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

Diff of /code/trunk/ChangeLog

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

<
revision 29 by nigel, Sat Feb 24 21:38:53 2007 UTC revision 1041 by ph10, Sun Sep 16 10:16:27 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    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62        structure could go wrong in environments where size_t is not the same size
63        as int.
64    
65    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
66        set.
67    
68    15. The EBCDIC support had decayed; later updates to the code had included
69        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
70        general tidy up of EBCDIC-related issues, and the documentation was also
71        not quite right. There is now a test that can be run on ASCII systems to
72        check some of the EBCDIC-related things (but is it not a full test).
73    
74    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
75        in a small tidy to the code.
76    
77    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
78    
79    18. If the --only-matching (-o) option in pcregrep is specified multiple
80        times, each one causes appropriate output. For example, -o1 -o2 outputs the
81        substrings matched by the 1st and 2nd capturing parentheses. A separating
82        string can be specified by --om-separator (default empty).
83    
84    19. Improving the first n character searches.
85    
86    20. Turn case lists for horizontal and vertical white space into macros so that
87        they are defined only once.
88    
89    
90    Version 8.31 06-July-2012
91    -------------------------
92    
93    1.  Fixing a wrong JIT test case and some compiler warnings.
94    
95    2.  Removed a bashism from the RunTest script.
96    
97    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
98        to unsigned type, result still unsigned" that was given by an MS compiler
99        on encountering the code "-sizeof(xxx)".
100    
101    4.  Partial matching support is added to the JIT compiler.
102    
103    5.  Fixed several bugs concerned with partial matching of items that consist
104        of more than one character:
105    
106        (a) /^(..)\1/ did not partially match "aba" because checking references was
107            done on an "all or nothing" basis. This also applied to repeated
108            references.
109    
110        (b) \R did not give a hard partial match if \r was found at the end of the
111            subject.
112    
113        (c) \X did not give a hard partial match after matching one or more
114            characters at the end of the subject.
115    
116        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
117            a partial match for the string "\r".
118    
119        (e) When newline was set to CRLF, the metacharacter "." did not recognize
120            a partial match for a CR character at the end of the subject string.
121    
122    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
123        running pcretest, the text "(JIT)" added to the output whenever JIT is
124        actually used to run the match.
125    
126    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
127        or /S+[+] with a digit between 1 and 7.
128    
129    8.  OP_NOT now supports any UTF character not just single-byte ones.
130    
131    9.  (*MARK) control verb is now supported by the JIT compiler.
132    
133    10. The command "./RunTest list" lists the available tests without actually
134        running any of them. (Because I keep forgetting what they all are.)
135    
136    11. Add PCRE_INFO_MAXLOOKBEHIND.
137    
138    12. Applied a (slightly modified) user-supplied patch that improves performance
139        when the heap is used for recursion (compiled with --disable-stack-for-
140        recursion). Instead of malloc and free for each heap frame each time a
141        logical recursion happens, frames are retained on a chain and re-used where
142        possible. This sometimes gives as much as 30% improvement.
143    
144    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
145        call.
146    
147    14. As documented, (*COMMIT) is now confined to within a positive assertion.
148    
149    15. It is now possible to link pcretest with libedit as an alternative to
150        libreadline.
151    
152    16. (*COMMIT) control verb is now supported by the JIT compiler.
153    
154    17. The Unicode data tables have been updated to Unicode 6.1.0.
155    
156    18. Added --file-list option to pcregrep.
157    
158    19. Added binary file support to pcregrep, including the -a, --binary-files,
159        -I, and --text options.
160    
161    20. The madvise function is renamed for posix_madvise for QNX compatibility
162        reasons. Fixed by Giuseppe D'Angelo.
163    
164    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
165    
166    22. Changed the option for creating symbolic links for 16-bit man pages from
167        -s to -sf so that re-installing does not cause issues.
168    
169    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
170    
171    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
172        matches in certain environments (the workspace was not being correctly
173        retained). Also added to pcre_dfa_exec() a simple plausibility check on
174        some of the workspace data at the beginning of a restart.
175    
176    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
177        was not doing so when it should - probably a typo introduced by SVN 528
178        (change 8.10/14).
179    
180    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
181        \w+ when the character tables indicated that \x{c4} was a word character.
182        There were several related cases, all because the tests for doing a table
183        lookup were testing for characters less than 127 instead of 255.
184    
185    27. If a pattern contains capturing parentheses that are not used in a match,
186        their slots in the ovector are set to -1. For those that are higher than
187        any matched groups, this happens at the end of processing. In the case when
188        there were back references that the ovector was too small to contain
189        (causing temporary malloc'd memory to be used during matching), and the
190        highest capturing number was not used, memory off the end of the ovector
191        was incorrectly being set to -1. (It was using the size of the temporary
192        memory instead of the true size.)
193    
194    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
195        ovector size, it uses memory at the end of the block that it has got.
196    
197    29. Check for an overlong MARK name and give an error at compile time. The
198        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
199    
200    30. JIT compiler update.
201    
202    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
203        Rill for the patch.
204    
205    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
206    
207    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
208    
209    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
210        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
211        was enabled.
212    
213    35. Improve JIT code generation for greedy plus quantifier.
214    
215    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
216        1 to "aa" instead of to an empty string. The bug affected repeated groups
217        that could potentially match an empty string.
218    
219    37. Optimizing single character iterators in JIT.
220    
221    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
222        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
223        codepoints that are too big for the mode are faulted, and in a UTF mode,
224        disallowed codepoints are also faulted.
225    
226    39. If PCRE was compiled with UTF support, in three places in the DFA
227        matcher there was code that should only have been obeyed in UTF mode, but
228        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
229        processing when bytes with values greater than 127 were present. In 16-bit
230        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
231        both cases the values are those that cannot be the first data item in a UTF
232        character. The three items that might have provoked this were recursions,
233        possessively repeated groups, and atomic groups.
234    
235    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
236        and pcregrep, because some OS require shared objects to be explicitly
237        passed to ld, causing the link step to fail if they are not.
238    
239    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
240        mode, patterns that started with \h* or \R* might be incorrectly matched.
241    
242    
243    Version 8.30 04-February-2012
244    -----------------------------
245    
246    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
247        name is defined in ctype.h.
248    
249    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
250        only in quite long subpatterns.
251    
252    3.  Removed the function pcre_info(), which has been obsolete and deprecated
253        since it was replaced by pcre_fullinfo() in February 2000.
254    
255    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
256        match a (*MARK), and the match failed at the start of the subject, a
257        reference to memory before the start of the subject could occur. This bug
258        was introduced by fix 17 of release 8.21.
259    
260    5.  A reference to an unset group with zero minimum repetition was giving
261        totally wrong answers (in non-JavaScript-compatibility mode). For example,
262        /(another)?(\1?)test/ matched against "hello world test". This bug was
263        introduced in release 8.13.
264    
265    6.  Add support for 16-bit character strings (a large amount of work involving
266        many changes and refactorings).
267    
268    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
269        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
270        from a file.
271    
272    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
273        rounding is not applied in this particular case).
274    
275    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
276        if they appear, or are escaped, in patterns.
277    
278    10. Get rid of a number of -Wunused-but-set-variable warnings.
279    
280    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
281        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
282        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
283        also returns the mark "x". This bug applied to capturing parentheses,
284        non-capturing parentheses, and atomic parentheses. It also applied to some
285        assertions.
286    
287    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
288        information out of configure.ac instead of relying on pcre.h.generic, which
289        is not stored in the repository.
290    
291    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
292        -lreadline.
293    
294    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
295    
296    15. Applied Graycode's patch to put the top-level frame on the stack rather
297        than the heap when not using the stack for recursion. This gives a
298        performance improvement in many cases when recursion is not deep.
299    
300    16. Experimental code added to "pcretest -C" to output the stack frame size.
301    
302    
303    Version 8.21 12-Dec-2011
304    ------------------------
305    
306    1.  Updating the JIT compiler.
307    
308    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
309        are added as well.
310    
311    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
312        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
313        calling _pcre_jit_exec. Some extra comments are added.
314    
315    4.  (*MARK) settings inside atomic groups that do not contain any capturing
316        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
317        was introduced by change 18 for 8.20.
318    
319    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
320        ECMA-262 standard.
321    
322    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
323        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
324        This bug was probably introduced by change 9 of 8.13.
325    
326    7.  While fixing 6 above, I noticed that a number of other items were being
327        incorrectly rejected as "not fixed length". This arose partly because newer
328        opcodes had not been added to the fixed-length checking code. I have (a)
329        corrected the bug and added tests for these items, and (b) arranged for an
330        error to occur if an unknown opcode is encountered while checking for fixed
331        length instead of just assuming "not fixed length". The items that were
332        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
333        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
334        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
335    
336    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
337        being incorrectly compiled and would have given unpredicatble results.
338    
339    9.  A possessively repeated subpattern with minimum repeat count greater than
340        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
341        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
342        the first (A) could occur when it should not.
343    
344    10. Add a cast and remove a redundant test from the code.
345    
346    11. JIT should use pcre_malloc/pcre_free for allocation.
347    
348    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
349        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
350        is anything other than /usr, -L is still shown).
351    
352    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
353    
354    14. Perl does not support \N without a following name in a [] class; PCRE now
355        also gives an error.
356    
357    15. If a forward reference was repeated with an upper limit of around 2000,
358        it caused the error "internal error: overran compiling workspace". The
359        maximum number of forward references (including repeats) was limited by the
360        internal workspace, and dependent on the LINK_SIZE. The code has been
361        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
362        the default depends on LINK_SIZE. There is a new upper limit (for safety)
363        of around 200,000 forward references. While doing this, I also speeded up
364        the filling in of repeated forward references.
365    
366    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
367        incorrectly expecting the subject to contain another "a" after the start.
368    
369    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
370        in the match, the SKIP should be ignored. This was not happening; instead
371        the SKIP was being treated as NOMATCH. For patterns such as
372        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
373        tested.
374    
375    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
376        now much more compatible with Perl, in particular in cases where the result
377        is a non-match for a non-anchored pattern. For example, if
378        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
379        "m", where previously it did not return a name. A side effect of this
380        change is that for partial matches, the last encountered mark name is
381        returned, as for non matches. A number of tests that were previously not
382        Perl-compatible have been moved into the Perl-compatible test files. The
383        refactoring has had the pleasing side effect of removing one argument from
384        the match() function, thus reducing its stack requirements.
385    
386    19. If the /S+ option was used in pcretest to study a pattern using JIT,
387        subsequent uses of /S (without +) incorrectly behaved like /S+.
388    
389    21. Retrieve executable code size support for the JIT compiler and fixing
390        some warnings.
391    
392    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
393        not work when the shorter character appeared right at the end of the
394        subject string.
395    
396    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
397        systems.
398    
399    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
400        output it when the /M option is used in pcretest.
401    
402    25. The CheckMan script was not being included in the distribution. Also, added
403        an explicit "perl" to run Perl scripts from the PrepareRelease script
404        because this is reportedly needed in Windows.
405    
406    26. If study data was being save in a file and studying had not found a set of
407        "starts with" bytes for the pattern, the data written to the file (though
408        never used) was taken from uninitialized memory and so caused valgrind to
409        complain.
410    
411    27. Updated RunTest.bat as provided by Sheri Pierce.
412    
413    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
414    
415    29. Computation of memory usage for the table of capturing group names was
416        giving an unnecessarily large value.
417    
418    
419    Version 8.20 21-Oct-2011
420    ------------------------
421    
422    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
423        a POSIX class. After further experiments with Perl, which convinced me that
424        Perl has bugs and confusions, a closing square bracket is no longer allowed
425        in a POSIX name. This bug also affected patterns with classes that started
426        with full stops.
427    
428    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
429        captured substring, but while checking the failing first alternative,
430        substring 1 is temporarily captured. If the output vector supplied to
431        pcre_exec() was not big enough for this capture, the yield of the function
432        was still zero ("insufficient space for captured substrings"). This cannot
433        be totally fixed without adding another stack variable, which seems a lot
434        of expense for a edge case. However, I have improved the situation in cases
435        such as /(a)(b)x|abc/ matched against "abc", where the return code
436        indicates that fewer than the maximum number of slots in the ovector have
437        been set.
438    
439    3.  Related to (2) above: when there are more back references in a pattern than
440        slots in the output vector, pcre_exec() uses temporary memory during
441        matching, and copies in the captures as far as possible afterwards. It was
442        using the entire output vector, but this conflicts with the specification
443        that only 2/3 is used for passing back captured substrings. Now it uses
444        only the first 2/3, for compatibility. This is, of course, another edge
445        case.
446    
447    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
448        main code base, and can be used by building with --enable-jit. When this is
449        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
450        runtime --no-jit option is given.
451    
452    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
453        ovector, the return from the function was zero, implying that there were
454        other matches that did not fit. The correct "exactly full" value is now
455        returned.
456    
457    6.  If a subpattern that was called recursively or as a subroutine contained
458        (*PRUNE) or any other control that caused it to give a non-standard return,
459        invalid errors such as "Error -26 (nested recursion at the same subject
460        position)" or even infinite loops could occur.
461    
462    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
463        computing the minimum length on reaching *ACCEPT, and so ended up with the
464        wrong value of 1 rather than 0. Further investigation indicates that
465        computing a minimum subject length in the presence of *ACCEPT is difficult
466        (think back references, subroutine calls), and so I have changed the code
467        so that no minimum is registered for a pattern that contains *ACCEPT.
468    
469    8.  If (*THEN) was present in the first (true) branch of a conditional group,
470        it was not handled as intended. [But see 16 below.]
471    
472    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
473        Sheri Pierce.
474    
475    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
476        the first byte in a match must be "a".
477    
478    11. Change 17 for 8.13 increased the recursion depth for patterns like
479        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
480        pattern contains any instances of (*THEN). If it does not, the old
481        optimizations are restored. It would be nice to do this on a per-group
482        basis, but at the moment that is not feasible.
483    
484    12. In some environments, the output of pcretest -C is CRLF terminated. This
485        broke RunTest's code that checks for the link size. A single white space
486        character after the value is now allowed for.
487    
488    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
489        For "fr", it uses the Windows-specific input and output files.
490    
491    14. If (*THEN) appeared in a group that was called recursively or as a
492        subroutine, it did not work as intended. [But see next item.]
493    
494    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
495        pattern fragments (but not containing any | characters). If A and B are
496        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
497        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
498        D. In other words, Perl considers parentheses that do not contain any |
499        characters to be part of a surrounding alternative, whereas PCRE was
500        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
501        differently. PCRE now behaves in the same way as Perl, except in the case
502        of subroutine/recursion calls such as (?1) which have in any case always
503        been different (but PCRE had them first :-).
504    
505    16. Related to 15 above: Perl does not treat the | in a conditional group as
506        creating alternatives. Such a group is treated in the same way as an
507        ordinary group without any | characters when processing (*THEN). PCRE has
508        been changed to match Perl's behaviour.
509    
510    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
511        RunGrepTest script failed.
512    
513    18. Change 22 for version 13 caused atomic groups to use more stack. This is
514        inevitable for groups that contain captures, but it can lead to a lot of
515        stack use in large patterns. The old behaviour has been restored for atomic
516        groups that do not contain any capturing parentheses.
517    
518    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
519        suppress the check for a minimum subject length at run time. (If it was
520        given to pcre_exec() or pcre_dfa_exec() it did work.)
521    
522    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
523        fail to work when decoding hex characters in data strings in EBCDIC
524        environments.
525    
526    21. It appears that in at least one Mac OS environment, the isxdigit() function
527        is implemented as a macro that evaluates to its argument more than once,
528        contravening the C 90 Standard (I haven't checked a later standard). There
529        was an instance in pcretest which caused it to go wrong when processing
530        \x{...} escapes in subject strings. The has been rewritten to avoid using
531        things like p++ in the argument of isxdigit().
532    
533    
534    Version 8.13 16-Aug-2011
535    ------------------------
536    
537    1.  The Unicode data tables have been updated to Unicode 6.0.0.
538    
539    2.  Two minor typos in pcre_internal.h have been fixed.
540    
541    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
542        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
543        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
544    
545    4.  There were a number of related bugs in the code for matching backrefences
546        caselessly in UTF-8 mode when codes for the characters concerned were
547        different numbers of bytes. For example, U+023A and U+2C65 are an upper
548        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
549        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
550        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
551        2-byte code at the end of the subject (it thought there wasn't enough data
552        left).
553    
554    5.  Comprehensive information about what went wrong is now returned by
555        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
556        as the output vector has at least 2 elements. The offset of the start of
557        the failing character and a reason code are placed in the vector.
558    
559    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
560        now returned is for the first byte of the failing character, instead of the
561        last byte inspected. This is an incompatible change, but I hope it is small
562        enough not to be a problem. It makes the returned offset consistent with
563        pcre_exec() and pcre_dfa_exec().
564    
565    7.  pcretest now gives a text phrase as well as the error number when
566        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
567        failure, the offset and reason code are output.
568    
569    8.  When \R was used with a maximizing quantifier it failed to skip backwards
570        over a \r\n pair if the subsequent match failed. Instead, it just skipped
571        back over a single character (\n). This seems wrong (because it treated the
572        two characters as a single entity when going forwards), conflicts with the
573        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
574        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
575        has been changed.
576    
577    9.  Some internal refactoring has changed the processing so that the handling
578        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
579        time (the PCRE_DOTALL option was changed this way some time ago: version
580        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
581        which was always a bit of a fudge. It also means that there is one less
582        argument for the match() function, which reduces its stack requirements
583        slightly. This change also fixes an incompatibility with Perl: the pattern
584        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
585    
586    10. More internal refactoring has drastically reduced the number of recursive
587        calls to match() for possessively repeated groups such as (abc)++ when
588        using pcre_exec().
589    
590    11. While implementing 10, a number of bugs in the handling of groups were
591        discovered and fixed:
592    
593        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
594        (a|)*(?1) gave a compile-time internal error.
595        ((a|)+)+  did not notice that the outer group could match an empty string.
596        (^a|^)+   was not marked as anchored.
597        (.*a|.*)+ was not marked as matching at start or after a newline.
598    
599    12. Yet more internal refactoring has removed another argument from the match()
600        function. Special calls to this function are now indicated by setting a
601        value in a variable in the "match data" data block.
602    
603    13. Be more explicit in pcre_study() instead of relying on "default" for
604        opcodes that mean there is no starting character; this means that when new
605        ones are added and accidentally left out of pcre_study(), testing should
606        pick them up.
607    
608    14. The -s option of pcretest has been documented for ages as being an old
609        synonym of -m (show memory usage). I have changed it to mean "force study
610        for every regex", that is, assume /S for every regex. This is similar to -i
611        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
612        using it. It makes it easier to run collections of tests with and without
613        study enabled, and thereby test pcre_study() more easily. All the standard
614        tests are now run with and without -s (but some patterns can be marked as
615        "never study" - see 20 below).
616    
617    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
618        restoration of the capturing data to the outer values was not happening
619        correctly.
620    
621    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
622        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
623        pattern had matched an empty string, and so incorrectly returned a no
624        match.
625    
626    17. There was optimizing code for the last branch of non-capturing parentheses,
627        and also for the obeyed branch of a conditional subexpression, which used
628        tail recursion to cut down on stack usage. Unfortunately, now that there is
629        the possibility of (*THEN) occurring in these branches, tail recursion is
630        no longer possible because the return has to be checked for (*THEN). These
631        two optimizations have therefore been removed. [But see 8.20/11 above.]
632    
633    18. If a pattern containing \R was studied, it was assumed that \R always
634        matched two bytes, thus causing the minimum subject length to be
635        incorrectly computed because \R can also match just one byte.
636    
637    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
638        was incorrectly computed.
639    
640    20. If /S is present twice on a test pattern in pcretest input, it now
641        *disables* studying, thereby overriding the use of -s on the command line
642        (see 14 above). This is necessary for one or two tests to keep the output
643        identical in both cases.
644    
645    21. When (*ACCEPT) was used in an assertion that matched an empty string and
646        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
647    
648    22. When an atomic group that contained a capturing parenthesis was
649        successfully matched, but the branch in which it appeared failed, the
650        capturing was not being forgotten if a higher numbered group was later
651        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
652        group 1 to "a", when in fact it should be unset. This applied to multi-
653        branched capturing and non-capturing groups, repeated or not, and also to
654        positive assertions (capturing in negative assertions does not happen
655        in PCRE) and also to nested atomic groups.
656    
657    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
658        subject after a captured substring, to make it easier to tell which of a
659        number of identical substrings has been captured.
660    
661    24. The way atomic groups are processed by pcre_exec() has been changed so that
662        if they are repeated, backtracking one repetition now resets captured
663        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
664        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
665        "aaa". Previously, it would have been "a". As part of this code
666        refactoring, the way recursive calls are handled has also been changed.
667    
668    25. If an assertion condition captured any substrings, they were not passed
669        back unless some other capturing happened later. For example, if
670        (?(?=(a))a) was matched against "a", no capturing was returned.
671    
672    26. When studying a pattern that contained subroutine calls or assertions,
673        the code for finding the minimum length of a possible match was handling
674        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
675        group 1 called group 2 while simultaneously a separate group 2 called group
676        1). A stack overflow occurred in this case. I have fixed this by limiting
677        the recursion depth to 10.
678    
679    27. Updated RunTest.bat in the distribution to the version supplied by Tom
680        Fortmann. This supports explicit test numbers on the command line, and has
681        argument validation and error reporting.
682    
683    28. An instance of \X with an unlimited repeat could fail if at any point the
684        first character it looked at was a mark character.
685    
686    29. Some minor code refactoring concerning Unicode properties and scripts
687        should reduce the stack requirement of match() slightly.
688    
689    30. Added the '=' option to pcretest to check the setting of unused capturing
690        slots at the end of the pattern, which are documented as being -1, but are
691        not included in the return count.
692    
693    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
694        compiled something random. Now it gives a compile-time error (as does
695        Perl).
696    
697    32. A *MARK encountered during the processing of a positive assertion is now
698        recorded and passed back (compatible with Perl).
699    
700    33. If --only-matching or --colour was set on a pcregrep call whose pattern
701        had alternative anchored branches, the search for a second match in a line
702        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
703        matched the line "0102" twice. The same bug affected patterns that started
704        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
705        twice.
706    
707    34. Previously, PCRE did not allow quantification of assertions. However, Perl
708        does, and because of capturing effects, quantifying parenthesized
709        assertions may at times be useful. Quantifiers are now allowed for
710        parenthesized assertions.
711    
712    35. A minor code tidy in pcre_compile() when checking options for \R usage.
713    
714    36. \g was being checked for fancy things in a character class, when it should
715        just be a literal "g".
716    
717    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
718        appearance of a nested POSIX class supersedes an apparent external class.
719        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
720        unescaped square brackets may also appear as part of class names. For
721        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
722        more like Perl. (But see 8.20/1 above.)
723    
724    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
725        was because it thought it was \N{name}, which is not supported).
726    
727    39. Add minix to OS list not supporting the -S option in pcretest.
728    
729    40. PCRE tries to detect cases of infinite recursion at compile time, but it
730        cannot analyze patterns in sufficient detail to catch mutual recursions
731        such as ((?1))((?2)). There is now a runtime test that gives an error if a
732        subgroup is called recursively as a subpattern for a second time at the
733        same position in the subject string. In previous releases this might have
734        been caught by the recursion limit, or it might have run out of stack.
735    
736    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
737        happen only once. PCRE was, however incorrectly giving a compile time error
738        "recursive call could loop indefinitely" because it cannot analyze the
739        pattern in sufficient detail. The compile time test no longer happens when
740        PCRE is compiling a conditional subpattern, but actual runaway loops are
741        now caught at runtime (see 40 above).
742    
743    42. It seems that Perl allows any characters other than a closing parenthesis
744        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
745        has been changed to be the same.
746    
747    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
748        as not to get warnings when autogen.sh is called. Also changed
749        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
750    
751    44. To help people who use pcregrep to scan files containing exceedingly long
752        lines, the following changes have been made:
753    
754        (a) The default value of the buffer size parameter has been increased from
755            8K to 20K. (The actual buffer used is three times this size.)
756    
757        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
758            PCRE is built.
759    
760        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
761            to be set at run time.
762    
763        (d) Numerical values in pcregrep options can be followed by K or M, for
764            example --buffer-size=50K.
765    
766        (e) If a line being scanned overflows pcregrep's buffer, an error is now
767            given and the return code is set to 2.
768    
769    45. Add a pointer to the latest mark to the callout data block.
770    
771    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
772        partial match of an empty string instead of no match. This was specific to
773        the use of ".".
774    
775    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
776        complete match instead of a partial match. This bug was dependent on both
777        the PCRE_UTF8 and PCRE_DOTALL options being set.
778    
779    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
780        starting byte set, because \b was not being ignored.
781    
782    
783    Version 8.12 15-Jan-2011
784    ------------------------
785    
786    1.  Fixed some typos in the markup of the man pages, and wrote a script that
787        checks for such things as part of the documentation building process.
788    
789    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
790        --match-limit and --recursion-limit options (added for 8.11). In
791        particular, this made one of the standard tests fail. (The integer value
792        went into the wrong half of a long int.)
793    
794    3.  If the --colour option was given to pcregrep with -v (invert match), it
795        did strange things, either producing crazy output, or crashing. It should,
796        of course, ignore a request for colour when reporting lines that do not
797        match.
798    
799    4.  Another pcregrep bug caused similar problems if --colour was specified with
800        -M (multiline) and the pattern match finished with a line ending.
801    
802    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
803        matched in multiline mode, the following line was shown as part of the
804        match. This seems wrong, so I have changed it.
805    
806    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
807        the check for further matches in the same line (so they could be coloured)
808        to overrun the end of the current line. If another match was found, it was
809        incorrectly shown (and then shown again when found in the next line).
810    
811    7.  If pcregrep was compiled under Windows, there was a reference to the
812        function pcregrep_exit() before it was defined. I am assuming this was
813        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
814        reported by a user. I've moved the definition above the reference.
815    
816    
817    Version 8.11 10-Dec-2010
818    ------------------------
819    
820    1.  (*THEN) was not working properly if there were untried alternatives prior
821        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
822        backtracked to try for "b" instead of moving to the next alternative branch
823        at the same level (in this case, to look for "c"). The Perl documentation
824        is clear that when (*THEN) is backtracked onto, it goes to the "next
825        alternative in the innermost enclosing group".
826    
827    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
828        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
829        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
830        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
831        (*THEN).
832    
833    3.  If \s appeared in a character class, it removed the VT character from
834        the class, even if it had been included by some previous item, for example
835        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
836        of \s, but is part of the POSIX "space" class.)
837    
838    4.  A partial match never returns an empty string (because you can always
839        match an empty string at the end of the subject); however the checking for
840        an empty string was starting at the "start of match" point. This has been
841        changed to the "earliest inspected character" point, because the returned
842        data for a partial match starts at this character. This means that, for
843        example, /(?<=abc)def/ gives a partial match for the subject "abc"
844        (previously it gave "no match").
845    
846    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
847        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
848        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
849        has an implication that the given string is incomplete (because a partial
850        match is preferred over a full match). For this reason, these items now
851        give a partial match in this situation. [Aside: previously, the one case
852        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
853        match rather than a full match, which was wrong by the old rules, but is
854        now correct.]
855    
856    6.  There was a bug in the handling of #-introduced comments, recognized when
857        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
858        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
859        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
860        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
861        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
862        places in pcre_compile().
863    
864    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
865        comments when looking ahead for named forward references to subpatterns,
866        the only newline sequence it recognized was NL. It now handles newlines
867        according to the set newline convention.
868    
869    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
870        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
871        cater for a lack of strerror(). These oversights have been fixed.
872    
873    9.  Added --match-limit and --recursion-limit to pcregrep.
874    
875    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
876    
877    11. When the -o option was used, pcregrep was setting a return code of 1, even
878        when matches were found, and --line-buffered was not being honoured.
879    
880    12. Added an optional parentheses number to the -o and --only-matching options
881        of pcregrep.
882    
883    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
884        can match an empty string. The code to do it in pcretest and pcredemo
885        needed fixing:
886    
887        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
888            only one byte after an empty string match just before CRLF (this case
889            just got forgotten; "any" and "anycrlf" were OK).
890    
891        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
892            mode when an empty string match preceded an ASCII character followed by
893            a non-ASCII character. (The code for advancing by one character rather
894            than one byte was nonsense.)
895    
896        (c) The pcredemo.c sample program did not have any code at all to handle
897            the cases when CRLF is a valid newline sequence.
898    
899    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
900        as a starting offset was within the subject string. There is now a new
901        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
902        negative or greater than the length of the string. In order to test this,
903        pcretest is extended to allow the setting of negative starting offsets.
904    
905    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
906        starting offset points to the beginning of a UTF-8 character was
907        unnecessarily clumsy. I tidied it up.
908    
909    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
910        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
911    
912    17. Nobody had reported that the --include_dir option, which was added in
913        release 7.7 should have been called --include-dir (hyphen, not underscore)
914        for compatibility with GNU grep. I have changed it to --include-dir, but
915        left --include_dir as an undocumented synonym, and the same for
916        --exclude-dir, though that is not available in GNU grep, at least as of
917        release 2.5.4.
918    
919    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
920        characters from a string of bytes) have been redefined so as not to use
921        loops, in order to improve performance in some environments. At the same
922        time, I abstracted some of the common code into auxiliary macros to save
923        repetition (this should not affect the compiled code).
924    
925    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
926        compile-time error is now given if \c is not followed by an ASCII
927        character, that is, a byte less than 128. (In EBCDIC mode, the code is
928        different, and any byte value is allowed.)
929    
930    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
931        START_OPTIMIZE option, which is now allowed at compile time - but just
932        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
933        to pcregrep and other applications that have no direct access to PCRE
934        options. The new /Y option in pcretest sets this option when calling
935        pcre_compile().
936    
937    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
938        back references. Groups containing recursive back references were forced to
939        be atomic by that change, but in the case of named groups, the amount of
940        memory required was incorrectly computed, leading to "Failed: internal
941        error: code overflow". This has been fixed.
942    
943    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
944        pcretest.c, to avoid build problems in some Borland environments.
945    
946    
947    Version 8.10 25-Jun-2010
948    ------------------------
949    
950    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
951        THEN.
952    
953    2.  (*ACCEPT) was not working when inside an atomic group.
954    
955    3.  Inside a character class, \B is treated as a literal by default, but
956        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
957        causes the error). The code is unchanged, but I tidied the documentation.
958    
959    4.  Inside a character class, PCRE always treated \R and \X as literals,
960        whereas Perl faults them if its -w option is set. I have changed PCRE so
961        that it faults them when PCRE_EXTRA is set.
962    
963    5.  Added support for \N, which always matches any character other than
964        newline. (It is the same as "." when PCRE_DOTALL is not set.)
965    
966    6.  When compiling pcregrep with newer versions of gcc which may have
967        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
968        declared with attribute warn_unused_result" were given. Just casting the
969        result to (void) does not stop the warnings; a more elaborate fudge is
970        needed. I've used a macro to implement this.
971    
972    7.  Minor change to pcretest.c to avoid a compiler warning.
973    
974    8.  Added four artifical Unicode properties to help with an option to make
975        \s etc use properties (see next item). The new properties are: Xan
976        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
977    
978    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
979        use Unicode properties. (*UCP) at the start of a pattern can be used to set
980        this option. Modified pcretest to add /W to test this facility. Added
981        REG_UCP to make it available via the POSIX interface.
982    
983    10. Added --line-buffered to pcregrep.
984    
985    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
986        studied, and the match started with a letter with a code point greater than
987        127 whose first byte was different to the first byte of the other case of
988        the letter, the other case of this starting letter was not recognized
989        (#976).
990    
991    12. If a pattern that was studied started with a repeated Unicode property
992        test, for example, \p{Nd}+, there was the theoretical possibility of
993        setting up an incorrect bitmap of starting bytes, but fortunately it could
994        not have actually happened in practice until change 8 above was made (it
995        added property types that matched character-matching opcodes).
996    
997    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
998        possible starting bytes for non-anchored patterns.
999    
1000    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1001        \R, and also a number of cases that involve Unicode properties, both
1002        explicit and implicit when PCRE_UCP is set.
1003    
1004    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1005        input, it could crash or give wrong results if characters with values
1006        greater than 0xc0 were present in the subject string. (Detail: it assumed
1007        UTF-8 input when processing these items.)
1008    
1009    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1010        size_t is 64-bit (#991).
1011    
1012    17. Added a check for running out of memory when PCRE is compiled with
1013        --disable-stack-for-recursion (#990).
1014    
1015    18. If the last data line in a file for pcretest does not have a newline on
1016        the end, a newline was missing in the output.
1017    
1018    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1019        less than 128) in its various bitmaps. However, there is a facility for
1020        generating tables according to the current locale when PCRE is compiled. It
1021        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1022        characters, are recognized by isspace() and therefore were getting set in
1023        these tables, and indeed these tables seem to approximate to ISO 8859. This
1024        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1025        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1026        which of course cannot start UTF-8 characters. I have changed the code so
1027        that only real ASCII characters (less than 128) and the correct starting
1028        bytes for UTF-8 encodings are set for characters greater than 127 when in
1029        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1030        altogether.)
1031    
1032    20. Added the /T option to pcretest so as to be able to run tests with non-
1033        standard character tables, thus making it possible to include the tests
1034        used for 19 above in the standard set of tests.
1035    
1036    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1037        reference to a subpattern the other side of a comment that contains an
1038        opening parenthesis caused either an internal compiling error, or a
1039        reference to the wrong subpattern.
1040    
1041    
1042    Version 8.02 19-Mar-2010
1043    ------------------------
1044    
1045    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1046    
1047    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1048        configured.
1049    
1050    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1051        original author of that file, following a query about its status.
1052    
1053    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1054        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1055    
1056    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1057        quantifier applied to a forward-referencing subroutine call, could compile
1058        incorrect code or give the error "internal error: previously-checked
1059        referenced subpattern not found".
1060    
1061    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1062        variables to point to external functions. For these systems, therefore,
1063        pcre_malloc etc. are now initialized to local functions that call the
1064        relevant global functions.
1065    
1066    7.  There were two entries missing in the vectors called coptable and poptable
1067        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1068        I've fixed the data, and added a kludgy way of testing at compile time that
1069        the lengths are correct (equal to the number of opcodes).
1070    
1071    8.  Following on from 7, I added a similar kludge to check the length of the
1072        eint vector in pcreposix.c.
1073    
1074    9.  Error texts for pcre_compile() are held as one long string to avoid too
1075        much relocation at load time. To find a text, the string is searched,
1076        counting zeros. There was no check for running off the end of the string,
1077        which could happen if a new error number was added without updating the
1078        string.
1079    
1080    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1081    
1082    11. \K was not working if it appeared in an atomic group or in a group that
1083        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1084        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1085        the assertion is positive, but not if it is negative.
1086    
1087    12. Change 11 fortuitously reduced the size of the stack frame used in the
1088        "match()" function of pcre_exec.c by one pointer. Forthcoming
1089        implementation of support for (*MARK) will need an extra pointer on the
1090        stack; I have reserved it now, so that the stack frame size does not
1091        decrease.
1092    
1093    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1094        item in branch that calls a recursion is a subroutine call - as in the
1095        second branch in the above example - was incorrectly given the compile-
1096        time error "recursive call could loop indefinitely" because pcre_compile()
1097        was not correctly checking the subroutine for matching a non-empty string.
1098    
1099    14. The checks for overrunning compiling workspace could trigger after an
1100        overrun had occurred. This is a "should never occur" error, but it can be
1101        triggered by pathological patterns such as hundreds of nested parentheses.
1102        The checks now trigger 100 bytes before the end of the workspace.
1103    
1104    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1105    
1106    
1107    Version 8.01 19-Jan-2010
1108    ------------------------
1109    
1110    1.  If a pattern contained a conditional subpattern with only one branch (in
1111        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1112        computed the wrong minimum data length (which is of course zero for such
1113        subpatterns). This could cause incorrect "no match" results.
1114    
1115    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1116        the pattern is reset in the first branch, pcre_compile() failed with
1117        "internal error: code overflow at offset...". This happened only when
1118        the reset was to the original external option setting. (An optimization
1119        abstracts leading options settings into an external setting, which was the
1120        cause of this.)
1121    
1122    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1123        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1124        assertion pattern did not match (meaning that the assertion was true), it
1125        was incorrectly treated as false if the SKIP had been reached during the
1126        matching. This also applied to assertions used as conditions.
1127    
1128    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1129        assertion subpattern, including such a pattern used as a condition,
1130        unpredictable results occurred, instead of the error return
1131        PCRE_ERROR_DFA_UITEM.
1132    
1133    5.  The C++ GlobalReplace function was not working like Perl for the special
1134        situation when an empty string is matched. It now does the fancy magic
1135        stuff that is necessary.
1136    
1137    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1138        removed. (These were left over from very, very early versions of PCRE.)
1139    
1140    7.  Some cosmetic changes to the code to make life easier when compiling it
1141        as part of something else:
1142    
1143        (a) Change DEBUG to PCRE_DEBUG.
1144    
1145        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1146            called "current" as "current_branch", to prevent a collision with the
1147            Linux macro when compiled as a kernel module.
1148    
1149        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1150            prevent a collision with the Linux macro when compiled as a kernel
1151            module.
1152    
1153    8.  In pcre_compile() there are some checks for integer overflows that used to
1154        cast potentially large values to (double). This has been changed to that
1155        when building, a check for int64_t is made, and if it is found, it is used
1156        instead, thus avoiding the use of floating point arithmetic. (There is no
1157        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1158        double.
1159    
1160    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1161        2005 (difference between two addresses compared to an unsigned value).
1162    
1163    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1164        custom one, because of the following reported problem in Windows:
1165    
1166          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1167              under Win32.
1168          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1169              therefore missing the function definition.
1170          - The compiler thus generates a "C" signature for the test function.
1171          - The linker fails to find the "C" function.
1172          - PCRE fails to configure if asked to do so against libbz2.
1173    
1174    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1175        messages were output:
1176    
1177          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1178          rerunning libtoolize, to keep the correct libtool macros in-tree.
1179          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1180    
1181        I have done both of these things.
1182    
1183    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1184        most of the time, it *can* run out if it is given a pattern that contains a
1185        runaway infinite recursion. I updated the discussion in the pcrestack man
1186        page.
1187    
1188    13. Now that we have gone to the x.xx style of version numbers, the minor
1189        version may start with zero. Using 08 or 09 is a bad idea because users
1190        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1191        interpreted as invalid octal numbers. I've updated the previous comment in
1192        configure.ac, and also added a check that gives an error if 08 or 09 are
1193        used.
1194    
1195    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1196        causing partial matching to fail when the end of the subject matched \W
1197        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1198    
1199    15. There were some discrepancies between the declarations in pcre_internal.h
1200        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1201        their definitions. The declarations used "const uschar *" and the
1202        definitions used USPTR. Even though USPTR is normally defined as "const
1203        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1204        reported that: "This difference in casting confuses some C++ compilers, for
1205        example, SunCC recognizes above declarations as different functions and
1206        generates broken code for hbpcre." I have changed the declarations to use
1207        USPTR.
1208    
1209    16. GNU libtool is named differently on some systems. The autogen.sh script now
1210        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1211        (FreeBSD).
1212    
1213    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1214        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1215        comment: "Figure out how to create a longlong from a string: strtoll and
1216        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1217        instance, but it only takes 2 args instead of 3!"
1218    
1219    18. A subtle bug concerned with back references has been fixed by a change of
1220        specification, with a corresponding code fix. A pattern such as
1221        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1222        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1223        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1224        same bug. Such groups have to be quantified to be useful, or contained
1225        inside another quantified group. (If there's no repetition, the reference
1226        can never match.) The problem arises because, having left the group and
1227        moved on to the rest of the pattern, a later failure that backtracks into
1228        the group uses the captured value from the final iteration of the group
1229        rather than the correct earlier one. I have fixed this in PCRE by forcing
1230        any group that contains a reference to itself to be an atomic group; that
1231        is, there cannot be any backtracking into it once it has completed. This is
1232        similar to recursive and subroutine calls.
1233    
1234    
1235    Version 8.00 19-Oct-09
1236    ----------------------
1237    
1238    1.  The table for translating pcre_compile() error codes into POSIX error codes
1239        was out-of-date, and there was no check on the pcre_compile() error code
1240        being within the table. This could lead to an OK return being given in
1241        error.
1242    
1243    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1244        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1245        in a Windows environment.
1246    
1247    3.  The pcregrep --count option prints the count for each file even when it is
1248        zero, as does GNU grep. However, pcregrep was also printing all files when
1249        --files-with-matches was added. Now, when both options are given, it prints
1250        counts only for those files that have at least one match. (GNU grep just
1251        prints the file name in this circumstance, but including the count seems
1252        more useful - otherwise, why use --count?) Also ensured that the
1253        combination -clh just lists non-zero counts, with no names.
1254    
1255    4.  The long form of the pcregrep -F option was incorrectly implemented as
1256        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1257        but it seems right to fix it, and I didn't think it was worth preserving
1258        the old behaviour.
1259    
1260    5.  The command line items --regex=pattern and --regexp=pattern were not
1261        recognized by pcregrep, which required --regex pattern or --regexp pattern
1262        (with a space rather than an '='). The man page documented the '=' forms,
1263        which are compatible with GNU grep; these now work.
1264    
1265    6.  No libpcreposix.pc file was created for pkg-config; there was just
1266        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1267    
1268    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1269        when UCP support is not needed, by modifying the Python script that
1270        generates it from Unicode data files. This should not matter if the module
1271        is correctly used as a library, but I received one complaint about 50K of
1272        unwanted data. My guess is that the person linked everything into his
1273        program rather than using a library. Anyway, it does no harm.
1274    
1275    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1276        was a minimum greater than 1 for a wide character in a possessive
1277        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1278        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1279        character. Chaos in the form of incorrect output or a compiling loop could
1280        result.
1281    
1282    9.  The restrictions on what a pattern can contain when partial matching is
1283        requested for pcre_exec() have been removed. All patterns can now be
1284        partially matched by this function. In addition, if there are at least two
1285        slots in the offset vector, the offset of the earliest inspected character
1286        for the match and the offset of the end of the subject are set in them when
1287        PCRE_ERROR_PARTIAL is returned.
1288    
1289    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1290        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1291        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1292        and may be more useful for multi-segment matching.
1293    
1294    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1295        used to be given if ever the end of the subject was reached; now it is
1296        given only if matching could not proceed because another character was
1297        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1298        string "Z", which now yields "no match" instead of "partial match". In the
1299        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1300        final character ended with (*FAIL).
1301    
1302    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1303        if the pattern had a "must contain" character that was already found in the
1304        earlier partial match, unless partial matching was again requested. For
1305        example, with the pattern /dog.(body)?/, the "must contain" character is
1306        "g". If the first part-match was for the string "dog", restarting with
1307        "sbody" failed. This bug has been fixed.
1308    
1309    13. The string returned by pcre_dfa_exec() after a partial match has been
1310        changed so that it starts at the first inspected character rather than the
1311        first character of the match. This makes a difference only if the pattern
1312        starts with a lookbehind assertion or \b or \B (\K is not supported by
1313        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1314        matching functions compatible, and I think it's the right thing to do.
1315    
1316    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1317        so that the demonstration program is easily available in environments where
1318        PCRE has not been installed from source.
1319    
1320    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1321        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1322        library.
1323    
1324    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1325        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1326        is not the first non-POSIX option to be added. Clearly some people find
1327        these options useful.
1328    
1329    17. If a caller to the POSIX matching function regexec() passes a non-zero
1330        value for nmatch with a NULL value for pmatch, the value of
1331        nmatch is forced to zero.
1332    
1333    18. RunGrepTest did not have a test for the availability of the -u option of
1334        the diff command, as RunTest does. It now checks in the same way as
1335        RunTest, and also checks for the -b option.
1336    
1337    19. If an odd number of negated classes containing just a single character
1338        interposed, within parentheses, between a forward reference to a named
1339        subpattern and the definition of the subpattern, compilation crashed with
1340        an internal error, complaining that it could not find the referenced
1341        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1342        [The bug was that it was starting one character too far in when skipping
1343        over the character class, thus treating the ] as data rather than
1344        terminating the class. This meant it could skip too much.]
1345    
1346    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1347        /g option in pcretest when the pattern contains \K, which makes it possible
1348        to have an empty string match not at the start, even when the pattern is
1349        anchored. Updated pcretest and pcredemo to use this option.
1350    
1351    21. If the maximum number of capturing subpatterns in a recursion was greater
1352        than the maximum at the outer level, the higher number was returned, but
1353        with unset values at the outer level. The correct (outer level) value is
1354        now given.
1355    
1356    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1357        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1358        make it do so. The string so far is captured, making this feature
1359        compatible with Perl.
1360    
1361    23. The tests have been re-organized, adding tests 11 and 12, to make it
1362        possible to check the Perl 5.10 features against Perl 5.10.
1363    
1364    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1365        pattern matches a fixed length string. PCRE did not allow this; now it
1366        does. Neither allows recursion.
1367    
1368    25. I finally figured out how to implement a request to provide the minimum
1369        length of subject string that was needed in order to match a given pattern.
1370        (It was back references and recursion that I had previously got hung up
1371        on.) This code has now been added to pcre_study(); it finds a lower bound
1372        to the length of subject needed. It is not necessarily the greatest lower
1373        bound, but using it to avoid searching strings that are too short does give
1374        some useful speed-ups. The value is available to calling programs via
1375        pcre_fullinfo().
1376    
1377    26. While implementing 25, I discovered to my embarrassment that pcretest had
1378        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1379        study optimizations had never been tested with that matching function.
1380        Oops. What is worse, even when it was passed study data, there was a bug in
1381        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1382        were also very few tests of studied patterns with pcre_dfa_exec().
1383    
1384    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1385        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1386        on the other side of the coin, they are no longer allowed to have different
1387        names, because these cannot be distinguished in PCRE, and this has caused
1388        confusion. (This is a difference from Perl.)
1389    
1390    28. When duplicate subpattern names are present (necessarily with different
1391        numbers, as required by 27 above), and a test is made by name in a
1392        conditional pattern, either for a subpattern having been matched, or for
1393        recursion in such a pattern, all the associated numbered subpatterns are
1394        tested, and the overall condition is true if the condition is true for any
1395        one of them. This is the way Perl works, and is also more like the way
1396        testing by number works.
1397    
1398    
1399    Version 7.9 11-Apr-09
1400    ---------------------
1401    
1402    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1403        (pcretest), all targets were linked against these libraries. This included
1404        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1405        libraries. This caused unwanted dependencies to be created. This problem
1406        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1407        pcretest is linked with readline.
1408    
1409    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1410        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1411        moved outside it again, because FALSE and TRUE are already defined in AIX,
1412        but BOOL is not.
1413    
1414    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1415        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1416    
1417    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1418        hyphens) following file names and line numbers when outputting matching
1419        lines. This is not true; no spaces are inserted. I have also clarified the
1420        wording for the --colour (or --color) option.
1421    
1422    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1423        was not coloured; this is different to GNU grep, so I have changed it to be
1424        the same.
1425    
1426    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1427        each matching line was coloured. Now it goes on to look for further matches
1428        of any of the test patterns, which is the same behaviour as GNU grep.
1429    
1430    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1431        doesn't make sense to accept an empty string match in pcregrep, so I have
1432        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1433        seems to be how GNU grep behaves.
1434    
1435    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1436        start or after a newline", because the conditional assertion was not being
1437        correctly handled. The rule now is that both the assertion and what follows
1438        in the first alternative must satisfy the test.
1439    
1440    9.  If auto-callout was enabled in a pattern with a conditional group whose
1441        condition was an assertion, PCRE could crash during matching, both with
1442        pcre_exec() and pcre_dfa_exec().
1443    
1444    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1445        used for matching.
1446    
1447    11. Unicode property support in character classes was not working for
1448        characters (bytes) greater than 127 when not in UTF-8 mode.
1449    
1450    12. Added the -M command line option to pcretest.
1451    
1452    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1453    
1454    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1455    
1456    16. Added comments and documentation about mis-use of no_arg in the C++
1457        wrapper.
1458    
1459    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1460        from Martin Jerabek that uses macro names for all relevant character and
1461        string constants.
1462    
1463    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1464        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1465        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1466        these, but not everybody uses configure.
1467    
1468    19. A conditional group that had only one branch was not being correctly
1469        recognized as an item that could match an empty string. This meant that an
1470        enclosing group might also not be so recognized, causing infinite looping
1471        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1472        with the subject "ab", where knowledge that the repeated group can match
1473        nothing is needed in order to break the loop.
1474    
1475    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1476        exec(), but without supplying a callout function, matching went wrong.
1477    
1478    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1479        leak if the size of the offset vector was greater than 30. When the vector
1480        is smaller, the saved offsets during recursion go onto a local stack
1481        vector, but for larger vectors malloc() is used. It was failing to free
1482        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1483        error, in fact).
1484    
1485    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1486        heapframe that is used only when UTF-8 support is enabled. This caused no
1487        problem, but was untidy.
1488    
1489    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1490        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1491        included within another project.
1492    
1493    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1494        slightly modified by me:
1495    
1496          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1497              not building pcregrep.
1498    
1499          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1500              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1501    
1502    25. Forward references, both numeric and by name, in patterns that made use of
1503        duplicate group numbers, could behave incorrectly or give incorrect errors,
1504        because when scanning forward to find the reference group, PCRE was not
1505        taking into account the duplicate group numbers. A pattern such as
1506        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1507    
1508    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1509        the feature of a custom pointer more persuasive (as requested by a user).
1510    
1511    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1512        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1513        pre-defined.
1514    
1515    28. Added support for (*UTF8) at the start of a pattern.
1516    
1517    29. Arrange for flags added by the "release type" setting in CMake to be shown
1518        in the configuration summary.
1519    
1520    
1521    Version 7.8 05-Sep-08
1522    ---------------------
1523    
1524    1.  Replaced UCP searching code with optimized version as implemented for Ad
1525        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1526        stage table and inline lookup instead of a function, giving speed ups of 2
1527        to 5 times on some simple patterns that I tested. Permission was given to
1528        distribute the MultiStage2.py script that generates the tables (it's not in
1529        the tarball, but is in the Subversion repository).
1530    
1531    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1532        scripts.
1533    
1534    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1535        a group with a zero qualifier. The result of the study could be incorrect,
1536        or the function might crash, depending on the pattern.
1537    
1538    4.  Caseless matching was not working for non-ASCII characters in back
1539        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1540        It now works when Unicode Property Support is available.
1541    
1542    5.  In pcretest, an escape such as \x{de} in the data was always generating
1543        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1544        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1545        truncation.
1546    
1547    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1548    
1549    7.  Added two (int) casts to pcregrep when printing the difference of two
1550        pointers, in case they are 64-bit values.
1551    
1552    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1553        test 2 if it fails.
1554    
1555    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1556        and a #define of that name to empty if it is not externally set. This is to
1557        allow users of MSVC to set it if necessary.
1558    
1559    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1560        the convenience functions in the pcre_get.c source file.
1561    
1562    11. An option change at the start of a pattern that had top-level alternatives
1563        could cause overwriting and/or a crash. This command provoked a crash in
1564        some environments:
1565    
1566          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1567    
1568        This potential security problem was recorded as CVE-2008-2371.
1569    
1570    12. For a pattern where the match had to start at the beginning or immediately
1571        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1572        pcre_dfa_exec() could read past the end of the passed subject if there was
1573        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1574        pcretest so that it places the subject at the end of its malloc-ed buffer.
1575    
1576    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1577        exec() might read past the end of the data buffer in UTF-8 mode.
1578    
1579    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1580        the data contained the byte 0x85 as part of a UTF-8 character within its
1581        first line. This applied both to normal and DFA matching.
1582    
1583    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1584        /^[^d]*?$/8 failed to match "abc".
1585    
1586    16. Added a missing copyright notice to pcrecpp_internal.h.
1587    
1588    17. Make it more clear in the documentation that values returned from
1589        pcre_exec() in ovector are byte offsets, not character counts.
1590    
1591    18. Tidied a few places to stop certain compilers from issuing warnings.
1592    
1593    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1594        supplied by Stefan Weber. I made a further small update for 7.8 because
1595        there is a change of source arrangements: the pcre_searchfuncs.c module is
1596        replaced by pcre_ucd.c.
1597    
1598    
1599    Version 7.7 07-May-08
1600    ---------------------
1601    
1602    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1603        a string to a long long, pretend we don't even have a long long." This is
1604        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1605    
1606    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1607        pre-7.6 versions, which defined a global no_arg variable instead of putting
1608        it in the RE class. (See also #8 below.)
1609    
1610    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1611        Lopes.
1612    
1613    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1614    
1615        (1) The include/exclude patterns were being applied to the whole pathnames
1616            of files, instead of just to the final components.
1617    
1618        (2) If there was more than one level of directory, the subdirectories were
1619            skipped unless they satisfied the include/exclude conditions. This is
1620            inconsistent with GNU grep (and could even be seen as contrary to the
1621            pcregrep specification - which I improved to make it absolutely clear).
1622            The action now is always to scan all levels of directory, and just
1623            apply the include/exclude patterns to regular files.
1624    
1625    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1626        --exclude_dir in the tests to avoid scanning .svn directories.
1627    
1628    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1629        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1630        doesn't support NULs in patterns.
1631    
1632    7.  Added some missing "const"s to declarations of static tables in
1633        pcre_compile.c and pcre_dfa_exec.c.
1634    
1635    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1636        caused by fix #2  above. (Subsequently also a second patch to fix the
1637        first patch. And a third patch - this was a messy problem.)
1638    
1639    9.  Applied Craig's patch to remove the use of push_back().
1640    
1641    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1642        matching function regexec().
1643    
1644    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1645        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1646        references. PCRE supports relative numbers with this syntax (I don't think
1647        Oniguruma does).
1648    
1649    12. Previously, a group with a zero repeat such as (...){0} was completely
1650        omitted from the compiled regex. However, this means that if the group
1651        was called as a subroutine from elsewhere in the pattern, things went wrong
1652        (an internal error was given). Such groups are now left in the compiled
1653        pattern, with a new opcode that causes them to be skipped at execution
1654        time.
1655    
1656    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1657        to the way PCRE behaves:
1658    
1659        (a) A lone ] character is dis-allowed (Perl treats it as data).
1660    
1661        (b) A back reference to an unmatched subpattern matches an empty string
1662            (Perl fails the current match path).
1663    
1664        (c) A data ] in a character class must be notated as \] because if the
1665            first data character in a class is ], it defines an empty class. (In
1666            Perl it is not possible to have an empty class.) The empty class []
1667            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1668            The negative empty class [^] matches any one character, independently
1669            of the DOTALL setting.
1670    
1671    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1672        non-existent subpattern following a character class starting with ']' and
1673        containing () gave an internal compiling error instead of "reference to
1674        non-existent subpattern". Fortunately, when the pattern did exist, the
1675        compiled code was correct. (When scanning forwards to check for the
1676        existencd of the subpattern, it was treating the data ']' as terminating
1677        the class, so got the count wrong. When actually compiling, the reference
1678        was subsequently set up correctly.)
1679    
1680    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1681        it was being rejected as not supported by pcre_dfa_exec(), even though
1682        other assertions are supported. I have made pcre_dfa_exec() support
1683        (*FAIL).
1684    
1685    16. The implementation of 13c above involved the invention of a new opcode,
1686        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1687        cannot be changed at match time, I realized I could make a small
1688        improvement to matching performance by compiling OP_ALLANY instead of
1689        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1690        on the OP_ANY path.
1691    
1692    17. Compiling pcretest on Windows with readline support failed without the
1693        following two fixes: (1) Make the unistd.h include conditional on
1694        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1695    
1696    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1697        ncurses library to be included for pcretest when ReadLine support is
1698        requested, but also to allow for it to be overridden. This patch came from
1699        Daniel Bergström.
1700    
1701    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1702        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1703        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1704        spotting this.
1705    
1706    
1707    Version 7.6 28-Jan-08
1708    ---------------------
1709    
1710    1.  A character class containing a very large number of characters with
1711        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1712        overflow.
1713    
1714    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1715        HAVE_LONG_LONG is not defined.
1716    
1717    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1718        bring them up to date and include new features. This patch includes:
1719    
1720        - Fixed PH's badly added libz and libbz2 support.
1721        - Fixed a problem with static linking.
1722        - Added pcredemo. [But later removed - see 7 below.]
1723        - Fixed dftables problem and added an option.
1724        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1725            HAVE_LONG_LONG.
1726        - Added readline support for pcretest.
1727        - Added an listing of the option settings after cmake has run.
1728    
1729    4.  A user submitted a patch to Makefile that makes it easy to create
1730        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1731        Makefile.am that cause it to include this special target, without
1732        affecting anything else. Note that the same mingw target plus all
1733        the other distribution libraries and programs are now supported
1734        when configuring with CMake (see 6 below) instead of with
1735        Configure/Make.
1736    
1737    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1738        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1739        exported in the Windows port". It has not yet been confirmed that the patch
1740        solves the problem, but it does no harm.
1741    
1742    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1743        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1744        with CMake, and also correct the comment about stack recursion.
1745    
1746    7.  Remove the automatic building of pcredemo from the ./configure system and
1747        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1748        of a program that users should build themselves after PCRE is installed, so
1749        building it automatically is not really right. What is more, it gave
1750        trouble in some build environments.
1751    
1752    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1753    
1754    
1755    Version 7.5 10-Jan-08
1756    ---------------------
1757    
1758    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1759        values in parens when parsing an RE using the C++ wrapper."
1760    
1761    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1762        Characters greater than 255 were excluded from the class instead of being
1763        included.
1764    
1765    3.  The same bug as (2) above applied to negated POSIX classes such as
1766        [:^space:].
1767    
1768    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1769        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1770        I have changed it.
1771    
1772    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1773        first named subpattern) and a construct such as (?&a) would reference the
1774        first named subpattern whose name started with "a" (in other words, the
1775        length check was missing). Both these problems are fixed. "Subpattern name
1776        expected" is now given for (?&) (a zero-length name), and this patch also
1777        makes it give the same error for \k'' (previously it complained that that
1778        was a reference to a non-existent subpattern).
1779    
1780    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1781        this is right because (?- can be followed by option settings as well as by
1782        digits. I have, however, made the messages clearer.
1783    
1784    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1785        than the number used in the conditional) now cause a compile-time error.
1786        This is actually not compatible with Perl, which accepts such patterns, but
1787        treats the conditional as always being FALSE (as PCRE used to), but it
1788        seems to me that giving a diagnostic is better.
1789    
1790    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1791        and messages.
1792    
1793    9.  Fix two occurrences of "backslash" in comments that should have been
1794        "backspace".
1795    
1796    10. Remove two redundant lines of code that can never be obeyed (their function
1797        was moved elsewhere).
1798    
1799    11. The program that makes PCRE's Unicode character property table had a bug
1800        which caused it to generate incorrect table entries for sequences of
1801        characters that have the same character type, but are in different scripts.
1802        It amalgamated them into a single range, with the script of the first of
1803        them. In other words, some characters were in the wrong script. There were
1804        thirteen such cases, affecting characters in the following ranges:
1805    
1806          U+002b0 - U+002c1
1807          U+0060c - U+0060d
1808          U+0061e - U+00612
1809          U+0064b - U+0065e
1810          U+0074d - U+0076d
1811          U+01800 - U+01805
1812          U+01d00 - U+01d77
1813          U+01d9b - U+01dbf
1814          U+0200b - U+0200f
1815          U+030fc - U+030fe
1816          U+03260 - U+0327f
1817          U+0fb46 - U+0fbb1
1818          U+10450 - U+1049d
1819    
1820    12. The -o option (show only the matching part of a line) for pcregrep was not
1821        compatible with GNU grep in that, if there was more than one match in a
1822        line, it showed only the first of them. It now behaves in the same way as
1823        GNU grep.
1824    
1825    13. If the -o and -v options were combined for pcregrep, it printed a blank
1826        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1827        does the same. The return code can be used to tell if there were any
1828        non-matching lines.
1829    
1830    14. Added --file-offsets and --line-offsets to pcregrep.
1831    
1832    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1833        infinitely looping recursion. The bug was that positive lookaheads were not
1834        being skipped when checking for a possible empty match (negative lookaheads
1835        and both kinds of lookbehind were skipped).
1836    
1837    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1838        inclusion of <windows.h> to before rather than after the definition of
1839        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1840    
1841    17. Specifying a possessive quantifier with a specific limit for a Unicode
1842        character property caused pcre_compile() to compile bad code, which led at
1843        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1844        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1845        caused the error; without that there was no problem.
1846    
1847    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1848    
1849    19. Added --enable-pcretest-libreadline.
1850    
1851    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1852        RE::GlobalReplace(). As a result, the number of replacements returned was
1853        double what it should be. I removed one of the increments, but Craig sent a
1854        later patch that removed the other one (the right fix) and added unit tests
1855        that check the return values (which was not done before).
1856    
1857    21. Several CMake things:
1858    
1859        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1860            the names libpcre and libpcreposix, not just pcre and pcreposix.
1861    
1862        (2) The above change means that pcretest and pcregrep are now correctly
1863            linked with the newly-built libraries, not previously installed ones.
1864    
1865        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1866    
1867    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1868        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1869        UTF-8 newline character). The key issue is that the pattern starts .*;
1870        this means that the match must be either at the beginning, or after a
1871        newline. The bug was in the code for advancing after a failed match and
1872        checking that the new position followed a newline. It was not taking
1873        account of UTF-8 characters correctly.
1874    
1875    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1876        character classes. PCRE was not treating the sequence [:...:] as a
1877        character class unless the ... were all letters. Perl, however, seems to
1878        allow any characters between [: and :], though of course it rejects as
1879        unknown any "names" that contain non-letters, because all the known class
1880        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1881        for example, whereas PCRE did not - it did not recognize a POSIX character
1882        class. This seemed a bit dangerous, so the code has been changed to be
1883        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1884        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1885        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1886        Perl does, and where it didn't before.
1887    
1888    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1889        Windows environments %n is disabled by default.
1890    
1891    
1892    Version 7.4 21-Sep-07
1893    ---------------------
1894    
1895    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1896        means that a class such as [\s] counted as "explicit reference to CR or
1897        LF". That isn't really right - the whole point of the change was to try to
1898        help when there was an actual mention of one of the two characters. So now
1899        the change happens only if \r or \n (or a literal CR or LF) character is
1900        encountered.
1901    
1902    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1903        of both had grown to the point where there were only 3 bits left.
1904        Fortunately, there was spare space in the data structure, and so I have
1905        moved the internal flags into a new 16-bit field to free up more option
1906        bits.
1907    
1908    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1909        but did not set the internal JCHANGED flag - either of these is enough to
1910        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1911        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1912        start sets both bits.
1913    
1914    4.  Added options (at build time, compile time, exec time) to change \R from
1915        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1916    
1917    5.  doc/pcresyntax.html was missing from the distribution.
1918    
1919    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1920        compatibility, even though it is no longer used.
1921    
1922    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1923        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1924        windows.h file is present (where different names are used). [This was
1925        reversed later after testing - see 16 below.]
1926    
1927    8.  Changed all #include <config.h> to #include "config.h". There were also
1928        some further <pcre.h> cases that I changed to "pcre.h".
1929    
1930    9.  When pcregrep was used with the --colour option, it missed the line ending
1931        sequence off the lines that it output.
1932    
1933    10. It was pointed out to me that arrays of string pointers cause lots of
1934        relocations when a shared library is dynamically loaded. A technique of
1935        using a single long string with a table of offsets can drastically reduce
1936        these. I have refactored PCRE in four places to do this. The result is
1937        dramatic:
1938    
1939          Originally:                          290
1940          After changing UCP table:            187
1941          After changing error message table:   43
1942          After changing table of "verbs"       36
1943          After changing table of Posix names   22
1944    
1945        Thanks to the folks working on Gregex for glib for this insight.
1946    
1947    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1948        unicode-properties was also set.
1949    
1950    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1951    
1952    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1953        checked only for CRLF.
1954    
1955    14. Added casts to pcretest.c to avoid compiler warnings.
1956    
1957    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1958    
1959    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1960        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1961        entirely. This removes changes made in 7 above.
1962    
1963    17. The CMake files have been updated, and there is now more information about
1964        building with CMake in the NON-UNIX-USE document.
1965    
1966    
1967    Version 7.3 28-Aug-07
1968    ---------------------
1969    
1970     1. In the rejigging of the build system that eventually resulted in 7.1, the
1971        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1972        brackets there is not right, since it causes compilers to look for an
1973        installed pcre.h, not the version that is in the source that is being
1974        compiled (which of course may be different). I have changed it back to:
1975    
1976          #include "pcre.h"
1977    
1978        I have a vague recollection that the change was concerned with compiling in
1979        different directories, but in the new build system, that is taken care of
1980        by the VPATH setting the Makefile.
1981    
1982     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1983        when the subject happened to end in the byte 0x85 (e.g. if the last
1984        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1985        characters but of course it shouldn't be taken as a newline when it is part
1986        of another character. The bug was that, for an unlimited repeat of . in
1987        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1988        characters when looking for a newline.
1989    
1990     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1991    
1992     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1993        in debug output.
1994    
1995     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1996        long printing in the pcrecpp unittest when running under MinGW.
1997    
1998     6. ESC_K was left out of the EBCDIC table.
1999    
2000     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2001        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2002        limit also applies to "virtual nesting" when a pattern is recursive, and in
2003        this case 1000 isn't so big. I have been able to remove this limit at the
2004        expense of backing off one optimization in certain circumstances. Normally,
2005        when pcre_exec() would call its internal match() function recursively and
2006        immediately return the result unconditionally, it uses a "tail recursion"
2007        feature to save stack. However, when a subpattern that can match an empty
2008        string has an unlimited repetition quantifier, it no longer makes this
2009        optimization. That gives it a stack frame in which to save the data for
2010        checking that an empty string has been matched. Previously this was taken
2011        from the 1000-entry workspace that had been reserved. So now there is no
2012        explicit limit, but more stack is used.
2013    
2014     8. Applied Daniel's patches to solve problems with the import/export magic
2015        syntax that is required for Windows, and which was going wrong for the
2016        pcreposix and pcrecpp parts of the library. These were overlooked when this
2017        problem was solved for the main library.
2018    
2019     9. There were some crude static tests to avoid integer overflow when computing
2020        the size of patterns that contain repeated groups with explicit upper
2021        limits. As the maximum quantifier is 65535, the maximum group length was
2022        set at 30,000 so that the product of these two numbers did not overflow a
2023        32-bit integer. However, it turns out that people want to use groups that
2024        are longer than 30,000 bytes (though not repeat them that many times).
2025        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2026        made it possible to implement the integer overflow checks in a much more
2027        dynamic way, which I have now done. The artificial limitation on group
2028        length has been removed - we now have only the limit on the total length of
2029        the compiled pattern, which depends on the LINK_SIZE setting.
2030    
2031    10. Fixed a bug in the documentation for get/copy named substring when
2032        duplicate names are permitted. If none of the named substrings are set, the
2033        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2034        empty string.
2035    
2036    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2037        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2038        because the ] is interpreted as the first data character and the
2039        terminating ] is not found. PCRE has been made compatible with Perl in this
2040        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2041        cause memory overwriting.
2042    
2043    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2044        string has been matched (to stop an infinite loop). It was not recognizing
2045        a conditional subpattern that could match an empty string if that
2046        subpattern was within another subpattern. For example, it looped when
2047        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2048        condition was not nested. This bug has been fixed.
2049    
2050    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2051        past the start of the subject in the presence of bytes with the top bit
2052        set, for example "\x8aBCD".
2053    
2054    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2055        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2056    
2057    14. Optimized (?!) to (*FAIL).
2058    
2059    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2060        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2061        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2062        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2063        does: it's just the validity check that is more restrictive.
2064    
2065    16. Inserted checks for integer overflows during escape sequence (backslash)
2066        processing, and also fixed erroneous offset values for syntax errors during
2067        backslash processing.
2068    
2069    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2070        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2071    
2072    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2073        caused an overrun.
2074    
2075    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2076        something other than just ASCII characters) inside a group that had an
2077        unlimited repeat caused a loop at compile time (while checking to see
2078        whether the group could match an empty string).
2079    
2080    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2081        [\P{Any}] did so. (Error in the code for printing property names.)
2082    
2083    21. An orphan \E inside a character class could cause a crash.
2084    
2085    22. A repeated capturing bracket such as (A)? could cause a wild memory
2086        reference during compilation.
2087    
2088    23. There are several functions in pcre_compile() that scan along a compiled
2089        expression for various reasons (e.g. to see if it's fixed length for look
2090        behind). There were bugs in these functions when a repeated \p or \P was
2091        present in the pattern. These operators have additional parameters compared
2092        with \d, etc, and these were not being taken into account when moving along
2093        the compiled data. Specifically:
2094    
2095        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2096            length.
2097    
2098        (b) An item such as \pL+ within a repeated group could cause crashes or
2099            loops.
2100    
2101        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2102            "reference to non-existent subpattern" error.
2103    
2104        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2105    
2106    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2107        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2108    
2109    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2110    
2111    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2112        character were causing crashes (broken optimization).
2113    
2114    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2115        \p or \P) caused a compile-time loop.
2116    
2117    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2118        break. For example, the unstudied pattern [\r\n]A does not match the string
2119        "\r\nA" because change 7.0/46 below moves the current point on by two
2120        characters after failing to match at the start. However, the pattern \nA
2121        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2122        the same is true. There doesn't seem any very clean way out of this, but
2123        what I have chosen to do makes the common cases work: PCRE now takes note
2124        of whether there can be an explicit match for \r or \n anywhere in the
2125        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2126        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2127        pattern has explicit CR or LF references.
2128    
2129    29. Added (*CR) etc for changing newline setting at start of pattern.
2130    
2131    
2132    Version 7.2 19-Jun-07
2133    ---------------------
2134    
2135     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2136        which is apparently normally available under Windows.
2137    
2138     2. Re-jig the pcregrep tests with different newline settings in an attempt
2139        to make them independent of the local environment's newline setting.
2140    
2141     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2142    
2143     4. Some of the "internals" tests were previously cut out when the link size
2144        was not 2, because the output contained actual offsets. The recent new
2145        "Z" feature of pcretest means that these can be cut out, making the tests
2146        usable with all link sizes.
2147    
2148     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2149        stack recursion. This gives a massive performance boost under BSD, but just
2150        a small improvement under Linux. However, it saves one field in the frame
2151        in all cases.
2152    
2153     6. Added more features from the forthcoming Perl 5.10:
2154    
2155        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2156            recursion call. It refers to the nth most recently opened parentheses.
2157    
2158        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2159            to be opened parentheses.
2160    
2161        (c) Conditions that refer to capturing parentheses can be specified
2162            relatively, for example, (?(-2)... or (?(+3)...
2163    
2164        (d) \K resets the start of the current match so that everything before
2165            is not part of it.
2166    
2167        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2168    
2169        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2170            reference syntax.
2171    
2172        (g) (?| introduces a group in which the numbering of parentheses in each
2173            alternative starts with the same number.
2174    
2175        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2176    
2177     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2178        PCRE_INFO_JCHANGED.
2179    
2180     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2181        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2182        for detecting groups that can match an empty string.
2183    
2184     9. A pattern with a very large number of alternatives (more than several
2185        hundred) was running out of internal workspace during the pre-compile
2186        phase, where pcre_compile() figures out how much memory will be needed. A
2187        bit of new cunning has reduced the workspace needed for groups with
2188        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2189        workspace instead of running out of the 4096 that are available.
2190    
2191    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2192    
2193    11. Applied patch from Google to remove an optimization that didn't quite work.
2194        The report of the bug said:
2195    
2196          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2197          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2198          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2199    
2200    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2201        it matched the wrong number of bytes.
2202    
2203    
2204    Version 7.1 24-Apr-07
2205    ---------------------
2206    
2207     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2208        that is more "standard", making use of automake and other Autotools. There
2209        is some re-arrangement of the files and adjustment of comments consequent
2210        on this.
2211    
2212     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2213        for recursive directory scanning broke on some systems because the files
2214        are not scanned in any specific order and on different systems the order
2215        was different. A call to "sort" has been inserted into RunGrepTest for the
2216        approprate test as a short-term fix. In the longer term there may be an
2217        alternative.
2218    
2219     3. I had an email from Eric Raymond about problems translating some of PCRE's
2220        man pages to HTML (despite the fact that I distribute HTML pages, some
2221        people do their own conversions for various reasons). The problems
2222        concerned the use of low-level troff macros .br and .in. I have therefore
2223        removed all such uses from the man pages (some were redundant, some could
2224        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2225        HTML has been updated to handle .nf/.fi and to complain if it encounters
2226        .br or .in.
2227    
2228     4. Updated comments in configure.ac that get placed in config.h.in and also
2229        arranged for config.h to be included in the distribution, with the name
2230        config.h.generic, for the benefit of those who have to compile without
2231        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2232    
2233     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2234        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2235        makevp.bat for the current PCRE, using the additional files
2236        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2237    
2238     6. A Windows user reported a minor discrepancy with test 2, which turned out
2239        to be caused by a trailing space on an input line that had got lost in his
2240        copy. The trailing space was an accident, so I've just removed it.
2241    
2242     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2243        that is needed.
2244    
2245     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2246        as "const" (a) because they are and (b) because it helps the PHP
2247        maintainers who have recently made a script to detect big data structures
2248        in the php code that should be moved to the .rodata section. I remembered
2249        to update Builducptable as well, so it won't revert if ucptable.h is ever
2250        re-created.
2251    
2252     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2253        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2254        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2255        support is not required. This saves 1.5-2K of code, which is important in
2256        some applications.
2257    
2258        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2259        so as not to refer to the tables, even though these functions will never be
2260        called when UTF-8 support is disabled. Otherwise there are problems with a
2261        shared library.
2262    
2263    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2264    
2265        (a) It was defining its arguments as char * instead of void *.
2266    
2267        (b) It was assuming that all moves were upwards in memory; this was true
2268            a long time ago when I wrote it, but is no longer the case.
2269    
2270        The emulated memove() is provided for those environments that have neither
2271        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2272        is clearly not the case, as these two bugs were recently reported.
2273    
2274    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2275        and Detrail to create the HTML documentation, the .txt form of the man
2276        pages, and it removes trailing spaces from listed files. It also creates
2277        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2278        case, it wraps all the #defines with #ifndefs. This script should be run
2279        before "make dist".
2280    
2281    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2282        with Unicode property support.
2283    
2284        (a) For a maximizing quantifier, if the two different cases of the
2285            character were of different lengths in their UTF-8 codings (there are
2286            some cases like this - I found 11), and the matching function had to
2287            back up over a mixture of the two cases, it incorrectly assumed they
2288            were both the same length.
2289    
2290        (b) When PCRE was configured to use the heap rather than the stack for
2291            recursion during matching, it was not correctly preserving the data for
2292            the other case of a UTF-8 character when checking ahead for a match
2293            while processing a minimizing repeat. If the check also involved
2294            matching a wide character, but failed, corruption could cause an
2295            erroneous result when trying to check for a repeat of the original
2296            character.
2297    
2298    13. Some tidying changes to the testing mechanism:
2299    
2300        (a) The RunTest script now detects the internal link size and whether there
2301            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2302            values substituted by "configure". (The RunGrepTest script already did
2303            this for UTF-8.) The configure.ac script no longer substitutes the
2304            relevant variables.
2305    
2306        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2307            with length and offset values. This means that the output is different
2308            for different internal link sizes. Test 2 is skipped for link sizes
2309            other than 2 because of this, bypassing the problem. Unfortunately,
2310            there was also a test in test 3 (the locale tests) that used /B and
2311            failed for link sizes other than 2. Rather than cut the whole test out,
2312            I have added a new /Z option to pcretest that replaces the length and
2313            offset values with spaces. This is now used to make test 3 independent
2314            of link size. (Test 2 will be tidied up later.)
2315    
2316    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2317        segmentation fault instead of returning the appropriate error message.
2318    
2319    15. In multiline mode when the newline sequence was set to "any", the pattern
2320        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2321        This doesn't seem right; it now treats the CRLF combination as the line
2322        ending, and so does not match in that case. It's only a pattern such as ^$
2323        that would hit this one: something like ^ABC$ would have failed after \r
2324        and then tried again after \r\n.
2325    
2326    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2327        in an attempt to make files that differ only in their line terminators
2328        compare equal. This works on Linux.
2329    
2330    17. Under certain error circumstances pcregrep might try to free random memory
2331        as it exited. This is now fixed, thanks to valgrind.
2332    
2333    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2334        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2335        was because its rules for how to advance for /g after matching an empty
2336        string at the end of a line did not allow for this case. They now check for
2337        it specially.
2338    
2339    20. pcretest is supposed to handle patterns and data of any length, by
2340        extending its buffers when necessary. It was getting this wrong when the
2341        buffer for a data line had to be extended.
2342    
2343    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2344        CRLF as a newline sequence.
2345    
2346    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2347        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2348        I have nevertheless tidied it up.
2349    
2350    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2351    
2352    24. Added a man page for pcre-config.
2353    
2354    
2355    Version 7.0 19-Dec-06
2356    ---------------------
2357    
2358     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2359        moving to gcc 4.1.1.
2360    
2361     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2362        sys/time.h, which is documented as needed for this function. It doesn't
2363        seem to matter on Linux, but it showed up on some releases of OS X.
2364    
2365     3. It seems that there are systems where bytes whose values are greater than
2366        127 match isprint() in the "C" locale. The "C" locale should be the
2367        default when a C program starts up. In most systems, only ASCII printing
2368        characters match isprint(). This difference caused the output from pcretest
2369        to vary, making some of the tests fail. I have changed pcretest so that:
2370    
2371        (a) When it is outputting text in the compiled version of a pattern, bytes
2372            other than 32-126 are always shown as hex escapes.
2373    
2374        (b) When it is outputting text that is a matched part of a subject string,
2375            it does the same, unless a different locale has been set for the match
2376            (using the /L modifier). In this case, it uses isprint() to decide.
2377    
2378     4. Fixed a major bug that caused incorrect computation of the amount of memory
2379        required for a compiled pattern when options that changed within the
2380        pattern affected the logic of the preliminary scan that determines the
2381        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2382        that the computed length was too small. The symptoms of this bug were
2383        either the PCRE error "internal error: code overflow" from pcre_compile(),
2384        or a glibc crash with a message such as "pcretest: free(): invalid next
2385        size (fast)". Examples of patterns that provoked this bug (shown in
2386        pcretest format) are:
2387    
2388          /(?-x: )/x
2389          /(?x)(?-x: \s*#\s*)/
2390          /((?i)[\x{c0}])/8
2391          /(?i:[\x{c0}])/8
2392    
2393        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2394        is now done differently.
2395    
2396     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2397        wrapper classes; (b) implement a new function in the C++ scanner that is
2398        more efficient than the old way of doing things because it avoids levels of
2399        recursion in the regex matching; (c) add a paragraph to the documentation
2400        for the FullMatch() function.
2401    
2402     6. The escape sequence \n was being treated as whatever was defined as
2403        "newline". Not only was this contrary to the documentation, which states
2404        that \n is character 10 (hex 0A), but it also went horribly wrong when
2405        "newline" was defined as CRLF. This has been fixed.
2406    
2407     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2408        was being set to -1 for the "end of line" case (supposedly a value that no
2409        character can have). Though this value is never used (the check for end of
2410        line is "zero bytes in current character"), it caused compiler complaints.
2411        I've changed it to 0xffffffff.
2412    
2413     8. In pcre_version.c, the version string was being built by a sequence of
2414        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2415        string (as it is for production releases) called a macro with an empty
2416        argument. The C standard says the result of this is undefined. The gcc
2417        compiler treats it as an empty string (which was what was wanted) but it is
2418        reported that Visual C gives an error. The source has been hacked around to
2419        avoid this problem.
2420    
2421     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2422        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2423        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2424        of them did).
2425    
2426    10. Originally, pcretest opened its input and output without "b"; then I was
2427        told that "b" was needed in some environments, so it was added for release
2428        5.0 to both the input and output. (It makes no difference on Unix-like
2429        systems.) Later I was told that it is wrong for the input on Windows. I've
2430        now abstracted the modes into two macros, to make it easier to fiddle with
2431        them, and removed "b" from the input mode under Windows.
2432    
2433    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2434    
2435    12. Added -help and --help to pcretest as an official way of being reminded
2436        of the options.
2437    
2438    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2439        and pcrecpp.cc because they annoy compilers at high warning levels.
2440    
2441    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2442    
2443    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2444        scripts are not C programs :-) and which was not noticed because it works
2445        on Linux.
2446    
2447    16. pcretest is supposed to handle any length of pattern and data line (as one
2448        line or as a continued sequence of lines) by extending its input buffer if
2449        necessary. This feature was broken for very long pattern lines, leading to
2450        a string of junk being passed to pcre_compile() if the pattern was longer
2451        than about 50K.
2452    
2453    17. I have done a major re-factoring of the way pcre_compile() computes the
2454        amount of memory needed for a compiled pattern. Previously, there was code
2455        that made a preliminary scan of the pattern in order to do this. That was
2456        OK when PCRE was new, but as the facilities have expanded, it has become
2457        harder and harder to keep it in step with the real compile phase, and there
2458        have been a number of bugs (see for example, 4 above). I have now found a
2459        cunning way of running the real compile function in a "fake" mode that
2460        enables it to compute how much memory it would need, while actually only
2461        ever using a few hundred bytes of working memory and without too many
2462        tests of the mode. This should make future maintenance and development
2463        easier. A side effect of this work is that the limit of 200 on the nesting
2464        depth of parentheses has been removed (though this was never a serious
2465        limitation, I suspect). However, there is a downside: pcre_compile() now
2466        runs more slowly than before (30% or more, depending on the pattern). I
2467        hope this isn't a big issue. There is no effect on runtime performance.
2468    
2469    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2470        newline (only possible for the last line of a file) and it was a
2471        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2472    
2473    19. Added additional timing features to pcretest. (1) The -tm option now times
2474        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2475        separate command line item, by a number that specifies the number of
2476        repeats to use when timing. The default is 50000; this gives better
2477        precision, but takes uncomfortably long for very large patterns.
2478    
2479    20. Extended pcre_study() to be more clever in cases where a branch of a
2480        subpattern has no definite first character. For example, (a*|b*)[cd] would
2481        previously give no result from pcre_study(). Now it recognizes that the
2482        first character must be a, b, c, or d.
2483    
2484    21. There was an incorrect error "recursive call could loop indefinitely" if
2485        a subpattern (or the entire pattern) that was being tested for matching an
2486        empty string contained only one non-empty item after a nested subpattern.
2487        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2488        incorrectly, because the \d was being skipped in the check.
2489    
2490    22. The pcretest program now has a new pattern option /B and a command line
2491        option -b, which is equivalent to adding /B to every pattern. This causes
2492        it to show the compiled bytecode, without the additional information that
2493        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2494        is the same as /B/I).
2495    
2496    23. A new optimization is now able automatically to treat some sequences such
2497        as a*b as a*+b. More specifically, if something simple (such as a character
2498        or a simple class like \d) has an unlimited quantifier, and is followed by
2499        something that cannot possibly match the quantified thing, the quantifier
2500        is automatically "possessified".
2501    
2502    24. A recursive reference to a subpattern whose number was greater than 39
2503        went wrong under certain circumstances in UTF-8 mode. This bug could also
2504        have affected the operation of pcre_study().
2505    
2506    25. Realized that a little bit of performance could be had by replacing
2507        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2508    
2509    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2510    
2511    27. Possessive quantifiers such as a++ were previously implemented by turning
2512        them into atomic groups such as ($>a+). Now they have their own opcodes,
2513        which improves performance. This includes the automatically created ones
2514        from 23 above.
2515    
2516    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2517        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2518        the first matched character to be a colon. This applied both to named and
2519        numbered groups.
2520    
2521    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2522    
2523    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2524        building PCRE on Windows easier, so I have included it in the distribution.
2525    
2526    31. There is now a check in pcretest against a ridiculously large number being
2527        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2528        loop, the loop is abandoned.
2529    
2530    32. Forward references to subpatterns in conditions such as (?(2)...) where
2531        subpattern 2 is defined later cause pcre_compile() to search forwards in
2532        the pattern for the relevant set of parentheses. This search went wrong
2533        when there were unescaped parentheses in a character class, parentheses
2534        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2535    
2536    33. "Subroutine" calls and backreferences were previously restricted to
2537        referencing subpatterns earlier in the regex. This restriction has now
2538        been removed.
2539    
2540    34. Added a number of extra features that are going to be in Perl 5.10. On the
2541        whole, these are just syntactic alternatives for features that PCRE had
2542        previously implemented using the Python syntax or my own invention. The
2543        other formats are all retained for compatibility.
2544    
2545        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2546            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2547            also .NET compatible.
2548    
2549        (b) A recursion or subroutine call to a named group can now be defined as
2550            (?&name) as well as (?P>name).
2551    
2552        (c) A backreference to a named group can now be defined as \k<name> or
2553            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2554            5.10, are also .NET compatible.
2555    
2556        (d) A conditional reference to a named group can now use the syntax
2557            (?(<name>) or (?('name') as well as (?(name).
2558    
2559        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2560            groups (named and numbered) that are never evaluated inline, but can be
2561            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2562            is always false. There may be only one alternative in such a group.
2563    
2564        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2565            as the simple (?(R). The condition is true only if the most recent
2566            recursion is that of the given number or name. It does not search out
2567            through the entire recursion stack.
2568    
2569        (g) The escape \gN or \g{N} has been added, where N is a positive or
2570            negative number, specifying an absolute or relative reference.
2571    
2572    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2573        some "unreachable code" warnings.
2574    
2575    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2576        things, this adds five new scripts.
2577    
2578    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2579        There were also incompatibilities regarding the handling of \Q..\E inside
2580        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2581        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2582    
2583    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2584        matches an empty string, and forcibly breaks the loop. There were bugs in
2585        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2586        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2587        separate and independent bugs (that affected different cases) have been
2588        fixed.
2589    
2590    39. Refactored the code to abolish the use of different opcodes for small
2591        capturing bracket numbers. This is a tidy that I avoided doing when I
2592        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2593        The new approach is not only tidier, it makes it possible to reduce the
2594        memory needed to fix the previous bug (38).
2595    
2596    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2597        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2598        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2599        mode.
2600    
2601    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2602        report.
2603    
2604    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2605        copy construction and assignment in the C++ wrapper.
2606    
2607    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2608        couple of bugs that could have given wrong results in the "--newline=crlf"
2609        case.
2610    
2611    44. Added a number of casts and did some reorganization of signed/unsigned int
2612        variables following suggestions from Dair Grant. Also renamed the variable
2613        "this" as "item" because it is a C++ keyword.
2614    
2615    45. Arranged for dftables to add
2616    
2617          #include "pcre_internal.h"
2618    
2619        to pcre_chartables.c because without it, gcc 4.x may remove the array
2620        definition from the final binary if PCRE is built into a static library and
2621        dead code stripping is activated.
2622    
2623    46. For an unanchored pattern, if a match attempt fails at the start of a
2624        newline sequence, and the newline setting is CRLF or ANY, and the next two
2625        characters are CRLF, advance by two characters instead of one.
2626    
2627    
2628    Version 6.7 04-Jul-06
2629    ---------------------
2630    
2631     1. In order to handle tests when input lines are enormously long, pcretest has
2632        been re-factored so that it automatically extends its buffers when
2633        necessary. The code is crude, but this _is_ just a test program. The
2634        default size has been increased from 32K to 50K.
2635    
2636     2. The code in pcre_study() was using the value of the re argument before
2637        testing it for NULL. (Of course, in any sensible call of the function, it
2638        won't be NULL.)
2639    
2640     3. The memmove() emulation function in pcre_internal.h, which is used on
2641        systems that lack both memmove() and bcopy() - that is, hardly ever -
2642        was missing a "static" storage class specifier.
2643    
2644     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2645        containing an extended class (one that cannot be represented by a bitmap
2646        because it contains high-valued characters or Unicode property items, e.g.
2647        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2648        pattern, but PCRE should not loop if you do not (it no longer does).
2649        [Detail: two cases were found: (a) a repeated subpattern containing an
2650        extended class; (b) a recursive reference to a subpattern that followed a
2651        previous extended class. It wasn't skipping over the extended class
2652        correctly when UTF-8 mode was not set.]
2653    
2654     5. A negated single-character class was not being recognized as fixed-length
2655        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2656        compile error "lookbehind assertion is not fixed length".
2657    
2658     6. The RunPerlTest auxiliary script was showing an unexpected difference
2659        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2660        write a Perl script that can interpret lines of an input file either as
2661        byte characters or as UTF-8, which is what "perltest" was being required to
2662        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2663        can't do is switch easily at run time between having the "use utf8;" pragma
2664        or not. In the end, I fudged it by using the RunPerlTest script to insert
2665        "use utf8;" explicitly for the UTF-8 tests.
2666    
2667     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2668        the end of the subject string, contrary to the documentation and to what
2669        Perl does. This was true of both matching functions. Now it matches only at
2670        the start of the subject and immediately after *internal* newlines.
2671    
2672     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2673        a pointer to an int instead of a pointer to an unsigned long int. This
2674        caused problems on 64-bit systems.
2675    
2676     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2677        instance of the 'standard' template library not being so standard".
2678    
2679    10. There was no check on the number of named subpatterns nor the maximum
2680        length of a subpattern name. The product of these values is used to compute
2681        the size of the memory block for a compiled pattern. By supplying a very
2682        long subpattern name and a large number of named subpatterns, the size
2683        computation could be caused to overflow. This is now prevented by limiting
2684        the length of names to 32 characters, and the number of named subpatterns
2685        to 10,000.
2686    
2687    11. Subpatterns that are repeated with specific counts have to be replicated in
2688        the compiled pattern. The size of memory for this was computed from the
2689        length of the subpattern and the repeat count. The latter is limited to
2690        65535, but there was no limit on the former, meaning that integer overflow
2691        could in principle occur. The compiled length of a repeated subpattern is
2692        now limited to 30,000 bytes in order to prevent this.
2693    
2694    12. Added the optional facility to have named substrings with the same name.
2695    
2696    13. Added the ability to use a named substring as a condition, using the
2697        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2698        are numbers (not recommended). Forward references are permitted.
2699    
2700    14. Added forward references in named backreferences (if you see what I mean).
2701    
2702    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2703        pattern could run off the end of the subject. For example, the pattern
2704        "(?s)(.{1,5})"8 did this with the subject "ab".
2705    
2706    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2707        PCRE_CASELESS was set when matching characters that were quantified with ?
2708        or *.
2709    
2710    17. A character class other than a single negated character that had a minimum
2711        but no maximum quantifier - for example [ab]{6,} - was not handled
2712        correctly by pce_dfa_exec(). It would match only one character.
2713    
2714    18. A valid (though odd) pattern that looked like a POSIX character
2715        class but used an invalid character after [ (for example [[,abc,]]) caused
2716        pcre_compile() to give the error "Failed: internal error: code overflow" or
2717        in some cases to crash with a glibc free() error. This could even happen if
2718        the pattern terminated after [[ but there just happened to be a sequence of
2719        letters, a binary zero, and a closing ] in the memory that followed.
2720    
2721    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2722        over the years. Originally (before any Unicode support), just the bottom 8
2723        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2724        output from "man perlunicode" includes this:
2725    
2726          The regular expression compiler produces polymorphic opcodes.  That
2727          is, the pattern adapts to the data and automatically switches to
2728          the Unicode character scheme when presented with Unicode data--or
2729          instead uses a traditional byte scheme when presented with byte
2730          data.
2731    
2732        Sadly, a wide octal escape does not cause a switch, and in a string with
2733        no other multibyte characters, these octal escapes are treated as before.
2734        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2735        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2736        Unicode string.
2737    
2738        I have not perpetrated such confusion in PCRE. Up till now, it took just
2739        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2740        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2741        translate to the appropriate multibyte character.
2742    
2743    29. Applied some refactoring to reduce the number of warnings from Microsoft
2744        and Borland compilers. This has included removing the fudge introduced
2745        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2746        a warning about an unused variable.
2747    
2748    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2749        characters since release 4.0, because Perl (from release 5.004) does not.
2750        [Or at least, is documented not to: some releases seem to be in conflict
2751        with the documentation.] However, when a pattern was studied with
2752        pcre_study() and all its branches started with \s, PCRE still included VT
2753        as a possible starting character. Of course, this did no harm; it just
2754        caused an unnecessary match attempt.
2755    
2756    22. Removed a now-redundant internal flag bit that recorded the fact that case
2757        dependency changed within the pattern. This was once needed for "required
2758        byte" processing, but is no longer used. This recovers a now-scarce options
2759        bit. Also moved the least significant internal flag bit to the most-
2760        significant bit of the word, which was not previously used (hangover from
2761        the days when it was an int rather than a uint) to free up another bit for
2762        the future.
2763    
2764    23. Added support for CRLF line endings as well as CR and LF. As well as the
2765        default being selectable at build time, it can now be changed at runtime
2766        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2767        specify that it is scanning data with non-default line endings.
2768    
2769    24. Changed the definition of CXXLINK to make it agree with the definition of
2770        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2771    
2772    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2773        recursions. This makes a big different to stack usage for some patterns.
2774    
2775    26. If a subpattern containing a named recursion or subroutine reference such
2776        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2777        the space required for the compiled pattern went wrong and gave too small a
2778        value. Depending on the environment, this could lead to "Failed: internal
2779        error: code overflow at offset 49" or "glibc detected double free or
2780        corruption" errors.
2781    
2782    27. Applied patches from Google (a) to support the new newline modes and (b) to
2783        advance over multibyte UTF-8 characters in GlobalReplace.
2784    
2785    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2786        difference for some implementation of PCRE in some Windows version.
2787    
2788    29. Added some extra testing facilities to pcretest:
2789    
2790        \q<number>   in a data line sets the "match limit" value
2791        \Q<number>   in a data line sets the "match recursion limt" value
2792        -S <number>  sets the stack size, where <number> is in megabytes
2793    
2794        The -S option isn't available for Windows.
2795    
2796    
2797    Version 6.6 06-Feb-06
2798    ---------------------
2799    
2800     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2801        in pcreposix.h. I have copied the definition from pcre.h.
2802    
2803     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2804        because pcre.h is no longer a built file.
2805    
2806     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2807        not normally included in the compiled code.
2808    
2809    
2810    Version 6.5 01-Feb-06
2811    ---------------------
2812    
2813     1. When using the partial match feature with pcre_dfa_exec(), it was not
2814        anchoring the second and subsequent partial matches at the new starting
2815        point. This could lead to incorrect results. For example, with the pattern
2816        /1234/, partially matching against "123" and then "a4" gave a match.
2817    
2818     2. Changes to pcregrep:
2819    
2820        (a) All non-match returns from pcre_exec() were being treated as failures
2821            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2822            error message is output. Some extra information is given for the
2823            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2824            probably the only errors that are likely to be caused by users (by
2825            specifying a regex that has nested indefinite repeats, for instance).
2826            If there are more than 20 of these errors, pcregrep is abandoned.
2827    
2828        (b) A binary zero was treated as data while matching, but terminated the
2829            output line if it was written out. This has been fixed: binary zeroes
2830            are now no different to any other data bytes.
2831    
2832        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2833            used to set a locale for matching. The --locale=xxxx long option has
2834            been added (no short equivalent) to specify a locale explicitly on the
2835            pcregrep command, overriding the environment variables.
2836    
2837        (d) When -B was used with -n, some line numbers in the output were one less
2838            than they should have been.
2839    
2840        (e) Added the -o (--only-matching) option.
2841    
2842        (f) If -A or -C was used with -c (count only), some lines of context were
2843            accidentally printed for the final match.
2844    
2845        (g) Added the -H (--with-filename) option.
2846    
2847        (h) The combination of options -rh failed to suppress file names for files
2848            that were found from directory arguments.
2849    
2850        (i) Added the -D (--devices) and -d (--directories) options.
2851    
2852        (j) Added the -F (--fixed-strings) option.
2853    
2854        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2855    
2856        (l) Added the --colo(u)r option.
2857    
2858        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2859            is not present by default.
2860    
2861     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2862        items such as (?R) or (?1), when the recursion could match a number of
2863        alternatives. If it matched one of the alternatives, but subsequently,
2864        outside the recursion, there was a failure, the code tried to back up into
2865        the recursion. However, because of the way PCRE is implemented, this is not
2866        possible, and the result was an incorrect result from the match.
2867    
2868        In order to prevent this happening, the specification of recursion has
2869        been changed so that all such subpatterns are automatically treated as
2870        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2871    
2872     4. I had overlooked the fact that, in some locales, there are characters for
2873        which isalpha() is true but neither isupper() nor islower() are true. In
2874        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2875        and ordfeminine) are like this. This affected the treatment of \w and \W
2876        when they appeared in character classes, but not when they appeared outside
2877        a character class. The bit map for "word" characters is now created
2878        separately from the results of isalnum() instead of just taking it from the
2879        upper, lower, and digit maps. (Plus the underscore character, of course.)
2880    
2881     5. The above bug also affected the handling of POSIX character classes such as
2882        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2883        permanent tables. Instead, the bit maps for such a class were previously
2884        created as the appropriate unions of the upper, lower, and digit bitmaps.
2885        Now they are created by subtraction from the [[:word:]] class, which has
2886        its own bitmap.
2887    
2888     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2889        It is created by subtracting the vertical space characters (\x09, \x0a,
2890        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2891        subtraction was done in the overall bitmap for a character class, meaning
2892        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2893        be recognized. This bug has been fixed.
2894    
2895     7. Patches from the folks at Google:
2896    
2897          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2898          real life, but is still worth protecting against".
2899    
2900          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2901          regular expressions".
2902    
2903          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2904          have it.
2905    
2906          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2907          "configure" and the latter not, in order to fix a problem somebody had
2908          with compiling the Arg class on HP-UX.
2909    
2910          (e) Improve the error-handling of the C++ wrapper a little bit.
2911    
2912          (f) New tests for checking recursion limiting.
2913    
2914     8. The pcre_memmove() function, which is used only if the environment does not
2915        have a standard memmove() function (and is therefore rarely compiled),
2916        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2917        returning a result (though PCRE never actually uses the result).
2918    
2919     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2920        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2921        returned instead of calling malloc() with an overflowing number that would
2922        most likely cause subsequent chaos.
2923    
2924    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2925    
2926    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2927        with this option is matched, the nmatch and pmatch options of regexec() are
2928        ignored.
2929    
2930    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2931        provided in case anyone wants to the the POSIX interface with UTF-8
2932        strings.
2933    
2934    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2935        C++ linking (needed for some HP-UX environments).
2936    
2937    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2938        (unused parameter) and in the pcre_printint() function (omitted "default"
2939        switch label when the default is to do nothing).
2940    
2941    15. Added some code to make it possible, when PCRE is compiled as a C++
2942        library, to replace subject pointers for pcre_exec() with a smart pointer
2943        class, thus making it possible to process discontinuous strings.
2944    
2945    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2946        much the same function. They were added by different people who were trying
2947        to make PCRE easy to compile on non-Unix systems. It has been suggested
2948        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2949        for compiling on Windows systems. I have therefore replaced it with
2950        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2951        defaults to "extern" for C or "extern C" for C++, which works fine on
2952        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2953        SCOPE with something explicit in config.h. In addition:
2954    
2955        (a) pcreposix.h still had just "extern" instead of either of these macros;
2956            I have replaced it with PCRE_DATA_SCOPE.
2957    
2958        (b) Functions such as _pcre_xclass(), which are internal to the library,
2959            but external in the C sense, all had PCRE_EXPORT in their definitions.
2960            This is apparently wrong for the Windows case, so I have removed it.
2961            (It makes no difference on Unix-like systems.)
2962    
2963    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2964        of recursive calls to match(). This is different to MATCH_LIMIT because
2965        that limits the total number of calls to match(), not all of which increase
2966        the depth of recursion. Limiting the recursion depth limits the amount of
2967        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2968        when PCRE is compiled, and changed at run time. A patch from Google adds
2969        this functionality to the C++ interface.
2970    
2971    18. Changes to the handling of Unicode character properties:
2972    
2973        (a) Updated the table to Unicode 4.1.0.
2974    
2975        (b) Recognize characters that are not in the table as "Cn" (undefined).
2976    
2977        (c) I revised the way the table is implemented to a much improved format
2978            which includes recognition of ranges. It now supports the ranges that
2979            are defined in UnicodeData.txt, and it also amalgamates other
2980            characters into ranges. This has reduced the number of entries in the
2981            table from around 16,000 to around 3,000, thus reducing its size
2982            considerably. I realized I did not need to use a tree structure after
2983            all - a binary chop search is just as efficient. Having reduced the
2984            number of entries, I extended their size from 6 bytes to 8 bytes to
2985            allow for more data.
2986    
2987        (d) Added support for Unicode script names via properties such as \p{Han}.
2988    
2989    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2990        matching that character.
2991    
2992    20. When matching a repeated Unicode property with a minimum greater than zero,
2993        (for example \pL{2,}), PCRE could look past the end of the subject if it
2994        reached it while seeking the minimum number of characters. This could
2995        happen only if some of the characters were more than one byte long, because
2996        there is a check for at least the minimum number of bytes.
2997    
2998    21. Refactored the implementation of \p and \P so as to be more general, to
2999        allow for more different types of property in future. This has changed the
3000        compiled form incompatibly. Anybody with saved compiled patterns that use
3001        \p or \P will have to recompile them.
3002    
3003    22. Added "Any" and "L&" to the supported property types.
3004    
3005    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3006        but give a compile time error if the value is greater than 0xff.
3007    
3008    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3009        accidentally not being installed or uninstalled.
3010    
3011    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3012        made were to insert the current release number. This seemed silly, because
3013        it made things harder for people building PCRE on systems that don't run
3014        "configure". I have turned pcre.h into a distributed file, no longer built
3015        by "configure", with the version identification directly included. There is
3016        no longer a pcre.h.in file.
3017    
3018        However, this change necessitated a change to the pcre-config script as
3019        well. It is built from pcre-config.in, and one of the substitutions was the
3020        release number. I have updated configure.ac so that ./configure now finds
3021        the release number by grepping pcre.h.
3022    
3023    26. Added the ability to run the tests under valgrind.
3024    
3025    
3026    Version 6.4 05-Sep-05
3027    ---------------------
3028    
3029     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3030        "--" to be printed when multiple files were scanned, even when none of the
3031        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3032        consider it to be a bug, and have restored the previous behaviour.
3033    
3034     2. A couple of code tidies to get rid of compiler warnings.
3035    
3036     3. The pcretest program used to cheat by referring to symbols in the library
3037        whose names begin with _pcre_. These are internal symbols that are not
3038        really supposed to be visible externally, and in some environments it is
3039        possible to suppress them. The cheating is now confined to including
3040        certain files from the library's source, which is a bit cleaner.
3041    
3042     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3043        file's purpose clearer.
3044    
3045     5. Reorganized pcre_ucp_findchar().
3046    
3047    
3048    Version 6.3 15-Aug-05
3049    ---------------------
3050    
3051     1. The file libpcre.pc.in did not have general read permission in the tarball.
3052    
3053     2. There were some problems when building without C++ support:
3054    
3055        (a) If C++ support was not built, "make install" and "make test" still
3056            tried to test it.
3057    
3058        (b) There were problems when the value of CXX was explicitly set. Some
3059            changes have been made to try to fix these, and ...
3060    
3061        (c) --disable-cpp can now be used to explicitly disable C++ support.
3062    
3063        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3064            backslash in a target when C++ was disabled. This confuses some
3065            versions of "make", apparently. Using an intermediate variable solves
3066            this. (Same for CPP_LOBJ.)
3067    
3068     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3069        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3070        necessary on certain architectures.
3071    
3072     4. Added a setting of -export-symbols-regex to the link command to remove
3073        those symbols that are exported in the C sense, but actually are local
3074        within the library, and not documented. Their names all begin with
3075        "_pcre_". This is not a perfect job, because (a) we have to except some
3076        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3077        available (and never for static libraries). I have made a note to try to
3078        find a way round (a) in the future.
3079    
3080    
3081    Version 6.2 01-Aug-05
3082    ---------------------
3083    
3084     1. There was no test for integer overflow of quantifier values. A construction
3085        such as {1111111111111111} would give undefined results. What is worse, if
3086        a minimum quantifier for a parenthesized subpattern overflowed and became
3087        negative, the calculation of the memory size went wrong. This could have
3088        led to memory overwriting.
3089    
3090     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3091    
3092     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3093        operating environments where this matters.
3094    
3095     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3096        PCRE options from within the C++ wrapper.
3097    
3098     5. Named capturing subpatterns were not being correctly counted when a pattern
3099        was compiled. This caused two problems: (a) If there were more than 100
3100        such subpatterns, the calculation of the memory needed for the whole
3101        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3102        back references of the form \12, where the number was greater than 9, were
3103        not recognized as back references, even though there were sufficient
3104        previous subpatterns.
3105    
3106     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3107        versions of gcc, e.g. 2.95.4.
3108    
3109    
3110    Version 6.1 21-Jun-05
3111    ---------------------
3112    
3113     1. There was one reference to the variable "posix" in pcretest.c that was not
3114        surrounded by "#if !defined NOPOSIX".
3115    
3116     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3117        the cross-check on the old pcre_info() function, for the benefit of the
3118        cut-down version of PCRE that is currently imported into Exim.
3119    
3120     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3121        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3122        patterns that start (?i) but I don't think that matters. The use of (?i) is
3123        just an example; this all applies to the other options as well.
3124    
3125     4. Since libtool seems to echo the compile commands it is issuing, the output
3126        from "make" can be reduced a bit by putting "@" in front of each libtool
3127        compile command.
3128    
3129     5. Patch from the folks at Google for configure.in to be a bit more thorough
3130        in checking for a suitable C++ installation before trying to compile the
3131        C++ stuff. This should fix a reported problem when a compiler was present,
3132        but no suitable headers.
3133    
3134     6. The man pages all had just "PCRE" as their title. I have changed them to
3135        be the relevant file name. I have also arranged that these names are
3136        retained in the file doc/pcre.txt, which is a concatenation in text format
3137        of all the man pages except the little individual ones for each function.
3138    
3139     7. The NON-UNIX-USE file had not been updated for the different set of source
3140        files that come with release 6. I also added a few comments about the C++
3141        wrapper.
3142    
3143    
3144    Version 6.0 07-Jun-05
3145    ---------------------
3146    
3147     1. Some minor internal re-organization to help with my DFA experiments.
3148    
3149     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3150        didn't matter for the library itself when fully configured, but did matter
3151        when compiling without UCP support, or within Exim, where the ucp files are
3152        not imported.
3153    
3154     3. Refactoring of the library code to split up the various functions into
3155        different source modules. The addition of the new DFA matching code (see
3156        below) to a single monolithic source would have made it really too
3157        unwieldy, quite apart from causing all the code to be include in a
3158        statically linked application, when only some functions are used. This is
3159        relevant even without the DFA addition now that patterns can be compiled in
3160        one application and matched in another.
3161    
3162        The downside of splitting up is that there have to be some external
3163        functions and data tables that are used internally in different modules of
3164        the library but which are not part of the API. These have all had their
3165        names changed to start with "_pcre_" so that they are unlikely to clash
3166        with other external names.
3167    
3168     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3169        a different (DFA) algorithm. Although it is slower than the original
3170        function, it does have some advantages for certain types of matching
3171        problem.
3172    
3173     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3174        including restarting after a partial match.
3175    
3176     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3177        defined when compiling for Windows was sent to me. I have put it into the
3178        code, though I have no means of testing or verifying it.
3179    
3180     7. Added the pcre_refcount() auxiliary function.
3181    
3182     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3183        match before or at the first newline in the subject string. In pcretest,
3184        the /f option on a pattern can be used to set this.
3185    
3186     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3187        would behave wrongly. This has been present in PCRE since release 4.0.
3188    
3189    10. A number of changes to the pcregrep command:
3190    
3191        (a) Refactored how -x works; insert ^(...)$ instead of setting
3192            PCRE_ANCHORED and checking the length, in preparation for adding
3193            something similar for -w.
3194    
3195        (b) Added the -w (match as a word) option.
3196    
3197        (c) Refactored the way lines are read and buffered so as to have more
3198            than one at a time available.
3199    
3200        (d) Implemented a pcregrep test script.
3201    
3202        (e) Added the -M (multiline match) option. This allows patterns to match
3203            over several lines of the subject. The buffering ensures that at least
3204            8K, or the rest of the document (whichever is the shorter) is available
3205            for matching (and similarly the previous 8K for lookbehind assertions).
3206    
3207        (f) Changed the --help output so that it now says
3208    
3209              -w, --word-regex(p)
3210    
3211            instead of two lines, one with "regex" and the other with "regexp"
3212            because that confused at least one person since the short forms are the
3213            same. (This required a bit of code, as the output is generated
3214            automatically from a table. It wasn't just a text change.)
3215    
3216        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3217            option but starts with a hyphen. Could be a pattern or a path name
3218            starting with a hyphen, for instance.
3219    
3220        (h) "-" can be given as a file name to represent stdin.
3221    
3222        (i) When file names are being printed, "(standard input)" is used for
3223            the standard input, for compatibility with GNU grep. Previously
3224            "<stdin>" was used.
3225    
3226        (j) The option --label=xxx can be used to supply a name to be used for
3227            stdin when file names are being printed. There is no short form.
3228    
3229        (k) Re-factored the options decoding logic because we are going to add
3230            two more options that take data. Such options can now be given in four
3231            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3232    
3233        (l) Added the -A, -B, and -C options for requesting that lines of context
3234            around matches be printed.
3235    
3236        (m) Added the -L option to print the names of files that do not contain
3237            any matching lines, that is, the complement of -l.
3238    
3239        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3240            continue to scan other files.
3241    
3242        (o) The -s option was incorrectly implemented. For compatibility with other
3243            greps, it now suppresses the error message for a non-existent or non-
3244            accessible file (but not the return code). There is a new option called
3245            -q that suppresses the output of matching lines, which was what -s was
3246            previously doing.
3247    
3248        (p) Added --include and --exclude options to specify files for inclusion
3249            and exclusion when recursing.
3250    
3251    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3252        Hopefully, it now does.
3253    
3254    12. Missing cast in pcre_study().
3255    
3256    13. Added an "uninstall" target to the makefile.
3257    
3258    14. Replaced "extern" in the function prototypes in Makefile.in with
3259        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3260        world, but is set differently for Windows.
3261    
3262    15. Added a second compiling function called pcre_compile2(). The only
3263        difference is that it has an extra argument, which is a pointer to an
3264        integer error code. When there is a compile-time failure, this is set
3265        non-zero, in addition to the error test pointer being set to point to an
3266        error message. The new argument may be NULL if no error number is required
3267        (but then you may as well call pcre_compile(), which is now just a
3268        wrapper). This facility is provided because some applications need a
3269        numeric error indication, but it has also enabled me to tidy up the way
3270        compile-time errors are handled in the POSIX wrapper.
3271    
3272    16. Added VPATH=.libs to the makefile; this should help when building with one
3273        prefix path and installing with another. (Or so I'm told by someone who
3274        knows more about this stuff than I do.)
3275    
3276    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3277        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3278        match everything, including newlines. This is not POSIX-compatible, but
3279        somebody wanted the feature. From pcretest it can be activated by using
3280        both the P and the s flags.
3281    
3282    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3283    
3284    19. libpcre.pc was being incorrectly installed as executable.
3285    
3286    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3287        it now also looks for '\r' so that it will work unmodified on Windows.
3288    
3289    21. Added Google's contributed C++ wrapper to the distribution.
3290    
3291    22. Added some untidy missing memory free() calls in pcretest, to keep
3292        Electric Fence happy when testing.
3293    
3294    
3295    
3296    Version 5.0 13-Sep-04
3297    ---------------------
3298    
3299     1. Internal change: literal characters are no longer packed up into items
3300        containing multiple characters in a single byte-string. Each character
3301        is now matched using a separate opcode. However, there may be more than one
3302        byte in the character in UTF-8 mode.
3303    
3304     2. The pcre_callout_block structure has two new fields: pattern_position and
3305        next_item_length. These contain the offset in the pattern to the next match
3306        item, and its length, respectively.
3307    
3308     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3309        insertion of callouts before each pattern item. Added the /C option to
3310        pcretest to make use of this.
3311    
3312     4. On the advice of a Windows user, the lines
3313    
3314          #if defined(_WIN32) || defined(WIN32)
3315          _setmode( _fileno( stdout ), 0x8000 );
3316          #endif  /* defined(_WIN32) || defined(WIN32) */
3317    
3318        have been added to the source of pcretest. This apparently does useful
3319        magic in relation to line terminators.
3320    
3321     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3322        for the benefit of those environments where the "b" makes a difference.
3323    
3324     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3325        to know about it. I have put a hack into configure.in that adds in code
3326        to set GCC=yes if CC=icc. This seems to end up at a point in the
3327        generated configure script that is early enough to affect the setting of
3328        compiler options, which is what is needed, but I have no means of testing
3329        whether it really works. (The user who reported this had patched the
3330        generated configure script, which of course I cannot do.)
3331    
3332        LATER: After change 22 below (new libtool files), the configure script
3333        seems to know about icc (and also ecc). Therefore, I have commented out
3334        this hack in configure.in.
3335    
3336     7. Added support for pkg-config (2 patches were sent in).
3337    
3338     8. Negated POSIX character classes that used a combination of internal tables
3339        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3340        [[:^ascii]]. Typically, they would match almost any characters. The other
3341        POSIX classes were not broken in this way.
3342    
3343     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3344        to find the match, as PCRE was deluded into thinking that the match had to
3345        start at the start point or following a newline. The same bug applied to
3346        patterns with negative forward assertions or any backward assertions
3347        preceding ".*" at the start, unless the pattern required a fixed first
3348        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3349    
3350    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3351        starting at the last subject character, bytes beyond the end of the subject
3352        string were read.
3353    
3354    11. Renamed the variable "class" as "classbits" to make life easier for C++
3355        users. (Previously there was a macro definition, but it apparently wasn't
3356        enough.)
3357    
3358    12. Added the new field "tables" to the extra data so that tables can be passed
3359        in at exec time, or the internal tables can be re-selected. This allows
3360        a compiled regex to be saved and re-used at a later time by a different
3361        program that might have everything at different addresses.
3362    
3363    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3364        -R library as well as a -L library.
3365    
3366    14. The debugging options of pcretest (-d on the command line or D on a
3367        pattern) showed incorrect output for anything following an extended class
3368        that contained multibyte characters and which was followed by a quantifier.
3369    
3370    15. Added optional support for general category Unicode character properties
3371        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3372        support. It adds about 90K to the size of the library. The meanings of the
3373        inbuilt class escapes such as \d and \s have NOT been changed.
3374    
3375    16. Updated pcredemo.c to include calls to free() to release the memory for the
3376        compiled pattern.
3377    
3378    17. The generated file chartables.c was being created in the source directory
3379        instead of in the building directory. This caused the build to fail if the
3380        source directory was different from the building directory, and was
3381        read-only.
3382    
3383    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3384        file. No doubt somebody will tell me if they don't make sense... Also added
3385        Dan Mooney's comments about building on OpenVMS.
3386    
3387    19. Added support for partial matching via the PCRE_PARTIAL option for
3388        pcre_exec() and the \P data escape in pcretest.
3389    
3390    20. Extended pcretest with 3 new pattern features:
3391    
3392        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3393              write the compiled pattern to the file whose name is "rest-of-line".
3394              This is a straight binary dump of the data, with the saved pointer to
3395              the character tables forced to be NULL. The study data, if any, is
3396              written too. After writing, pcretest reads a new pattern.
3397    
3398        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3399              compiled pattern from the given file. There must not be any
3400              occurrences of "<" in the file name (pretty unlikely); if there are,
3401              pcretest will instead treat the initial "<" as a pattern delimiter.
3402              After reading in the pattern, pcretest goes on to read data lines as
3403              usual.
3404    
3405        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3406              and 16-bit fields in a compiled pattern, to simulate a pattern that
3407              was compiled on a host of opposite endianness.
3408    
3409    21. The pcre-exec() function can now cope with patterns that were compiled on
3410        hosts of opposite endianness, with this restriction:
3411    
3412          As for any compiled expression that is saved and used later, the tables
3413          pointer field cannot be preserved; the extra_data field in the arguments
3414          to pcre_exec() should be used to pass in a tables address if a value
3415          other than the default internal tables were used at compile time.
3416    
3417    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3418        now diagnosed as an error. Previously, most of the time, a negative number
3419        would have been treated as zero, but if in addition "ovector" was passed as
3420        NULL, a crash could occur.
3421    
3422    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3423        new versions from the libtool 1.5 distribution (the last one is a copy of
3424        a file called libtool.m4). This seems to have fixed the need to patch
3425        "configure" to support Darwin 1.3 (which I used to do). However, I still
3426        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3427        workstation).
3428    
3429    24. Changed the PCRE licence to be the more standard "BSD" licence.
3430    
3431    
3432    Version 4.5 01-Dec-03
3433    ---------------------
3434    
3435     1. There has been some re-arrangement of the code for the match() function so
3436        that it can be compiled in a version that does not call itself recursively.
3437        Instead, it keeps those local variables that need separate instances for
3438        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3439        needs to "recurse". Keeping track of where control must go is done by means
3440        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3441        hide most of the details from the main code, and operates only if
3442        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3443        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3444        operating.
3445    
3446        To make it easier for callers to provide specially tailored get/free
3447        functions for this usage, two new functions, pcre_stack_malloc, and
3448        pcre_stack_free, are used. They are always called in strict stacking order,
3449        and the size of block requested is always the same.
3450    
3451        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3452        PCRE has been compiled to use the stack or the heap for recursion. The
3453        -C option of pcretest uses this to show which version is compiled.
3454    
3455        A new data escape \S, is added to pcretest; it causes the amounts of store
3456        obtained and freed by both kinds of malloc/free at match time to be added
3457        to the output.
3458    
3459     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3460        what's available on my current Linux desktop machine.
3461    
3462     3. When matching a UTF-8 string, the test for a valid string at the start has
3463        been extended. If start_offset is not zero, PCRE now checks that it points
3464        to a byte that is the start of a UTF-8 character. If not, it returns
3465        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3466        this is necessary because there may be backward assertions in the pattern.
3467        When matching the same subject several times, it may save resources to use
3468        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3469    
3470     4. The code for checking the validity of UTF-8 strings has been tightened so
3471        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3472        containing "overlong sequences".
3473    
3474     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3475        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3476        should have been "&", but it just so happened that all the cases this let
3477        through by mistake were picked up later in the function.
3478    
3479     6. I had used a variable called "isblank" - this is a C99 function, causing
3480        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3481    
3482     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3483        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3484        so the version is shown; (c) stop "make test" echoing "./RunTest".
3485    
3486     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3487    
3488     9. The prototype for memmove() for systems that don't have it was using
3489        size_t, but the inclusion of the header that defines size_t was later. I've
3490        moved the #includes for the C headers earlier to avoid this.
3491    
3492    10. Added some adjustments to the code to make it easier to compiler on certain
3493        special systems:
3494    
3495          (a) Some "const" qualifiers were missing.
3496          (b) Added the macro EXPORT before all exported functions; by default this
3497              is defined to be empty.
3498          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3499              that it reads its output file name as an argument instead of writing
3500              to the standard output and assuming this can be redirected.
3501    
3502    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3503        class containing characters with values greater than 255, PCRE compilation
3504        went into a loop.
3505    
3506    12. A recursive reference to a subpattern that was within another subpattern
3507        that had a minimum quantifier of zero caused PCRE to crash. For example,
3508        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3509        recursion. If the recursively-called subpattern itself had a zero repeat,
3510        that was OK.
3511    
3512    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3513        buffer into which it was copied (for escape processing) was still set at
3514        1024, so long lines caused crashes.
3515    
3516    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3517        "internal error: code overflow...". This applied to any character class
3518        that was followed by a possessive quantifier.
3519    
3520    15. Modified the Makefile to add libpcre.la as a prerequisite for
3521        libpcreposix.la because I was told this is needed for a parallel build to
3522        work.
3523    
3524    16. If a pattern that contained .* following optional items at the start was
3525        studied, the wrong optimizing data was generated, leading to matching
3526        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3527        matching string must start with a or b or c. The correct conclusion for
3528        this pattern is that a match can start with any character.
3529    
3530    
3531    Version 4.4 13-Aug-03
3532    ---------------------
3533    
3534     1. In UTF-8 mode, a character class containing characters with values between
3535        127 and 255 was not handled correctly if the compiled pattern was studied.
3536        In fixing this, I have also improved the studying algorithm for such
3537        classes (slightly).
3538    
3539     2. Three internal functions had redundant arguments passed to them. Removal
3540        might give a very teeny performance improvement.
3541    
3542     3. Documentation bug: the value of the capture_top field in a callout is *one
3543        more than* the number of the hightest numbered captured substring.
3544    
3545     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3546        in incorrectly linking with a previously installed version. They now link
3547        explicitly with libpcre.la.
3548    
3549     5. configure.in no longer needs to recognize Cygwin specially.
3550    
3551     6. A problem in pcre.in for Windows platforms is fixed.
3552    
3553     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3554        pcretest, it used to include the size of the study block as part of its
3555        output. Unfortunately, the structure contains a field that has a different
3556        size on different hardware architectures. This meant that the tests that
3557        showed this size failed. As the block is currently always of a fixed size,
3558        this information isn't actually particularly useful in pcretest output, so
3559        I have just removed it.
3560    
3561     8. Three pre-processor statements accidentally did not start in column 1.
3562        Sadly, there are *still* compilers around that complain, even though
3563        standard C has not required this for well over a decade. Sigh.
3564    
3565     9. In pcretest, the code for checking callouts passed small integers in the
3566        callout_data field, which is a void * field. However, some picky compilers
3567        complained about the casts involved for this on 64-bit systems. Now
3568        pcretest passes the address of the small integer instead, which should get
3569        rid of the warnings.
3570    
3571    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3572        both compile and run time, and gives an error if an invalid UTF-8 sequence
3573        is found. There is a option for disabling this check in cases where the
3574        string is known to be correct and/or the maximum performance is wanted.
3575    
3576    11. In response to a bug report, I changed one line in Makefile.in from
3577    
3578            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3579        to
3580            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3581    
3582        to look similar to other lines, but I have no way of telling whether this
3583        is the right thing to do, as I do not use Windows. No doubt I'll get told
3584        if it's wrong...
3585    
3586    
3587    Version 4.3 21-May-03
3588    ---------------------
3589    
3590    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3591       Makefile.
3592    
3593    2. Some refactoring to improve the quality of the code:
3594    
3595       (i)   The utf8_table... variables are now declared "const".
3596    
3597       (ii)  The code for \cx, which used the "case flipping" table to upper case
3598             lower case letters, now just substracts 32. This is ASCII-specific,
3599             but the whole concept of \cx is ASCII-specific, so it seems
3600             reasonable.
3601    
3602       (iii) PCRE was using its character types table to recognize decimal and
3603             hexadecimal digits in the pattern. This is silly, because it handles
3604             only 0-9, a-f, and A-F, but the character types table is locale-
3605             specific, which means strange things might happen. A private
3606             table is now used for this - though it costs 256 bytes, a table is
3607             much faster than multiple explicit tests. Of course, the standard
3608             character types table is still used for matching digits in subject
3609             strings against \d.
3610    
3611       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3612             ending in _t are). So I've renamed it as ESC_tee.
3613    
3614    3. The first argument for regexec() in the POSIX wrapper should have been
3615       defined as "const".
3616    
3617    4. Changed pcretest to use malloc() for its buffers so that they can be
3618       Electric Fenced for debugging.
3619    
3620    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3621       to read one or more bytes before the start of the subject string. Often this
3622       had no effect on PCRE's behaviour, but in some circumstances it could
3623       provoke a segmentation fault.
3624    
3625    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3626       to try to read one or more bytes before the start of the subject string.
3627    
3628    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3629       UTF-8 support could misbehave in various ways if the subject string
3630       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3631       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3632       back over UTF-8 characters.)
3633    
3634    
3635    Version 4.2 14-Apr-03
3636    ---------------------
3637    
3638    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3639    
3640    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3641         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3642         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3643         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3644         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3645           and BUILD_EXEEXT
3646         Note: automatic setting of the BUILD variables is not yet working
3647         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3648           compile-time but not at link-time
3649         [LINK]: use for linking executables only
3650         make different versions for Windows and non-Windows
3651         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3652           libraries
3653         [LINK_FOR_BUILD]: new variable
3654         [OBJEXT]: use throughout
3655         [EXEEXT]: use throughout
3656         <winshared>: new target
3657         <wininstall>: new target
3658         <dftables.o>: use native compiler
3659         <dftables>: use native linker
3660         <install>: handle Windows platform correctly
3661         <clean>: ditto
3662         <check>: ditto
3663         copy DLL to top builddir before testing
3664    
3665       As part of these changes, -no-undefined was removed again. This was reported
3666       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3667       in any case.
3668    
3669    3. Some tidies to get rid of compiler warnings:
3670    
3671       . In the match_data structure, match_limit was an unsigned long int, whereas
3672         match_call_count was an int. I've made them both unsigned long ints.
3673    
3674       . In pcretest the fact that a const uschar * doesn't automatically cast to
3675         a void * provoked a warning.
3676    
3677       . Turning on some more compiler warnings threw up some "shadow" variables
3678         and a few more missing casts.
3679    
3680    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3681       option, a class that contained a single character with a value between 128
3682       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3683    
3684    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3685       option, a class that contained several characters, but with at least one
3686       whose value was between 128 and 255 caused PCRE to crash.
3687    
3688    
3689    Version 4.1 12-Mar-03
3690    ---------------------
3691    
3692    1. Compiling with gcc -pedantic found a couple of places where casts were
3693    needed, and a string in dftables.c that was longer than standard compilers are
3694    required to support.
3695    
3696    2. Compiling with Sun's compiler found a few more places where the code could
3697    be tidied up in order to avoid warnings.
3698    
3699    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3700    first of these names is deprecated in the latest Autoconf in favour of the name
3701    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3702    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3703    analogy I have changed it to CFLAGS_FOR_BUILD.
3704    
3705    4. Added -no-undefined to the linking command in the Makefile, because this is
3706    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3707    linking step for the pcreposix library.
3708    
3709    5. PCRE was failing to diagnose the case of two named groups with the same
3710    name.
3711    
3712    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3713    literal character that is needed in the subject for a match, and scans along to
3714    ensure that it is present before embarking on the full matching process. This
3715    saves time in cases of nested unlimited repeats that are never going to match.
3716    Problem: the scan can take a lot of time if the subject is very long (e.g.
3717    megabytes), thus penalizing straightforward matches. It is now done only if the
3718    amount of subject to be scanned is less than 1000 bytes.
3719    
3720    7. A lesser problem with the same optimization is that it was recording the
3721    first character of an anchored pattern as "needed", thus provoking a search
3722    right along the subject, even when the first match of the pattern was going to
3723    fail. The "needed" character is now not set for anchored patterns, unless it
3724    follows something in the pattern that is of non-fixed length. Thus, it still
3725    fulfils its original purpose of finding quick non-matches in cases of nested
3726    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3727    
3728    
3729    Version 4.0 17-Feb-03
3730    ---------------------
3731    
3732    1. If a comment in an extended regex that started immediately after a meta-item
3733    extended to the end of string, PCRE compiled incorrect data. This could lead to
3734    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3735    
3736    2. Moved to autoconf 2.53 and libtool 1.4.2.
3737    
3738    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3739    the special perltest8 script is no longer needed - all the tests can be run
3740    from a single perltest script.
3741    
3742    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3743    by \s. It has now been removed in PCRE. This means it isn't recognized as
3744    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3745    class [:space:] *does* include VT, thereby creating a mess.
3746    
3747    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3748    space and tab.
3749    
3750    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3751    its new features into the main test script, reducing the number of scripts.
3752    
3753    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3754    were backward compatible, and made the (?i) apply to the whole pattern, as if
3755    /i were given. Now it behaves more logically, and applies the option setting
3756    only to what follows. PCRE has been changed to follow suit. However, if it
3757    finds options settings right at the start of the pattern, it extracts them into
3758    the global options, as before. Thus, they show up in the info data.
3759    
3760    8. Added support for the \Q...\E escape sequence. Characters in between are
3761    treated as literals. This is slightly different from Perl in that $ and @ are
3762    also handled as literals inside the quotes. In Perl, they will cause variable
3763    interpolation. Note the following examples:
3764    
3765        Pattern            PCRE matches      Perl matches
3766    
3767        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3768        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3769        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3770    
3771    For compatibility with Perl, \Q...\E sequences are recognized inside character
3772    classes as well as outside them.
3773    
3774    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3775    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3776    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3777    signed/unsigned warnings.
3778    
3779    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3780    option for pcretest, so I've replaced it by a simple function that does just
3781    that job.
3782    
3783    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3784    "pcregrep -".
3785    
3786    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3787    Java package. This provides some syntactic sugar for simple cases of what my
3788    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3789