so as not to refer to the tables, even though these functions will never be
2233        called when UTF-8 support is disabled. Otherwise there are problems with a
2234        shared library.
2235    
2236    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2237    
2238        (a) It was defining its arguments as char * instead of void *.
2239    
2240        (b) It was assuming that all moves were upwards in memory; this was true
2241            a long time ago when I wrote it, but is no longer the case.
2242    
2243        The emulated memove() is provided for those environments that have neither
2244        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2245        is clearly not the case, as these two bugs were recently reported.
2246    
2247    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2248        and Detrail to create the HTML documentation, the .txt form of the man
2249        pages, and it removes trailing spaces from listed files. It also creates
2250        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2251        case, it wraps all the #defines with #ifndefs. This script should be run
2252        before "make dist".
2253    
2254    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2255        with Unicode property support.
2256    
2257        (a) For a maximizing quantifier, if the two different cases of the
2258            character were of different lengths in their UTF-8 codings (there are
2259            some cases like this - I found 11), and the matching function had to
2260            back up over a mixture of the two cases, it incorrectly assumed they
2261            were both the same length.
2262    
2263        (b) When PCRE was configured to use the heap rather than the stack for
2264            recursion during matching, it was not correctly preserving the data for
2265            the other case of a UTF-8 character when checking ahead for a match
2266            while processing a minimizing repeat. If the check also involved
2267            matching a wide character, but failed, corruption could cause an
2268            erroneous result when trying to check for a repeat of the original
2269            character.
2270    
2271    13. Some tidying changes to the testing mechanism:
2272    
2273        (a) The RunTest script now detects the internal link size and whether there
2274            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2275            values substituted by "configure". (The RunGrepTest script already did
2276            this for UTF-8.) The configure.ac script no longer substitutes the
2277            relevant variables.
2278    
2279        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2280            with length and offset values. This means that the output is different
2281            for different internal link sizes. Test 2 is skipped for link sizes
2282            other than 2 because of this, bypassing the problem. Unfortunately,
2283            there was also a test in test 3 (the locale tests) that used /B and
2284            failed for link sizes other than 2. Rather than cut the whole test out,
2285            I have added a new /Z option to pcretest that replaces the length and
2286            offset values with spaces. This is now used to make test 3 independent
2287            of link size. (Test 2 will be tidied up later.)
2288    
2289    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2290        segmentation fault instead of returning the appropriate error message.
2291    
2292    15. In multiline mode when the newline sequence was set to "any", the pattern
2293        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2294        This doesn't seem right; it now treats the CRLF combination as the line
2295        ending, and so does not match in that case. It's only a pattern such as ^$
2296        that would hit this one: something like ^ABC$ would have failed after \r
2297        and then tried again after \r\n.
2298    
2299    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2300        in an attempt to make files that differ only in their line terminators
2301        compare equal. This works on Linux.
2302    
2303    17. Under certain error circumstances pcregrep might try to free random memory
2304        as it exited. This is now fixed, thanks to valgrind.
2305    
2306    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2307        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2308        was because its rules for how to advance for /g after matching an empty
2309        string at the end of a line did not allow for this case. They now check for
2310        it specially.
2311    
2312    20. pcretest is supposed to handle patterns and data of any length, by
2313        extending its buffers when necessary. It was getting this wrong when the
2314        buffer for a data line had to be extended.
2315    
2316    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2317        CRLF as a newline sequence.
2318    
2319    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2320        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2321        I have nevertheless tidied it up.
2322    
2323    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2324    
2325    24. Added a man page for pcre-config.
2326    
2327    
2328    Version 7.0 19-Dec-06
2329    ---------------------
2330    
2331     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2332        moving to gcc 4.1.1.
2333    
2334     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2335        sys/time.h, which is documented as needed for this function. It doesn't
2336        seem to matter on Linux, but it showed up on some releases of OS X.
2337    
2338     3. It seems that there are systems where bytes whose values are greater than
2339        127 match isprint() in the "C" locale. The "C" locale should be the
2340        default when a C program starts up. In most systems, only ASCII printing
2341        characters match isprint(). This difference caused the output from pcretest
2342        to vary, making some of the tests fail. I have changed pcretest so that:
2343    
2344        (a) When it is outputting text in the compiled version of a pattern, bytes
2345            other than 32-126 are always shown as hex escapes.
2346    
2347        (b) When it is outputting text that is a matched part of a subject string,
2348            it does the same, unless a different locale has been set for the match
2349            (using the /L modifier). In this case, it uses isprint() to decide.
2350    
2351     4. Fixed a major bug that caused incorrect computation of the amount of memory
2352        required for a compiled pattern when options that changed within the
2353        pattern affected the logic of the preliminary scan that determines the
2354        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2355        that the computed length was too small. The symptoms of this bug were
2356        either the PCRE error "internal error: code overflow" from pcre_compile(),
2357        or a glibc crash with a message such as "pcretest: free(): invalid next
2358        size (fast)". Examples of patterns that provoked this bug (shown in
2359        pcretest format) are:
2360    
2361          /(?-x: )/x
2362          /(?x)(?-x: \s*#\s*)/
2363          /((?i)[\x{c0}])/8
2364          /(?i:[\x{c0}])/8
2365    
2366        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2367        is now done differently.
2368    
2369     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2370        wrapper classes; (b) implement a new function in the C++ scanner that is
2371        more efficient than the old way of doing things because it avoids levels of
2372        recursion in the regex matching; (c) add a paragraph to the documentation
2373        for the FullMatch() function.
2374    
2375     6. The escape sequence \n was being treated as whatever was defined as
2376        "newline". Not only was this contrary to the documentation, which states
2377        that \n is character 10 (hex 0A), but it also went horribly wrong when
2378        "newline" was defined as CRLF. This has been fixed.
2379    
2380     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2381        was being set to -1 for the "end of line" case (supposedly a value that no
2382        character can have). Though this value is never used (the check for end of
2383        line is "zero bytes in current character"), it caused compiler complaints.
2384        I've changed it to 0xffffffff.
2385    
2386     8. In pcre_version.c, the version string was being built by a sequence of
2387        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2388        string (as it is for production releases) called a macro with an empty
2389        argument. The C standard says the result of this is undefined. The gcc
2390        compiler treats it as an empty string (which was what was wanted) but it is
2391        reported that Visual C gives an error. The source has been hacked around to
2392        avoid this problem.
2393    
2394     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2395        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2396        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2397        of them did).
2398    
2399    10. Originally, pcretest opened its input and output without "b"; then I was
2400        told that "b" was needed in some environments, so it was added for release
2401        5.0 to both the input and output. (It makes no difference on Unix-like
2402        systems.) Later I was told that it is wrong for the input on Windows. I've
2403        now abstracted the modes into two macros, to make it easier to fiddle with
2404        them, and removed "b" from the input mode under Windows.
2405    
2406    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2407    
2408    12. Added -help and --help to pcretest as an official way of being reminded
2409        of the options.
2410    
2411    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2412        and pcrecpp.cc because they annoy compilers at high warning levels.
2413    
2414    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2415    
2416    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2417        scripts are not C programs :-) and which was not noticed because it works
2418        on Linux.
2419    
2420    16. pcretest is supposed to handle any length of pattern and data line (as one
2421        line or as a continued sequence of lines) by extending its input buffer if
2422        necessary. This feature was broken for very long pattern lines, leading to
2423        a string of junk being passed to pcre_compile() if the pattern was longer
2424        than about 50K.
2425    
2426    17. I have done a major re-factoring of the way pcre_compile() computes the
2427        amount of memory needed for a compiled pattern. Previously, there was code
2428        that made a preliminary scan of the pattern in order to do this. That was
2429        OK when PCRE was new, but as the facilities have expanded, it has become
2430        harder and harder to keep it in step with the real compile phase, and there
2431        have been a number of bugs (see for example, 4 above). I have now found a
2432        cunning way of running the real compile function in a "fake" mode that
2433        enables it to compute how much memory it would need, while actually only
2434        ever using a few hundred bytes of working memory and without too many
2435        tests of the mode. This should make future maintenance and development
2436        easier. A side effect of this work is that the limit of 200 on the nesting
2437        depth of parentheses has been removed (though this was never a serious
2438        limitation, I suspect). However, there is a downside: pcre_compile() now
2439        runs more slowly than before (30% or more, depending on the pattern). I
2440        hope this isn't a big issue. There is no effect on runtime performance.
2441    
2442    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2443        newline (only possible for the last line of a file) and it was a
2444        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2445    
2446    19. Added additional timing features to pcretest. (1) The -tm option now times
2447        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2448        separate command line item, by a number that specifies the number of
2449        repeats to use when timing. The default is 50000; this gives better
2450        precision, but takes uncomfortably long for very large patterns.
2451    
2452    20. Extended pcre_study() to be more clever in cases where a branch of a
2453        subpattern has no definite first character. For example, (a*|b*)[cd] would
2454        previously give no result from pcre_study(). Now it recognizes that the
2455        first character must be a, b, c, or d.
2456    
2457    21. There was an incorrect error "recursive call could loop indefinitely" if
2458        a subpattern (or the entire pattern) that was being tested for matching an
2459        empty string contained only one non-empty item after a nested subpattern.
2460        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2461        incorrectly, because the \d was being skipped in the check.
2462    
2463    22. The pcretest program now has a new pattern option /B and a command line
2464        option -b, which is equivalent to adding /B to every pattern. This causes
2465        it to show the compiled bytecode, without the additional information that
2466        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2467        is the same as /B/I).
2468    
2469    23. A new optimization is now able automatically to treat some sequences such
2470        as a*b as a*+b. More specifically, if something simple (such as a character
2471        or a simple class like \d) has an unlimited quantifier, and is followed by
2472        something that cannot possibly match the quantified thing, the quantifier
2473        is automatically "possessified".
2474    
2475    24. A recursive reference to a subpattern whose number was greater than 39
2476        went wrong under certain circumstances in UTF-8 mode. This bug could also
2477        have affected the operation of pcre_study().
2478    
2479    25. Realized that a little bit of performance could be had by replacing
2480        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2481    
2482    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2483    
2484    27. Possessive quantifiers such as a++ were previously implemented by turning
2485        them into atomic groups such as ($>a+). Now they have their own opcodes,
2486        which improves performance. This includes the automatically created ones
2487        from 23 above.
2488    
2489    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2490        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2491        the first matched character to be a colon. This applied both to named and
2492        numbered groups.
2493    
2494    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2495    
2496    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2497        building PCRE on Windows easier, so I have included it in the distribution.
2498    
2499    31. There is now a check in pcretest against a ridiculously large number being
2500        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2501        loop, the loop is abandoned.
2502    
2503    32. Forward references to subpatterns in conditions such as (?(2)...) where
2504        subpattern 2 is defined later cause pcre_compile() to search forwards in
2505        the pattern for the relevant set of parentheses. This search went wrong
2506        when there were unescaped parentheses in a character class, parentheses
2507        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2508    
2509    33. "Subroutine" calls and backreferences were previously restricted to
2510        referencing subpatterns earlier in the regex. This restriction has now
2511        been removed.
2512    
2513    34. Added a number of extra features that are going to be in Perl 5.10. On the
2514        whole, these are just syntactic alternatives for features that PCRE had
2515        previously implemented using the Python syntax or my own invention. The
2516        other formats are all retained for compatibility.
2517    
2518        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2519            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2520            also .NET compatible.
2521    
2522        (b) A recursion or subroutine call to a named group can now be defined as
2523            (?&name) as well as (?P>name).
2524    
2525        (c) A backreference to a named group can now be defined as \k<name> or
2526            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2527            5.10, are also .NET compatible.
2528    
2529        (d) A conditional reference to a named group can now use the syntax
2530            (?(<name>) or (?('name') as well as (?(name).
2531    
2532        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2533            groups (named and numbered) that are never evaluated inline, but can be
2534            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2535            is always false. There may be only one alternative in such a group.
2536    
2537        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2538            as the simple (?(R). The condition is true only if the most recent
2539            recursion is that of the given number or name. It does not search out
2540            through the entire recursion stack.
2541    
2542        (g) The escape \gN or \g{N} has been added, where N is a positive or
2543            negative number, specifying an absolute or relative reference.
2544    
2545    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2546        some "unreachable code" warnings.
2547    
2548    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2549        things, this adds five new scripts.
2550    
2551    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2552        There were also incompatibilities regarding the handling of \Q..\E inside
2553        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2554        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2555    
2556    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2557        matches an empty string, and forcibly breaks the loop. There were bugs in
2558        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2559        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2560        separate and independent bugs (that affected different cases) have been
2561        fixed.
2562    
2563    39. Refactored the code to abolish the use of different opcodes for small
2564        capturing bracket numbers. This is a tidy that I avoided doing when I
2565        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2566        The new approach is not only tidier, it makes it possible to reduce the
2567        memory needed to fix the previous bug (38).
2568    
2569    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2570        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2571        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2572        mode.
2573    
2574    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2575        report.
2576    
2577    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2578        copy construction and assignment in the C++ wrapper.
2579    
2580    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2581        couple of bugs that could have given wrong results in the "--newline=crlf"
2582        case.
2583    
2584    44. Added a number of casts and did some reorganization of signed/unsigned int
2585        variables following suggestions from Dair Grant. Also renamed the variable
2586        "this" as "item" because it is a C++ keyword.
2587    
2588    45. Arranged for dftables to add
2589    
2590          #include "pcre_internal.h"
2591    
2592        to pcre_chartables.c because without it, gcc 4.x may remove the array
2593        definition from the final binary if PCRE is built into a static library and
2594        dead code stripping is activated.
2595    
2596    46. For an unanchored pattern, if a match attempt fails at the start of a
2597        newline sequence, and the newline setting is CRLF or ANY, and the next two
2598        characters are CRLF, advance by two characters instead of one.
2599    
2600    
2601    Version 6.7 04-Jul-06
2602    ---------------------
2603    
2604     1. In order to handle tests when input lines are enormously long, pcretest has
2605        been re-factored so that it automatically extends its buffers when
2606        necessary. The code is crude, but this _is_ just a test program. The
2607        default size has been increased from 32K to 50K.
2608    
2609     2. The code in pcre_study() was using the value of the re argument before
2610        testing it for NULL. (Of course, in any sensible call of the function, it
2611        won't be NULL.)
2612    
2613     3. The memmove() emulation function in pcre_internal.h, which is used on
2614        systems that lack both memmove() and bcopy() - that is, hardly ever -
2615        was missing a "static" storage class specifier.
2616    
2617     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2618        containing an extended class (one that cannot be represented by a bitmap
2619        because it contains high-valued characters or Unicode property items, e.g.
2620        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2621        pattern, but PCRE should not loop if you do not (it no longer does).
2622        [Detail: two cases were found: (a) a repeated subpattern containing an
2623        extended class; (b) a recursive reference to a subpattern that followed a
2624        previous extended class. It wasn't skipping over the extended class
2625        correctly when UTF-8 mode was not set.]
2626    
2627     5. A negated single-character class was not being recognized as fixed-length
2628        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2629        compile error "lookbehind assertion is not fixed length".
2630    
2631     6. The RunPerlTest auxiliary script was showing an unexpected difference
2632        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2633        write a Perl script that can interpret lines of an input file either as
2634        byte characters or as UTF-8, which is what "perltest" was being required to
2635        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2636        can't do is switch easily at run time between having the "use utf8;" pragma
2637        or not. In the end, I fudged it by using the RunPerlTest script to insert
2638        "use utf8;" explicitly for the UTF-8 tests.
2639    
2640     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2641        the end of the subject string, contrary to the documentation and to what
2642        Perl does. This was true of both matching functions. Now it matches only at
2643        the start of the subject and immediately after *internal* newlines.
2644    
2645     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2646        a pointer to an int instead of a pointer to an unsigned long int. This
2647        caused problems on 64-bit systems.
2648    
2649     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2650        instance of the 'standard' template library not being so standard".
2651    
2652    10. There was no check on the number of named subpatterns nor the maximum
2653        length of a subpattern name. The product of these values is used to compute
2654        the size of the memory block for a compiled pattern. By supplying a very
2655        long subpattern name and a large number of named subpatterns, the size
2656        computation could be caused to overflow. This is now prevented by limiting
2657        the length of names to 32 characters, and the number of named subpatterns
2658        to 10,000.
2659    
2660    11. Subpatterns that are repeated with specific counts have to be replicated in
2661        the compiled pattern. The size of memory for this was computed from the
2662        length of the subpattern and the repeat count. The latter is limited to
2663        65535, but there was no limit on the former, meaning that integer overflow
2664        could in principle occur. The compiled length of a repeated subpattern is
2665        now limited to 30,000 bytes in order to prevent this.
2666    
2667    12. Added the optional facility to have named substrings with the same name.
2668    
2669    13. Added the ability to use a named substring as a condition, using the
2670        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2671        are numbers (not recommended). Forward references are permitted.
2672    
2673    14. Added forward references in named backreferences (if you see what I mean).
2674    
2675    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2676        pattern could run off the end of the subject. For example, the pattern
2677        "(?s)(.{1,5})"8 did this with the subject "ab".
2678    
2679    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2680        PCRE_CASELESS was set when matching characters that were quantified with ?
2681        or *.
2682    
2683    17. A character class other than a single negated character that had a minimum
2684        but no maximum quantifier - for example [ab]{6,} - was not handled
2685        correctly by pce_dfa_exec(). It would match only one character.
2686    
2687    18. A valid (though odd) pattern that looked like a POSIX character
2688        class but used an invalid character after [ (for example [[,abc,]]) caused
2689        pcre_compile() to give the error "Failed: internal error: code overflow" or
2690        in some cases to crash with a glibc free() error. This could even happen if
2691        the pattern terminated after [[ but there just happened to be a sequence of
2692        letters, a binary zero, and a closing ] in the memory that followed.
2693    
2694    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2695        over the years. Originally (before any Unicode support), just the bottom 8
2696        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2697        output from "man perlunicode" includes this:
2698    
2699          The regular expression compiler produces polymorphic opcodes.  That
2700          is, the pattern adapts to the data and automatically switches to
2701          the Unicode character scheme when presented with Unicode data--or
2702          instead uses a traditional byte scheme when presented with byte
2703          data.
2704    
2705        Sadly, a wide octal escape does not cause a switch, and in a string with
2706        no other multibyte characters, these octal escapes are treated as before.
2707        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2708        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2709        Unicode string.
2710    
2711        I have not perpetrated such confusion in PCRE. Up till now, it took just
2712        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2713        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2714        translate to the appropriate multibyte character.
2715    
2716    29. Applied some refactoring to reduce the number of warnings from Microsoft
2717        and Borland compilers. This has included removing the fudge introduced
2718        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2719        a warning about an unused variable.
2720    
2721    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2722        characters since release 4.0, because Perl (from release 5.004) does not.
2723        [Or at least, is documented not to: some releases seem to be in conflict
2724        with the documentation.] However, when a pattern was studied with
2725        pcre_study() and all its branches started with \s, PCRE still included VT
2726        as a possible starting character. Of course, this did no harm; it just
2727        caused an unnecessary match attempt.
2728    
2729    22. Removed a now-redundant internal flag bit that recorded the fact that case
2730        dependency changed within the pattern. This was once needed for "required
2731        byte" processing, but is no longer used. This recovers a now-scarce options
2732        bit. Also moved the least significant internal flag bit to the most-
2733        significant bit of the word, which was not previously used (hangover from
2734        the days when it was an int rather than a uint) to free up another bit for
2735        the future.
2736    
2737    23. Added support for CRLF line endings as well as CR and LF. As well as the
2738        default being selectable at build time, it can now be changed at runtime
2739        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2740        specify that it is scanning data with non-default line endings.
2741    
2742    24. Changed the definition of CXXLINK to make it agree with the definition of
2743        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2744    
2745    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2746        recursions. This makes a big different to stack usage for some patterns.
2747    
2748    26. If a subpattern containing a named recursion or subroutine reference such
2749        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2750        the space required for the compiled pattern went wrong and gave too small a
2751        value. Depending on the environment, this could lead to "Failed: internal
2752        error: code overflow at offset 49" or "glibc detected double free or
2753        corruption" errors.
2754    
2755    27. Applied patches from Google (a) to support the new newline modes and (b) to
2756        advance over multibyte UTF-8 characters in GlobalReplace.
2757    
2758    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2759        difference for some implementation of PCRE in some Windows version.
2760    
2761    29. Added some extra testing facilities to pcretest:
2762    
2763        \q<number>   in a data line sets the "match limit" value
2764        \Q<number>   in a data line sets the "match recursion limt" value
2765        -S <number>  sets the stack size, where <number> is in megabytes
2766    
2767        The -S option isn't available for Windows.
2768    
2769    
2770    Version 6.6 06-Feb-06
2771    ---------------------
2772    
2773     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2774        in pcreposix.h. I have copied the definition from pcre.h.
2775    
2776     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2777        because pcre.h is no longer a built file.
2778    
2779     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2780        not normally included in the compiled code.
2781    
2782    
2783    Version 6.5 01-Feb-06
2784    ---------------------
2785    
2786     1. When using the partial match feature with pcre_dfa_exec(), it was not
2787        anchoring the second and subsequent partial matches at the new starting
2788        point. This could lead to incorrect results. For example, with the pattern
2789        /1234/, partially matching against "123" and then "a4" gave a match.
2790    
2791     2. Changes to pcregrep:
2792    
2793        (a) All non-match returns from pcre_exec() were being treated as failures
2794            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2795            error message is output. Some extra information is given for the
2796            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2797            probably the only errors that are likely to be caused by users (by
2798            specifying a regex that has nested indefinite repeats, for instance).
2799            If there are more than 20 of these errors, pcregrep is abandoned.
2800    
2801        (b) A binary zero was treated as data while matching, but terminated the
2802            output line if it was written out. This has been fixed: binary zeroes
2803            are now no different to any other data bytes.
2804    
2805        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2806            used to set a locale for matching. The --locale=xxxx long option has
2807            been added (no short equivalent) to specify a locale explicitly on the
2808            pcregrep command, overriding the environment variables.
2809    
2810        (d) When -B was used with -n, some line numbers in the output were one less
2811            than they should have been.
2812    
2813        (e) Added the -o (--only-matching) option.
2814    
2815        (f) If -A or -C was used with -c (count only), some lines of context were
2816            accidentally printed for the final match.
2817    
2818        (g) Added the -H (--with-filename) option.
2819    
2820        (h) The combination of options -rh failed to suppress file names for files
2821            that were found from directory arguments.
2822    
2823        (i) Added the -D (--devices) and -d (--directories) options.
2824    
2825        (j) Added the -F (--fixed-strings) option.
2826    
2827        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2828    
2829        (l) Added the --colo(u)r option.
2830    
2831        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2832            is not present by default.
2833    
2834     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2835        items such as (?R) or (?1), when the recursion could match a number of
2836        alternatives. If it matched one of the alternatives, but subsequently,
2837        outside the recursion, there was a failure, the code tried to back up into
2838        the recursion. However, because of the way PCRE is implemented, this is not
2839        possible, and the result was an incorrect result from the match.
2840    
2841        In order to prevent this happening, the specification of recursion has
2842        been changed so that all such subpatterns are automatically treated as
2843        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2844    
2845     4. I had overlooked the fact that, in some locales, there are characters for
2846        which isalpha() is true but neither isupper() nor islower() are true. In
2847        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2848        and ordfeminine) are like this. This affected the treatment of \w and \W
2849        when they appeared in character classes, but not when they appeared outside
2850        a character class. The bit map for "word" characters is now created
2851        separately from the results of isalnum() instead of just taking it from the
2852        upper, lower, and digit maps. (Plus the underscore character, of course.)
2853    
2854     5. The above bug also affected the handling of POSIX character classes such as
2855        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2856        permanent tables. Instead, the bit maps for such a class were previously
2857        created as the appropriate unions of the upper, lower, and digit bitmaps.
2858        Now they are created by subtraction from the [[:word:]] class, which has
2859        its own bitmap.
2860    
2861     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2862        It is created by subtracting the vertical space characters (\x09, \x0a,
2863        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2864        subtraction was done in the overall bitmap for a character class, meaning
2865        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2866        be recognized. This bug has been fixed.
2867    
2868     7. Patches from the folks at Google:
2869    
2870          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2871          real life, but is still worth protecting against".
2872    
2873          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2874          regular expressions".
2875    
2876          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2877          have it.
2878    
2879          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2880          "configure" and the latter not, in order to fix a problem somebody had
2881          with compiling the Arg class on HP-UX.
2882    
2883          (e) Improve the error-handling of the C++ wrapper a little bit.
2884    
2885          (f) New tests for checking recursion limiting.
2886    
2887     8. The pcre_memmove() function, which is used only if the environment does not
2888        have a standard memmove() function (and is therefore rarely compiled),
2889        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2890        returning a result (though PCRE never actually uses the result).
2891    
2892     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2893        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2894        returned instead of calling malloc() with an overflowing number that would
2895        most likely cause subsequent chaos.
2896    
2897    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2898    
2899    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2900        with this option is matched, the nmatch and pmatch options of regexec() are
2901        ignored.
2902    
2903    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2904        provided in case anyone wants to the the POSIX interface with UTF-8
2905        strings.
2906    
2907    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2908        C++ linking (needed for some HP-UX environments).
2909    
2910    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2911        (unused parameter) and in the pcre_printint() function (omitted "default"
2912        switch label when the default is to do nothing).
2913    
2914    15. Added some code to make it possible, when PCRE is compiled as a C++
2915        library, to replace subject pointers for pcre_exec() with a smart pointer
2916        class, thus making it possible to process discontinuous strings.
2917    
2918    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2919        much the same function. They were added by different people who were trying
2920        to make PCRE easy to compile on non-Unix systems. It has been suggested
2921        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2922        for compiling on Windows systems. I have therefore replaced it with
2923        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2924        defaults to "extern" for C or "extern C" for C++, which works fine on
2925        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2926        SCOPE with something explicit in config.h. In addition:
2927    
2928        (a) pcreposix.h still had just "extern" instead of either of these macros;
2929            I have replaced it with PCRE_DATA_SCOPE.
2930    
2931        (b) Functions such as _pcre_xclass(), which are internal to the library,
2932            but external in the C sense, all had PCRE_EXPORT in their definitions.
2933            This is apparently wrong for the Windows case, so I have removed it.
2934            (It makes no difference on Unix-like systems.)
2935    
2936    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2937        of recursive calls to match(). This is different to MATCH_LIMIT because
2938        that limits the total number of calls to match(), not all of which increase
2939        the depth of recursion. Limiting the recursion depth limits the amount of
2940        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2941        when PCRE is compiled, and changed at run time. A patch from Google adds
2942        this functionality to the C++ interface.
2943    
2944    18. Changes to the handling of Unicode character properties:
2945    
2946        (a) Updated the table to Unicode 4.1.0.
2947    
2948        (b) Recognize characters that are not in the table as "Cn" (undefined).
2949    
2950        (c) I revised the way the table is implemented to a much improved format
2951            which includes recognition of ranges. It now supports the ranges that
2952            are defined in UnicodeData.txt, and it also amalgamates other
2953            characters into ranges. This has reduced the number of entries in the
2954            table from around 16,000 to around 3,000, thus reducing its size
2955            considerably. I realized I did not need to use a tree structure after
2956            all - a binary chop search is just as efficient. Having reduced the
2957            number of entries, I extended their size from 6 bytes to 8 bytes to
2958            allow for more data.
2959    
2960        (d) Added support for Unicode script names via properties such as \p{Han}.
2961    
2962    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2963        matching that character.
2964    
2965    20. When matching a repeated Unicode property with a minimum greater than zero,
2966        (for example \pL{2,}), PCRE could look past the end of the subject if it
2967        reached it while seeking the minimum number of characters. This could
2968        happen only if some of the characters were more than one byte long, because
2969        there is a check for at least the minimum number of bytes.
2970    
2971    21. Refactored the implementation of \p and \P so as to be more general, to
2972        allow for more different types of property in future. This has changed the
2973        compiled form incompatibly. Anybody with saved compiled patterns that use
2974        \p or \P will have to recompile them.
2975    
2976    22. Added "Any" and "L&" to the supported property types.
2977    
2978    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2979        but give a compile time error if the value is greater than 0xff.
2980    
2981    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2982        accidentally not being installed or uninstalled.
2983    
2984    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2985        made were to insert the current release number. This seemed silly, because
2986        it made things harder for people building PCRE on systems that don't run
2987        "configure". I have turned pcre.h into a distributed file, no longer built
2988        by "configure", with the version identification directly included. There is
2989        no longer a pcre.h.in file.
2990    
2991        However, this change necessitated a change to the pcre-config script as
2992        well. It is built from pcre-config.in, and one of the substitutions was the
2993        release number. I have updated configure.ac so that ./configure now finds
2994        the release number by grepping pcre.h.
2995    
2996    26. Added the ability to run the tests under valgrind.
2997    
2998    
2999    Version 6.4 05-Sep-05
3000    ---------------------
3001    
3002     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3003        "--" to be printed when multiple files were scanned, even when none of the
3004        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3005        consider it to be a bug, and have restored the previous behaviour.
3006    
3007     2. A couple of code tidies to get rid of compiler warnings.
3008    
3009     3. The pcretest program used to cheat by referring to symbols in the library
3010        whose names begin with _pcre_. These are internal symbols that are not
3011        really supposed to be visible externally, and in some environments it is
3012        possible to suppress them. The cheating is now confined to including
3013        certain files from the library's source, which is a bit cleaner.
3014    
3015     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3016        file's purpose clearer.
3017    
3018     5. Reorganized pcre_ucp_findchar().
3019    
3020    
3021    Version 6.3 15-Aug-05
3022    ---------------------
3023    
3024     1. The file libpcre.pc.in did not have general read permission in the tarball.
3025    
3026     2. There were some problems when building without C++ support:
3027    
3028        (a) If C++ support was not built, "make install" and "make test" still
3029            tried to test it.
3030    
3031        (b) There were problems when the value of CXX was explicitly set. Some
3032            changes have been made to try to fix these, and ...
3033    
3034        (c) --disable-cpp can now be used to explicitly disable C++ support.
3035    
3036        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3037            backslash in a target when C++ was disabled. This confuses some
3038            versions of "make", apparently. Using an intermediate variable solves
3039            this. (Same for CPP_LOBJ.)
3040    
3041     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3042        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3043        necessary on certain architectures.
3044    
3045     4. Added a setting of -export-symbols-regex to the link command to remove
3046        those symbols that are exported in the C sense, but actually are local
3047        within the library, and not documented. Their names all begin with
3048        "_pcre_". This is not a perfect job, because (a) we have to except some
3049        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3050        available (and never for static libraries). I have made a note to try to
3051        find a way round (a) in the future.
3052    
3053    
3054    Version 6.2 01-Aug-05
3055    ---------------------
3056    
3057     1. There was no test for integer overflow of quantifier values. A construction
3058        such as {1111111111111111} would give undefined results. What is worse, if
3059        a minimum quantifier for a parenthesized subpattern overflowed and became
3060        negative, the calculation of the memory size went wrong. This could have
3061        led to memory overwriting.
3062    
3063     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3064    
3065     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3066        operating environments where this matters.
3067    
3068     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3069        PCRE options from within the C++ wrapper.
3070    
3071     5. Named capturing subpatterns were not being correctly counted when a pattern
3072        was compiled. This caused two problems: (a) If there were more than 100
3073        such subpatterns, the calculation of the memory needed for the whole
3074        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3075        back references of the form \12, where the number was greater than 9, were
3076        not recognized as back references, even though there were sufficient
3077        previous subpatterns.
3078    
3079     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3080        versions of gcc, e.g. 2.95.4.
3081    
3082    
3083    Version 6.1 21-Jun-05
3084    ---------------------
3085    
3086     1. There was one reference to the variable "posix" in pcretest.c that was not
3087        surrounded by "#if !defined NOPOSIX".
3088    
3089     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3090        the cross-check on the old pcre_info() function, for the benefit of the
3091        cut-down version of PCRE that is currently imported into Exim.
3092    
3093     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3094        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3095        patterns that start (?i) but I don't think that matters. The use of (?i) is
3096        just an example; this all applies to the other options as well.
3097    
3098     4. Since libtool seems to echo the compile commands it is issuing, the output
3099        from "make" can be reduced a bit by putting "@" in front of each libtool
3100        compile command.
3101    
3102     5. Patch from the folks at Google for configure.in to be a bit more thorough
3103        in checking for a suitable C++ installation before trying to compile the
3104        C++ stuff. This should fix a reported problem when a compiler was present,
3105        but no suitable headers.
3106    
3107     6. The man pages all had just "PCRE" as their title. I have changed them to
3108        be the relevant file name. I have also arranged that these names are
3109        retained in the file doc/pcre.txt, which is a concatenation in text format
3110        of all the man pages except the little individual ones for each function.
3111    
3112     7. The NON-UNIX-USE file had not been updated for the different set of source
3113        files that come with release 6. I also added a few comments about the C++
3114        wrapper.
3115    
3116    
3117    Version 6.0 07-Jun-05
3118    ---------------------
3119    
3120     1. Some minor internal re-organization to help with my DFA experiments.
3121    
3122     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3123        didn't matter for the library itself when fully configured, but did matter
3124        when compiling without UCP support, or within Exim, where the ucp files are
3125        not imported.
3126    
3127     3. Refactoring of the library code to split up the various functions into
3128        different source modules. The addition of the new DFA matching code (see
3129        below) to a single monolithic source would have made it really too
3130        unwieldy, quite apart from causing all the code to be include in a
3131        statically linked application, when only some functions are used. This is
3132        relevant even without the DFA addition now that patterns can be compiled in
3133        one application and matched in another.
3134    
3135        The downside of splitting up is that there have to be some external
3136        functions and data tables that are used internally in different modules of
3137        the library but which are not part of the API. These have all had their
3138        names changed to start with "_pcre_" so that they are unlikely to clash
3139        with other external names.
3140    
3141     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3142        a different (DFA) algorithm. Although it is slower than the original
3143        function, it does have some advantages for certain types of matching
3144        problem.
3145    
3146     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3147        including restarting after a partial match.
3148    
3149     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3150        defined when compiling for Windows was sent to me. I have put it into the
3151        code, though I have no means of testing or verifying it.
3152    
3153     7. Added the pcre_refcount() auxiliary function.
3154    
3155     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3156        match before or at the first newline in the subject string. In pcretest,
3157        the /f option on a pattern can be used to set this.
3158    
3159     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3160        would behave wrongly. This has been present in PCRE since release 4.0.
3161    
3162    10. A number of changes to the pcregrep command:
3163    
3164        (a) Refactored how -x works; insert ^(...)$ instead of setting
3165            PCRE_ANCHORED and checking the length, in preparation for adding
3166            something similar for -w.
3167    
3168        (b) Added the -w (match as a word) option.
3169    
3170        (c) Refactored the way lines are read and buffered so as to have more
3171            than one at a time available.
3172    
3173        (d) Implemented a pcregrep test script.
3174    
3175        (e) Added the -M (multiline match) option. This allows patterns to match
3176            over several lines of the subject. The buffering ensures that at least
3177            8K, or the rest of the document (whichever is the shorter) is available
3178            for matching (and similarly the previous 8K for lookbehind assertions).
3179    
3180        (f) Changed the --help output so that it now says
3181    
3182              -w, --word-regex(p)
3183    
3184            instead of two lines, one with "regex" and the other with "regexp"
3185            because that confused at least one person since the short forms are the
3186            same. (This required a bit of code, as the output is generated
3187            automatically from a table. It wasn't just a text change.)
3188    
3189        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3190            option but starts with a hyphen. Could be a pattern or a path name
3191            starting with a hyphen, for instance.
3192    
3193        (h) "-" can be given as a file name to represent stdin.
3194    
3195        (i) When file names are being printed, "(standard input)" is used for
3196            the standard input, for compatibility with GNU grep. Previously
3197            "<stdin>" was used.
3198    
3199        (j) The option --label=xxx can be used to supply a name to be used for
3200            stdin when file names are being printed. There is no short form.
3201    
3202        (k) Re-factored the options decoding logic because we are going to add
3203            two more options that take data. Such options can now be given in four
3204            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3205    
3206        (l) Added the -A, -B, and -C options for requesting that lines of context
3207            around matches be printed.
3208    
3209        (m) Added the -L option to print the names of files that do not contain
3210            any matching lines, that is, the complement of -l.
3211    
3212        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3213            continue to scan other files.
3214    
3215        (o) The -s option was incorrectly implemented. For compatibility with other
3216            greps, it now suppresses the error message for a non-existent or non-
3217            accessible file (but not the return code). There is a new option called
3218            -q that suppresses the output of matching lines, which was what -s was
3219            previously doing.
3220    
3221        (p) Added --include and --exclude options to specify files for inclusion
3222            and exclusion when recursing.
3223    
3224    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3225        Hopefully, it now does.
3226    
3227    12. Missing cast in pcre_study().
3228    
3229    13. Added an "uninstall" target to the makefile.
3230    
3231    14. Replaced "extern" in the function prototypes in Makefile.in with
3232        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3233        world, but is set differently for Windows.
3234    
3235    15. Added a second compiling function called pcre_compile2(). The only
3236        difference is that it has an extra argument, which is a pointer to an
3237        integer error code. When there is a compile-time failure, this is set
3238        non-zero, in addition to the error test pointer being set to point to an
3239        error message. The new argument may be NULL if no error number is required
3240        (but then you may as well call pcre_compile(), which is now just a
3241        wrapper). This facility is provided because some applications need a
3242        numeric error indication, but it has also enabled me to tidy up the way
3243        compile-time errors are handled in the POSIX wrapper.
3244    
3245    16. Added VPATH=.libs to the makefile; this should help when building with one
3246        prefix path and installing with another. (Or so I'm told by someone who
3247        knows more about this stuff than I do.)
3248    
3249    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3250        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3251        match everything, including newlines. This is not POSIX-compatible, but
3252        somebody wanted the feature. From pcretest it can be activated by using
3253        both the P and the s flags.
3254    
3255    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3256    
3257    19. libpcre.pc was being incorrectly installed as executable.
3258    
3259    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3260        it now also looks for '\r' so that it will work unmodified on Windows.
3261    
3262    21. Added Google's contributed C++ wrapper to the distribution.
3263    
3264    22. Added some untidy missing memory free() calls in pcretest, to keep
3265        Electric Fence happy when testing.
3266    
3267    
3268    
3269    Version 5.0 13-Sep-04
3270    ---------------------
3271    
3272     1. Internal change: literal characters are no longer packed up into items
3273        containing multiple characters in a single byte-string. Each character
3274        is now matched using a separate opcode. However, there may be more than one
3275        byte in the character in UTF-8 mode.
3276    
3277     2. The pcre_callout_block structure has two new fields: pattern_position and
3278        next_item_length. These contain the offset in the pattern to the next match
3279        item, and its length, respectively.
3280    
3281     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3282        insertion of callouts before each pattern item. Added the /C option to
3283        pcretest to make use of this.
3284    
3285     4. On the advice of a Windows user, the lines
3286    
3287          #if defined(_WIN32) || defined(WIN32)
3288          _setmode( _fileno( stdout ), 0x8000 );
3289          #endif  /* defined(_WIN32) || defined(WIN32) */
3290    
3291        have been added to the source of pcretest. This apparently does useful
3292        magic in relation to line terminators.
3293    
3294     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3295        for the benefit of those environments where the "b" makes a difference.
3296    
3297     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3298        to know about it. I have put a hack into configure.in that adds in code
3299        to set GCC=yes if CC=icc. This seems to end up at a point in the
3300        generated configure script that is early enough to affect the setting of
3301        compiler options, which is what is needed, but I have no means of testing
3302        whether it really works. (The user who reported this had patched the
3303        generated configure script, which of course I cannot do.)
3304    
3305        LATER: After change 22 below (new libtool files), the configure script
3306        seems to know about icc (and also ecc). Therefore, I have commented out
3307        this hack in configure.in.
3308    
3309     7. Added support for pkg-config (2 patches were sent in).
3310    
3311     8. Negated POSIX character classes that used a combination of internal tables
3312        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3313        [[:^ascii]]. Typically, they would match almost any characters. The other
3314        POSIX classes were not broken in this way.
3315    
3316     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3317        to find the match, as PCRE was deluded into thinking that the match had to
3318        start at the start point or following a newline. The same bug applied to
3319        patterns with negative forward assertions or any backward assertions
3320        preceding ".*" at the start, unless the pattern required a fixed first
3321        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3322    
3323    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3324        starting at the last subject character, bytes beyond the end of the subject
3325        string were read.
3326    
3327    11. Renamed the variable "class" as "classbits" to make life easier for C++
3328        users. (Previously there was a macro definition, but it apparently wasn't
3329        enough.)
3330    
3331    12. Added the new field "tables" to the extra data so that tables can be passed
3332        in at exec time, or the internal tables can be re-selected. This allows
3333        a compiled regex to be saved and re-used at a later time by a different
3334        program that might have everything at different addresses.
3335    
3336    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3337        -R library as well as a -L library.
3338    
3339    14. The debugging options of pcretest (-d on the command line or D on a
3340        pattern) showed incorrect output for anything following an extended class
3341        that contained multibyte characters and which was followed by a quantifier.
3342    
3343    15. Added optional support for general category Unicode character properties
3344        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3345        support. It adds about 90K to the size of the library. The meanings of the
3346        inbuilt class escapes such as \d and \s have NOT been changed.
3347    
3348    16. Updated pcredemo.c to include calls to free() to release the memory for the
3349        compiled pattern.
3350    
3351    17. The generated file chartables.c was being created in the source directory
3352        instead of in the building directory. This caused the build to fail if the
3353        source directory was different from the building directory, and was
3354        read-only.
3355    
3356    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3357        file. No doubt somebody will tell me if they don't make sense... Also added
3358        Dan Mooney's comments about building on OpenVMS.
3359    
3360    19. Added support for partial matching via the PCRE_PARTIAL option for
3361        pcre_exec() and the \P data escape in pcretest.
3362    
3363    20. Extended pcretest with 3 new pattern features:
3364    
3365        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3366              write the compiled pattern to the file whose name is "rest-of-line".
3367              This is a straight binary dump of the data, with the saved pointer to
3368              the character tables forced to be NULL. The study data, if any, is
3369              written too. After writing, pcretest reads a new pattern.
3370    
3371        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3372              compiled pattern from the given file. There must not be any
3373              occurrences of "<" in the file name (pretty unlikely); if there are,
3374              pcretest will instead treat the initial "<" as a pattern delimiter.
3375              After reading in the pattern, pcretest goes on to read data lines as
3376              usual.
3377    
3378        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3379              and 16-bit fields in a compiled pattern, to simulate a pattern that
3380              was compiled on a host of opposite endianness.
3381    
3382    21. The pcre-exec() function can now cope with patterns that were compiled on
3383        hosts of opposite endianness, with this restriction:
3384    
3385          As for any compiled expression that is saved and used later, the tables
3386          pointer field cannot be preserved; the extra_data field in the arguments
3387          to pcre_exec() should be used to pass in a tables address if a value
3388          other than the default internal tables were used at compile time.
3389    
3390    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3391        now diagnosed as an error. Previously, most of the time, a negative number
3392        would have been treated as zero, but if in addition "ovector" was passed as
3393        NULL, a crash could occur.
3394    
3395    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3396        new versions from the libtool 1.5 distribution (the last one is a copy of
3397        a file called libtool.m4). This seems to have fixed the need to patch
3398        "configure" to support Darwin 1.3 (which I used to do). However, I still
3399        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3400        workstation).
3401    
3402    24. Changed the PCRE licence to be the more standard "BSD" licence.
3403    
3404    
3405    Version 4.5 01-Dec-03
3406    ---------------------
3407    
3408     1. There has been some re-arrangement of the code for the match() function so
3409        that it can be compiled in a version that does not call itself recursively.
3410        Instead, it keeps those local variables that need separate instances for
3411        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3412        needs to "recurse". Keeping track of where control must go is done by means
3413        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3414        hide most of the details from the main code, and operates only if
3415        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3416        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3417        operating.
3418    
3419        To make it easier for callers to provide specially tailored get/free
3420        functions for this usage, two new functions, pcre_stack_malloc, and
3421        pcre_stack_free, are used. They are always called in strict stacking order,
3422        and the size of block requested is always the same.
3423    
3424        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3425        PCRE has been compiled to use the stack or the heap for recursion. The
3426        -C option of pcretest uses this to show which version is compiled.
3427    
3428        A new data escape \S, is added to pcretest; it causes the amounts of store
3429        obtained and freed by both kinds of malloc/free at match time to be added
3430        to the output.
3431    
3432     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3433        what's available on my current Linux desktop machine.
3434    
3435     3. When matching a UTF-8 string, the test for a valid string at the start has
3436        been extended. If start_offset is not zero, PCRE now checks that it points
3437        to a byte that is the start of a UTF-8 character. If not, it returns
3438        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3439        this is necessary because there may be backward assertions in the pattern.
3440        When matching the same subject several times, it may save resources to use
3441        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3442    
3443     4. The code for checking the validity of UTF-8 strings has been tightened so
3444        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3445        containing "overlong sequences".
3446    
3447     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3448        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3449        should have been "&", but it just so happened that all the cases this let
3450        through by mistake were picked up later in the function.
3451    
3452     6. I had used a variable called "isblank" - this is a C99 function, causing
3453        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3454    
3455     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3456        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3457        so the version is shown; (c) stop "make test" echoing "./RunTest".
3458    
3459     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3460    
3461     9. The prototype for memmove() for systems that don't have it was using
3462        size_t, but the inclusion of the header that defines size_t was later. I've
3463        moved the #includes for the C headers earlier to avoid this.
3464    
3465    10. Added some adjustments to the code to make it easier to compiler on certain
3466        special systems:
3467    
3468          (a) Some "const" qualifiers were missing.
3469          (b) Added the macro EXPORT before all exported functions; by default this
3470              is defined to be empty.
3471          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3472              that it reads its output file name as an argument instead of writing
3473              to the standard output and assuming this can be redirected.
3474    
3475    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3476        class containing characters with values greater than 255, PCRE compilation
3477        went into a loop.
3478    
3479    12. A recursive reference to a subpattern that was within another subpattern
3480        that had a minimum quantifier of zero caused PCRE to crash. For example,
3481        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3482        recursion. If the recursively-called subpattern itself had a zero repeat,
3483        that was OK.
3484    
3485    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3486        buffer into which it was copied (for escape processing) was still set at
3487        1024, so long lines caused crashes.
3488    
3489    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3490        "internal error: code overflow...". This applied to any character class
3491        that was followed by a possessive quantifier.
3492    
3493    15. Modified the Makefile to add libpcre.la as a prerequisite for
3494        libpcreposix.la because I was told this is needed for a parallel build to
3495        work.
3496    
3497    16. If a pattern that contained .* following optional items at the start was
3498        studied, the wrong optimizing data was generated, leading to matching
3499        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3500        matching string must start with a or b or c. The correct conclusion for
3501        this pattern is that a match can start with any character.
3502    
3503    
3504    Version 4.4 13-Aug-03
3505    ---------------------
3506    
3507     1. In UTF-8 mode, a character class containing characters with values between
3508        127 and 255 was not handled correctly if the compiled pattern was studied.
3509        In fixing this, I have also improved the studying algorithm for such
3510        classes (slightly).
3511    
3512     2. Three internal functions had redundant arguments passed to them. Removal
3513        might give a very teeny performance improvement.
3514    
3515     3. Documentation bug: the value of the capture_top field in a callout is *one
3516        more than* the number of the hightest numbered captured substring.
3517    
3518     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3519        in incorrectly linking with a previously installed version. They now link
3520        explicitly with libpcre.la.
3521    
3522     5. configure.in no longer needs to recognize Cygwin specially.
3523    
3524     6. A problem in pcre.in for Windows platforms is fixed.
3525    
3526     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3527        pcretest, it used to include the size of the study block as part of its
3528        output. Unfortunately, the structure contains a field that has a different
3529        size on different hardware architectures. This meant that the tests that
3530        showed this size failed. As the block is currently always of a fixed size,
3531        this information isn't actually particularly useful in pcretest output, so
3532        I have just removed it.
3533    
3534     8. Three pre-processor statements accidentally did not start in column 1.
3535        Sadly, there are *still* compilers around that complain, even though
3536        standard C has not required this for well over a decade. Sigh.
3537    
3538     9. In pcretest, the code for checking callouts passed small integers in the
3539        callout_data field, which is a void * field. However, some picky compilers
3540        complained about the casts involved for this on 64-bit systems. Now
3541        pcretest passes the address of the small integer instead, which should get
3542        rid of the warnings.
3543    
3544    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3545        both compile and run time, and gives an error if an invalid UTF-8 sequence
3546        is found. There is a option for disabling this check in cases where the
3547        string is known to be correct and/or the maximum performance is wanted.
3548    
3549    11. In response to a bug report, I changed one line in Makefile.in from
3550    
3551            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3552        to
3553            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3554    
3555        to look similar to other lines, but I have no way of telling whether this
3556        is the right thing to do, as I do not use Windows. No doubt I'll get told
3557        if it's wrong...
3558    
3559    
3560    Version 4.3 21-May-03
3561    ---------------------
3562    
3563    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3564       Makefile.
3565    
3566    2. Some refactoring to improve the quality of the code:
3567    
3568       (i)   The utf8_table... variables are now declared "const".
3569    
3570       (ii)  The code for \cx, which used the "case flipping" table to upper case
3571             lower case letters, now just substracts 32. This is ASCII-specific,
3572             but the whole concept of \cx is ASCII-specific, so it seems
3573             reasonable.
3574    
3575       (iii) PCRE was using its character types table to recognize decimal and
3576             hexadecimal digits in the pattern. This is silly, because it handles
3577             only 0-9, a-f, and A-F, but the character types table is locale-
3578             specific, which means strange things might happen. A private
3579             table is now used for this - though it costs 256 bytes, a table is
3580             much faster than multiple explicit tests. Of course, the standard
3581             character types table is still used for matching digits in subject
3582             strings against \d.
3583    
3584       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3585             ending in _t are). So I've renamed it as ESC_tee.
3586    
3587    3. The first argument for regexec() in the POSIX wrapper should have been
3588       defined as "const".
3589    
3590    4. Changed pcretest to use malloc() for its buffers so that they can be
3591       Electric Fenced for debugging.
3592    
3593    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3594       to read one or more bytes before the start of the subject string. Often this
3595       had no effect on PCRE's behaviour, but in some circumstances it could
3596       provoke a segmentation fault.
3597    
3598    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3599       to try to read one or more bytes before the start of the subject string.
3600    
3601    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3602       UTF-8 support could misbehave in various ways if the subject string
3603       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3604       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3605       back over UTF-8 characters.)
3606    
3607    
3608    Version 4.2 14-Apr-03
3609    ---------------------
3610    
3611    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3612    
3613    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3614         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3615         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3616         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3617         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3618           and BUILD_EXEEXT
3619         Note: automatic setting of the BUILD variables is not yet working
3620         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3621           compile-time but not at link-time
3622         [LINK]: use for linking executables only
3623         make different versions for Windows and non-Windows
3624         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3625           libraries
3626         [LINK_FOR_BUILD]: new variable
3627         [OBJEXT]: use throughout
3628         [EXEEXT]: use throughout
3629         <winshared>: new target
3630         <wininstall>: new target
3631         <dftables.o>: use native compiler
3632         <dftables>: use native linker
3633         <install>: handle Windows platform correctly
3634         <clean>: ditto
3635         <check>: ditto
3636         copy DLL to top builddir before testing
3637    
3638       As part of these changes, -no-undefined was removed again. This was reported
3639       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3640       in any case.
3641    
3642    3. Some tidies to get rid of compiler warnings:
3643    
3644       . In the match_data structure, match_limit was an unsigned long int, whereas
3645         match_call_count was an int. I've made them both unsigned long ints.
3646    
3647       . In pcretest the fact that a const uschar * doesn't automatically cast to
3648         a void * provoked a warning.
3649    
3650       . Turning on some more compiler warnings threw up some "shadow" variables
3651         and a few more missing casts.
3652    
3653    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3654       option, a class that contained a single character with a value between 128
3655       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3656    
3657    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3658       option, a class that contained several characters, but with at least one
3659       whose value was between 128 and 255 caused PCRE to crash.
3660    
3661    
3662    Version 4.1 12-Mar-03
3663    ---------------------
3664    
3665    1. Compiling with gcc -pedantic found a couple of places where casts were
3666    needed, and a string in dftables.c that was longer than standard compilers are
3667    required to support.
3668    
3669    2. Compiling with Sun's compiler found a few more places where the code could
3670    be tidied up in order to avoid warnings.
3671    
3672    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3673    first of these names is deprecated in the latest Autoconf in favour of the name
3674    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3675    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3676    analogy I have changed it to CFLAGS_FOR_BUILD.
3677    
3678    4. Added -no-undefined to the linking command in the Makefile, because this is
3679    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3680    linking step for the pcreposix library.
3681    
3682    5. PCRE was failing to diagnose the case of two named groups with the same
3683    name.
3684    
3685    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3686    literal character that is needed in the subject for a match, and scans along to
3687    ensure that it is present before embarking on the full matching process. This
3688    saves time in cases of nested unlimited repeats that are never going to match.
3689    Problem: the scan can take a lot of time if the subject is very long (e.g.
3690    megabytes), thus penalizing straightforward matches. It is now done only if the
3691    amount of subject to be scanned is less than 1000 bytes.
3692    
3693    7. A lesser problem with the same optimization is that it was recording the
3694    first character of an anchored pattern as "needed", thus provoking a search
3695    right along the subject, even when the first match of the pattern was going to
3696    fail. The "needed" character is now not set for anchored patterns, unless it
3697    follows something in the pattern that is of non-fixed length. Thus, it still
3698    fulfils its original purpose of finding quick non-matches in cases of nested
3699    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3700    
3701    
3702    Version 4.0 17-Feb-03
3703    ---------------------
3704    
3705    1. If a comment in an extended regex that started immediately after a meta-item
3706    extended to the end of string, PCRE compiled incorrect data. This could lead to
3707    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3708    
3709    2. Moved to autoconf 2.53 and libtool 1.4.2.
3710    
3711    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3712    the special perltest8 script is no longer needed - all the tests can be run
3713    from a single perltest script.
3714    
3715    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3716    by \s. It has now been removed in PCRE. This means it isn't recognized as
3717    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3718    class [:space:] *does* include VT, thereby creating a mess.
3719    
3720    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3721    space and tab.
3722    
3723    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3724    its new features into the main test script, reducing the number of scripts.
3725    
3726    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3727    were backward compatible, and made the (?i) apply to the whole pattern, as if
3728    /i were given. Now it behaves more logically, and applies the option setting
3729    only to what follows. PCRE has been changed to follow suit. However, if it
3730    finds options settings right at the start of the pattern, it extracts them into
3731    the global options, as before. Thus, they show up in the info data.
3732    
3733    8. Added support for the \Q...\E escape sequence. Characters in between are
3734    treated as literals. This is slightly different from Perl in that $ and @ are
3735    also handled as literals inside the quotes. In Perl, they will cause variable
3736    interpolation. Note the following examples:
3737    
3738        Pattern            PCRE matches      Perl matches
3739    
3740        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3741        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3742        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3743    
3744    For compatibility with Perl, \Q...\E sequences are recognized inside character
3745    classes as well as outside them.
3746    
3747    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3748    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3749    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3750    signed/unsigned warnings.
3751    
3752    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3753    option for pcretest, so I've replaced it by a simple function that does just
3754    that job.
3755    
3756    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3757    "pcregrep -".
3758    
3759    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3760    Java package. This provides some syntactic sugar for simple cases of what my
3761    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3762    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3763    item, you can use this simplified notation. Note that only makes sense with
3764    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3765    greediness, whatever the setting of the PCRE_UNGREEDY option.
3766    
3767    13. A change of greediness default within a pattern was not taking effect at
3768    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3769    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3770    was abstracted outside.
3771    
3772    14. PCRE now supports the \G assertion. It is true when the current matching
3773    position is at the start point of the match. This differs from \A when the
3774    starting offset is non-zero. Used with the /g option of pcretest (or similar
3775    code), it works in the same way as it does for Perl's /g option. If all
3776    alternatives of a regex begin with \G, the expression is anchored to the start
3777    match position, and the "anchored" flag is set in the compiled expression.
3778    
3779    15. Some bugs concerning the handling of certain option changes within patterns
3780    have been fixed. These applied to options other than (?ims). For example,
3781    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3782    been the other way round. Some of this was related to change 7 above.
3783    
3784    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3785    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3786    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3787    POSIX classes only within a class (e.g. /[[:alpha:]]/).