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

Diff of /code/trunk/ChangeLog

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

revision 131 by ph10, Mon Mar 26 15:10:12 2007 UTC revision 1239 by zherczeg, Fri Jan 18 08:20:44 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 12-Mar-07  Version 8.33 xx-xxxx-201x
5    -------------------------
6    
7    1.  Added 'U' to some constants that are compared to unsigned integers, to
8        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9        variables that are added to signed variables, to ensure the result is
10        signed and can be negated.
11    
12    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
13        CMake config files.
14    
15    3.  Revise the creation of config.h.generic so that all boolean macros are
16        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17        overriding via -D on the command line possible.
18    
19    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20        to unsigned int is reported to make a quite noticeable speed difference in
21        a specific Windows environment. Testing on Linux did also appear to show
22        some benefit (and it is clearly not harmful). Also fixed the definition of
23        Xop which should be unsigned.
24    
25    5.  Related to (4), changing the definition of the intermediate variable cc
26        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27        improvements.
28    
29    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
30        unnecessary spaces.
31    
32    
33    Version 8.32 30-November-2012
34    -----------------------------
35    
36    1.  Improved JIT compiler optimizations for first character search and single
37        character iterators.
38    
39    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
40        Patch by Daniel Richard G.
41    
42    3.  Single character iterator optimizations in the JIT compiler.
43    
44    4.  Improved JIT compiler optimizations for character ranges.
45    
46    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
47        Reported by Giuseppe D'Angelo.
48    
49    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
50        of a line, was being set incorrectly in cases where .* appeared inside
51        atomic brackets at the start of a pattern, or where there was a subsequent
52        *PRUNE or *SKIP.
53    
54    7.  Improved instruction cache flush for POWER/PowerPC.
55        Patch by Daniel Richard G.
56    
57    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
58        grep:
59    
60        (a) There is now no limit to the number of patterns to be matched.
61    
62        (b) An error is given if a pattern is too long.
63    
64        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
65            are now supported.
66    
67        (d) --exclude-from and --include-from (multiple use) have been added.
68    
69        (e) Exclusions and inclusions now apply to all files and directories, not
70            just to those obtained from scanning a directory recursively.
71    
72        (f) Multiple uses of -f and --file-list are now supported.
73    
74        (g) In a Windows environment, the default for -d has been changed from
75            "read" (the GNU grep default) to "skip", because otherwise the presence
76            of a directory in the file list provokes an error.
77    
78        (h) The documentation has been revised and clarified in places.
79    
80    9.  Improve the matching speed of capturing brackets.
81    
82    10. Changed the meaning of \X so that it now matches a Unicode extended
83        grapheme cluster.
84    
85    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
86        out POSIX threads when JIT support is configured.
87    
88    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
89    
90    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
91        structure could go wrong in environments where size_t is not the same size
92        as int.
93    
94    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
95        set.
96    
97    15. The EBCDIC support had decayed; later updates to the code had included
98        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
99        general tidy up of EBCDIC-related issues, and the documentation was also
100        not quite right. There is now a test that can be run on ASCII systems to
101        check some of the EBCDIC-related things (but is it not a full test).
102    
103    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
104        in a small tidy to the code.
105    
106    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
107    
108    18. If the --only-matching (-o) option in pcregrep is specified multiple
109        times, each one causes appropriate output. For example, -o1 -o2 outputs the
110        substrings matched by the 1st and 2nd capturing parentheses. A separating
111        string can be specified by --om-separator (default empty).
112    
113    19. Improving the first n character searches.
114    
115    20. Turn case lists for horizontal and vertical white space into macros so that
116        they are defined only once.
117    
118    21. This set of changes together give more compatible Unicode case-folding
119        behaviour for characters that have more than one other case when UCP
120        support is available.
121    
122        (a) The Unicode property table now has offsets into a new table of sets of
123            three or more characters that are case-equivalent. The MultiStage2.py
124            script that generates these tables (the pcre_ucd.c file) now scans
125            CaseFolding.txt instead of UnicodeData.txt for character case
126            information.
127    
128        (b) The code for adding characters or ranges of characters to a character
129            class has been abstracted into a generalized function that also handles
130            case-independence. In UTF-mode with UCP support, this uses the new data
131            to handle characters with more than one other case.
132    
133        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
134            whose other case is greater than 256 are now correctly matched
135            caselessly. Previously, the high codepoint matched the low one, but not
136            vice versa.
137    
138        (d) The processing of \h, \H, \v, and \ in character classes now makes use
139            of the new class addition function, using character lists defined as
140            macros alongside the case definitions of 20 above.
141    
142        (e) Caseless back references now work with characters that have more than
143            one other case.
144    
145        (f) General caseless matching of characters with more than one other case
146            is supported.
147    
148    22. Unicode character properties were updated from Unicode 6.2.0
149    
150    23. Improved CMake support under Windows. Patch by Daniel Richard G.
151    
152    24. Add support for 32-bit character strings, and UTF-32
153    
154    25. Major JIT compiler update (code refactoring and bugfixing).
155        Experimental Sparc 32 support is added.
156    
157    26. Applied a modified version of Daniel Richard G's patch to create
158        pcre.h.generic and config.h.generic by "make" instead of in the
159        PrepareRelease script.
160    
161    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
162        pcre_compile.c when checking for a zero character.
163    
164    28. Introducing a native interface for JIT. Through this interface, the compiled
165        machine code can be directly executed. The purpose of this interface is to
166        provide fast pattern matching, so several sanity checks are not performed.
167        However, feature tests are still performed. The new interface provides
168        1.4x speedup compared to the old one.
169    
170    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
171        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
172        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
173    
174    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
175        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
176        the "old" RFC 2279). Instead, it ended up passing a negative length to
177        pcre_exec().
178    
179    31. Add support for GCC's visibility feature to hide internal functions.
180    
181    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
182        "unknown -C option" after outputting 0 or 1.
183    
184    33. There is now support for generating a code coverage report for the test
185        suite in environments where gcc is the compiler and lcov is installed. This
186        is mainly for the benefit of the developers.
187    
188    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
189        unaddressable using valgrind annotations, allowing valgrind to detect
190        invalid memory accesses. This is mainly for the benefit of the developers.
191    
192    25. (*UTF) can now be used to start a pattern in any of the three libraries.
193    
194    26. Give configure error if --enable-cpp but no C++ compiler found.
195    
196    
197    Version 8.31 06-July-2012
198    -------------------------
199    
200    1.  Fixing a wrong JIT test case and some compiler warnings.
201    
202    2.  Removed a bashism from the RunTest script.
203    
204    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
205        to unsigned type, result still unsigned" that was given by an MS compiler
206        on encountering the code "-sizeof(xxx)".
207    
208    4.  Partial matching support is added to the JIT compiler.
209    
210    5.  Fixed several bugs concerned with partial matching of items that consist
211        of more than one character:
212    
213        (a) /^(..)\1/ did not partially match "aba" because checking references was
214            done on an "all or nothing" basis. This also applied to repeated
215            references.
216    
217        (b) \R did not give a hard partial match if \r was found at the end of the
218            subject.
219    
220        (c) \X did not give a hard partial match after matching one or more
221            characters at the end of the subject.
222    
223        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
224            a partial match for the string "\r".
225    
226        (e) When newline was set to CRLF, the metacharacter "." did not recognize
227            a partial match for a CR character at the end of the subject string.
228    
229    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
230        running pcretest, the text "(JIT)" added to the output whenever JIT is
231        actually used to run the match.
232    
233    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
234        or /S+[+] with a digit between 1 and 7.
235    
236    8.  OP_NOT now supports any UTF character not just single-byte ones.
237    
238    9.  (*MARK) control verb is now supported by the JIT compiler.
239    
240    10. The command "./RunTest list" lists the available tests without actually
241        running any of them. (Because I keep forgetting what they all are.)
242    
243    11. Add PCRE_INFO_MAXLOOKBEHIND.
244    
245    12. Applied a (slightly modified) user-supplied patch that improves performance
246        when the heap is used for recursion (compiled with --disable-stack-for-
247        recursion). Instead of malloc and free for each heap frame each time a
248        logical recursion happens, frames are retained on a chain and re-used where
249        possible. This sometimes gives as much as 30% improvement.
250    
251    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
252        call.
253    
254    14. As documented, (*COMMIT) is now confined to within a positive assertion.
255    
256    15. It is now possible to link pcretest with libedit as an alternative to
257        libreadline.
258    
259    16. (*COMMIT) control verb is now supported by the JIT compiler.
260    
261    17. The Unicode data tables have been updated to Unicode 6.1.0.
262    
263    18. Added --file-list option to pcregrep.
264    
265    19. Added binary file support to pcregrep, including the -a, --binary-files,
266        -I, and --text options.
267    
268    20. The madvise function is renamed for posix_madvise for QNX compatibility
269        reasons. Fixed by Giuseppe D'Angelo.
270    
271    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
272    
273    22. Changed the option for creating symbolic links for 16-bit man pages from
274        -s to -sf so that re-installing does not cause issues.
275    
276    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
277    
278    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
279        matches in certain environments (the workspace was not being correctly
280        retained). Also added to pcre_dfa_exec() a simple plausibility check on
281        some of the workspace data at the beginning of a restart.
282    
283    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
284        was not doing so when it should - probably a typo introduced by SVN 528
285        (change 8.10/14).
286    
287    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
288        \w+ when the character tables indicated that \x{c4} was a word character.
289        There were several related cases, all because the tests for doing a table
290        lookup were testing for characters less than 127 instead of 255.
291    
292    27. If a pattern contains capturing parentheses that are not used in a match,
293        their slots in the ovector are set to -1. For those that are higher than
294        any matched groups, this happens at the end of processing. In the case when
295        there were back references that the ovector was too small to contain
296        (causing temporary malloc'd memory to be used during matching), and the
297        highest capturing number was not used, memory off the end of the ovector
298        was incorrectly being set to -1. (It was using the size of the temporary
299        memory instead of the true size.)
300    
301    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
302        ovector size, it uses memory at the end of the block that it has got.
303    
304    29. Check for an overlong MARK name and give an error at compile time. The
305        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
306    
307    30. JIT compiler update.
308    
309    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
310        Rill for the patch.
311    
312    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
313    
314    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
315    
316    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
317        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
318        was enabled.
319    
320    35. Improve JIT code generation for greedy plus quantifier.
321    
322    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
323        1 to "aa" instead of to an empty string. The bug affected repeated groups
324        that could potentially match an empty string.
325    
326    37. Optimizing single character iterators in JIT.
327    
328    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
329        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
330        codepoints that are too big for the mode are faulted, and in a UTF mode,
331        disallowed codepoints are also faulted.
332    
333    39. If PCRE was compiled with UTF support, in three places in the DFA
334        matcher there was code that should only have been obeyed in UTF mode, but
335        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
336        processing when bytes with values greater than 127 were present. In 16-bit
337        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
338        both cases the values are those that cannot be the first data item in a UTF
339        character. The three items that might have provoked this were recursions,
340        possessively repeated groups, and atomic groups.
341    
342    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
343        and pcregrep, because some OS require shared objects to be explicitly
344        passed to ld, causing the link step to fail if they are not.
345    
346    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
347        mode, patterns that started with \h* or \R* might be incorrectly matched.
348    
349    
350    Version 8.30 04-February-2012
351    -----------------------------
352    
353    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
354        name is defined in ctype.h.
355    
356    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
357        only in quite long subpatterns.
358    
359    3.  Removed the function pcre_info(), which has been obsolete and deprecated
360        since it was replaced by pcre_fullinfo() in February 2000.
361    
362    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
363        match a (*MARK), and the match failed at the start of the subject, a
364        reference to memory before the start of the subject could occur. This bug
365        was introduced by fix 17 of release 8.21.
366    
367    5.  A reference to an unset group with zero minimum repetition was giving
368        totally wrong answers (in non-JavaScript-compatibility mode). For example,
369        /(another)?(\1?)test/ matched against "hello world test". This bug was
370        introduced in release 8.13.
371    
372    6.  Add support for 16-bit character strings (a large amount of work involving
373        many changes and refactorings).
374    
375    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
376        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
377        from a file.
378    
379    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
380        rounding is not applied in this particular case).
381    
382    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
383        if they appear, or are escaped, in patterns.
384    
385    10. Get rid of a number of -Wunused-but-set-variable warnings.
386    
387    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
388        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
389        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
390        also returns the mark "x". This bug applied to capturing parentheses,
391        non-capturing parentheses, and atomic parentheses. It also applied to some
392        assertions.
393    
394    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
395        information out of configure.ac instead of relying on pcre.h.generic, which
396        is not stored in the repository.
397    
398    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
399        -lreadline.
400    
401    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
402    
403    15. Applied Graycode's patch to put the top-level frame on the stack rather
404        than the heap when not using the stack for recursion. This gives a
405        performance improvement in many cases when recursion is not deep.
406    
407    16. Experimental code added to "pcretest -C" to output the stack frame size.
408    
409    
410    Version 8.21 12-Dec-2011
411    ------------------------
412    
413    1.  Updating the JIT compiler.
414    
415    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
416        are added as well.
417    
418    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
419        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
420        calling _pcre_jit_exec. Some extra comments are added.
421    
422    4.  (*MARK) settings inside atomic groups that do not contain any capturing
423        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
424        was introduced by change 18 for 8.20.
425    
426    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
427        ECMA-262 standard.
428    
429    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
430        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
431        This bug was probably introduced by change 9 of 8.13.
432    
433    7.  While fixing 6 above, I noticed that a number of other items were being
434        incorrectly rejected as "not fixed length". This arose partly because newer
435        opcodes had not been added to the fixed-length checking code. I have (a)
436        corrected the bug and added tests for these items, and (b) arranged for an
437        error to occur if an unknown opcode is encountered while checking for fixed
438        length instead of just assuming "not fixed length". The items that were
439        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
440        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
441        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
442    
443    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
444        being incorrectly compiled and would have given unpredicatble results.
445    
446    9.  A possessively repeated subpattern with minimum repeat count greater than
447        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
448        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
449        the first (A) could occur when it should not.
450    
451    10. Add a cast and remove a redundant test from the code.
452    
453    11. JIT should use pcre_malloc/pcre_free for allocation.
454    
455    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
456        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
457        is anything other than /usr, -L is still shown).
458    
459    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
460    
461    14. Perl does not support \N without a following name in a [] class; PCRE now
462        also gives an error.
463    
464    15. If a forward reference was repeated with an upper limit of around 2000,
465        it caused the error "internal error: overran compiling workspace". The
466        maximum number of forward references (including repeats) was limited by the
467        internal workspace, and dependent on the LINK_SIZE. The code has been
468        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
469        the default depends on LINK_SIZE. There is a new upper limit (for safety)
470        of around 200,000 forward references. While doing this, I also speeded up
471        the filling in of repeated forward references.
472    
473    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
474        incorrectly expecting the subject to contain another "a" after the start.
475    
476    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
477        in the match, the SKIP should be ignored. This was not happening; instead
478        the SKIP was being treated as NOMATCH. For patterns such as
479        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
480        tested.
481    
482    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
483        now much more compatible with Perl, in particular in cases where the result
484        is a non-match for a non-anchored pattern. For example, if
485        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
486        "m", where previously it did not return a name. A side effect of this
487        change is that for partial matches, the last encountered mark name is
488        returned, as for non matches. A number of tests that were previously not
489        Perl-compatible have been moved into the Perl-compatible test files. The
490        refactoring has had the pleasing side effect of removing one argument from
491        the match() function, thus reducing its stack requirements.
492    
493    19. If the /S+ option was used in pcretest to study a pattern using JIT,
494        subsequent uses of /S (without +) incorrectly behaved like /S+.
495    
496    21. Retrieve executable code size support for the JIT compiler and fixing
497        some warnings.
498    
499    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
500        not work when the shorter character appeared right at the end of the
501        subject string.
502    
503    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
504        systems.
505    
506    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
507        output it when the /M option is used in pcretest.
508    
509    25. The CheckMan script was not being included in the distribution. Also, added
510        an explicit "perl" to run Perl scripts from the PrepareRelease script
511        because this is reportedly needed in Windows.
512    
513    26. If study data was being save in a file and studying had not found a set of
514        "starts with" bytes for the pattern, the data written to the file (though
515        never used) was taken from uninitialized memory and so caused valgrind to
516        complain.
517    
518    27. Updated RunTest.bat as provided by Sheri Pierce.
519    
520    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
521    
522    29. Computation of memory usage for the table of capturing group names was
523        giving an unnecessarily large value.
524    
525    
526    Version 8.20 21-Oct-2011
527    ------------------------
528    
529    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
530        a POSIX class. After further experiments with Perl, which convinced me that
531        Perl has bugs and confusions, a closing square bracket is no longer allowed
532        in a POSIX name. This bug also affected patterns with classes that started
533        with full stops.
534    
535    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
536        captured substring, but while checking the failing first alternative,
537        substring 1 is temporarily captured. If the output vector supplied to
538        pcre_exec() was not big enough for this capture, the yield of the function
539        was still zero ("insufficient space for captured substrings"). This cannot
540        be totally fixed without adding another stack variable, which seems a lot
541        of expense for a edge case. However, I have improved the situation in cases
542        such as /(a)(b)x|abc/ matched against "abc", where the return code
543        indicates that fewer than the maximum number of slots in the ovector have
544        been set.
545    
546    3.  Related to (2) above: when there are more back references in a pattern than
547        slots in the output vector, pcre_exec() uses temporary memory during
548        matching, and copies in the captures as far as possible afterwards. It was
549        using the entire output vector, but this conflicts with the specification
550        that only 2/3 is used for passing back captured substrings. Now it uses
551        only the first 2/3, for compatibility. This is, of course, another edge
552        case.
553    
554    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
555        main code base, and can be used by building with --enable-jit. When this is
556        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
557        runtime --no-jit option is given.
558    
559    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
560        ovector, the return from the function was zero, implying that there were
561        other matches that did not fit. The correct "exactly full" value is now
562        returned.
563    
564    6.  If a subpattern that was called recursively or as a subroutine contained
565        (*PRUNE) or any other control that caused it to give a non-standard return,
566        invalid errors such as "Error -26 (nested recursion at the same subject
567        position)" or even infinite loops could occur.
568    
569    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
570        computing the minimum length on reaching *ACCEPT, and so ended up with the
571        wrong value of 1 rather than 0. Further investigation indicates that
572        computing a minimum subject length in the presence of *ACCEPT is difficult
573        (think back references, subroutine calls), and so I have changed the code
574        so that no minimum is registered for a pattern that contains *ACCEPT.
575    
576    8.  If (*THEN) was present in the first (true) branch of a conditional group,
577        it was not handled as intended. [But see 16 below.]
578    
579    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
580        Sheri Pierce.
581    
582    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
583        the first byte in a match must be "a".
584    
585    11. Change 17 for 8.13 increased the recursion depth for patterns like
586        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
587        pattern contains any instances of (*THEN). If it does not, the old
588        optimizations are restored. It would be nice to do this on a per-group
589        basis, but at the moment that is not feasible.
590    
591    12. In some environments, the output of pcretest -C is CRLF terminated. This
592        broke RunTest's code that checks for the link size. A single white space
593        character after the value is now allowed for.
594    
595    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
596        For "fr", it uses the Windows-specific input and output files.
597    
598    14. If (*THEN) appeared in a group that was called recursively or as a
599        subroutine, it did not work as intended. [But see next item.]
600    
601    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
602        pattern fragments (but not containing any | characters). If A and B are
603        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
604        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
605        D. In other words, Perl considers parentheses that do not contain any |
606        characters to be part of a surrounding alternative, whereas PCRE was
607        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
608        differently. PCRE now behaves in the same way as Perl, except in the case
609        of subroutine/recursion calls such as (?1) which have in any case always
610        been different (but PCRE had them first :-).
611    
612    16. Related to 15 above: Perl does not treat the | in a conditional group as
613        creating alternatives. Such a group is treated in the same way as an
614        ordinary group without any | characters when processing (*THEN). PCRE has
615        been changed to match Perl's behaviour.
616    
617    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
618        RunGrepTest script failed.
619    
620    18. Change 22 for version 13 caused atomic groups to use more stack. This is
621        inevitable for groups that contain captures, but it can lead to a lot of
622        stack use in large patterns. The old behaviour has been restored for atomic
623        groups that do not contain any capturing parentheses.
624    
625    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
626        suppress the check for a minimum subject length at run time. (If it was
627        given to pcre_exec() or pcre_dfa_exec() it did work.)
628    
629    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
630        fail to work when decoding hex characters in data strings in EBCDIC
631        environments.
632    
633    21. It appears that in at least one Mac OS environment, the isxdigit() function
634        is implemented as a macro that evaluates to its argument more than once,
635        contravening the C 90 Standard (I haven't checked a later standard). There
636        was an instance in pcretest which caused it to go wrong when processing
637        \x{...} escapes in subject strings. The has been rewritten to avoid using
638        things like p++ in the argument of isxdigit().
639    
640    
641    Version 8.13 16-Aug-2011
642    ------------------------
643    
644    1.  The Unicode data tables have been updated to Unicode 6.0.0.
645    
646    2.  Two minor typos in pcre_internal.h have been fixed.
647    
648    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
649        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
650        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
651    
652    4.  There were a number of related bugs in the code for matching backrefences
653        caselessly in UTF-8 mode when codes for the characters concerned were
654        different numbers of bytes. For example, U+023A and U+2C65 are an upper
655        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
656        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
657        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
658        2-byte code at the end of the subject (it thought there wasn't enough data
659        left).
660    
661    5.  Comprehensive information about what went wrong is now returned by
662        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
663        as the output vector has at least 2 elements. The offset of the start of
664        the failing character and a reason code are placed in the vector.
665    
666    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
667        now returned is for the first byte of the failing character, instead of the
668        last byte inspected. This is an incompatible change, but I hope it is small
669        enough not to be a problem. It makes the returned offset consistent with
670        pcre_exec() and pcre_dfa_exec().
671    
672    7.  pcretest now gives a text phrase as well as the error number when
673        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
674        failure, the offset and reason code are output.
675    
676    8.  When \R was used with a maximizing quantifier it failed to skip backwards
677        over a \r\n pair if the subsequent match failed. Instead, it just skipped
678        back over a single character (\n). This seems wrong (because it treated the
679        two characters as a single entity when going forwards), conflicts with the
680        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
681        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
682        has been changed.
683    
684    9.  Some internal refactoring has changed the processing so that the handling
685        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
686        time (the PCRE_DOTALL option was changed this way some time ago: version
687        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
688        which was always a bit of a fudge. It also means that there is one less
689        argument for the match() function, which reduces its stack requirements
690        slightly. This change also fixes an incompatibility with Perl: the pattern
691        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
692    
693    10. More internal refactoring has drastically reduced the number of recursive
694        calls to match() for possessively repeated groups such as (abc)++ when
695        using pcre_exec().
696    
697    11. While implementing 10, a number of bugs in the handling of groups were
698        discovered and fixed:
699    
700        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
701        (a|)*(?1) gave a compile-time internal error.
702        ((a|)+)+  did not notice that the outer group could match an empty string.
703        (^a|^)+   was not marked as anchored.
704        (.*a|.*)+ was not marked as matching at start or after a newline.
705    
706    12. Yet more internal refactoring has removed another argument from the match()
707        function. Special calls to this function are now indicated by setting a
708        value in a variable in the "match data" data block.
709    
710    13. Be more explicit in pcre_study() instead of relying on "default" for
711        opcodes that mean there is no starting character; this means that when new
712        ones are added and accidentally left out of pcre_study(), testing should
713        pick them up.
714    
715    14. The -s option of pcretest has been documented for ages as being an old
716        synonym of -m (show memory usage). I have changed it to mean "force study
717        for every regex", that is, assume /S for every regex. This is similar to -i
718        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
719        using it. It makes it easier to run collections of tests with and without
720        study enabled, and thereby test pcre_study() more easily. All the standard
721        tests are now run with and without -s (but some patterns can be marked as
722        "never study" - see 20 below).
723    
724    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
725        restoration of the capturing data to the outer values was not happening
726        correctly.
727    
728    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
729        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
730        pattern had matched an empty string, and so incorrectly returned a no
731        match.
732    
733    17. There was optimizing code for the last branch of non-capturing parentheses,
734        and also for the obeyed branch of a conditional subexpression, which used
735        tail recursion to cut down on stack usage. Unfortunately, now that there is
736        the possibility of (*THEN) occurring in these branches, tail recursion is
737        no longer possible because the return has to be checked for (*THEN). These
738        two optimizations have therefore been removed. [But see 8.20/11 above.]
739    
740    18. If a pattern containing \R was studied, it was assumed that \R always
741        matched two bytes, thus causing the minimum subject length to be
742        incorrectly computed because \R can also match just one byte.
743    
744    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
745        was incorrectly computed.
746    
747    20. If /S is present twice on a test pattern in pcretest input, it now
748        *disables* studying, thereby overriding the use of -s on the command line
749        (see 14 above). This is necessary for one or two tests to keep the output
750        identical in both cases.
751    
752    21. When (*ACCEPT) was used in an assertion that matched an empty string and
753        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
754    
755    22. When an atomic group that contained a capturing parenthesis was
756        successfully matched, but the branch in which it appeared failed, the
757        capturing was not being forgotten if a higher numbered group was later
758        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
759        group 1 to "a", when in fact it should be unset. This applied to multi-
760        branched capturing and non-capturing groups, repeated or not, and also to
761        positive assertions (capturing in negative assertions does not happen
762        in PCRE) and also to nested atomic groups.
763    
764    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
765        subject after a captured substring, to make it easier to tell which of a
766        number of identical substrings has been captured.
767    
768    24. The way atomic groups are processed by pcre_exec() has been changed so that
769        if they are repeated, backtracking one repetition now resets captured
770        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
771        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
772        "aaa". Previously, it would have been "a". As part of this code
773        refactoring, the way recursive calls are handled has also been changed.
774    
775    25. If an assertion condition captured any substrings, they were not passed
776        back unless some other capturing happened later. For example, if
777        (?(?=(a))a) was matched against "a", no capturing was returned.
778    
779    26. When studying a pattern that contained subroutine calls or assertions,
780        the code for finding the minimum length of a possible match was handling
781        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
782        group 1 called group 2 while simultaneously a separate group 2 called group
783        1). A stack overflow occurred in this case. I have fixed this by limiting
784        the recursion depth to 10.
785    
786    27. Updated RunTest.bat in the distribution to the version supplied by Tom
787        Fortmann. This supports explicit test numbers on the command line, and has
788        argument validation and error reporting.
789    
790    28. An instance of \X with an unlimited repeat could fail if at any point the
791        first character it looked at was a mark character.
792    
793    29. Some minor code refactoring concerning Unicode properties and scripts
794        should reduce the stack requirement of match() slightly.
795    
796    30. Added the '=' option to pcretest to check the setting of unused capturing
797        slots at the end of the pattern, which are documented as being -1, but are
798        not included in the return count.
799    
800    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
801        compiled something random. Now it gives a compile-time error (as does
802        Perl).
803    
804    32. A *MARK encountered during the processing of a positive assertion is now
805        recorded and passed back (compatible with Perl).
806    
807    33. If --only-matching or --colour was set on a pcregrep call whose pattern
808        had alternative anchored branches, the search for a second match in a line
809        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
810        matched the line "0102" twice. The same bug affected patterns that started
811        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
812        twice.
813    
814    34. Previously, PCRE did not allow quantification of assertions. However, Perl
815        does, and because of capturing effects, quantifying parenthesized
816        assertions may at times be useful. Quantifiers are now allowed for
817        parenthesized assertions.
818    
819    35. A minor code tidy in pcre_compile() when checking options for \R usage.
820    
821    36. \g was being checked for fancy things in a character class, when it should
822        just be a literal "g".
823    
824    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
825        appearance of a nested POSIX class supersedes an apparent external class.
826        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
827        unescaped square brackets may also appear as part of class names. For
828        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
829        more like Perl. (But see 8.20/1 above.)
830    
831    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
832        was because it thought it was \N{name}, which is not supported).
833    
834    39. Add minix to OS list not supporting the -S option in pcretest.
835    
836    40. PCRE tries to detect cases of infinite recursion at compile time, but it
837        cannot analyze patterns in sufficient detail to catch mutual recursions
838        such as ((?1))((?2)). There is now a runtime test that gives an error if a
839        subgroup is called recursively as a subpattern for a second time at the
840        same position in the subject string. In previous releases this might have
841        been caught by the recursion limit, or it might have run out of stack.
842    
843    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
844        happen only once. PCRE was, however incorrectly giving a compile time error
845        "recursive call could loop indefinitely" because it cannot analyze the
846        pattern in sufficient detail. The compile time test no longer happens when
847        PCRE is compiling a conditional subpattern, but actual runaway loops are
848        now caught at runtime (see 40 above).
849    
850    42. It seems that Perl allows any characters other than a closing parenthesis
851        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
852        has been changed to be the same.
853    
854    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
855        as not to get warnings when autogen.sh is called. Also changed
856        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
857    
858    44. To help people who use pcregrep to scan files containing exceedingly long
859        lines, the following changes have been made:
860    
861        (a) The default value of the buffer size parameter has been increased from
862            8K to 20K. (The actual buffer used is three times this size.)
863    
864        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
865            PCRE is built.
866    
867        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
868            to be set at run time.
869    
870        (d) Numerical values in pcregrep options can be followed by K or M, for
871            example --buffer-size=50K.
872    
873        (e) If a line being scanned overflows pcregrep's buffer, an error is now
874            given and the return code is set to 2.
875    
876    45. Add a pointer to the latest mark to the callout data block.
877    
878    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
879        partial match of an empty string instead of no match. This was specific to
880        the use of ".".
881    
882    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
883        complete match instead of a partial match. This bug was dependent on both
884        the PCRE_UTF8 and PCRE_DOTALL options being set.
885    
886    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
887        starting byte set, because \b was not being ignored.
888    
889    
890    Version 8.12 15-Jan-2011
891    ------------------------
892    
893    1.  Fixed some typos in the markup of the man pages, and wrote a script that
894        checks for such things as part of the documentation building process.
895    
896    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
897        --match-limit and --recursion-limit options (added for 8.11). In
898        particular, this made one of the standard tests fail. (The integer value
899        went into the wrong half of a long int.)
900    
901    3.  If the --colour option was given to pcregrep with -v (invert match), it
902        did strange things, either producing crazy output, or crashing. It should,
903        of course, ignore a request for colour when reporting lines that do not
904        match.
905    
906    4.  Another pcregrep bug caused similar problems if --colour was specified with
907        -M (multiline) and the pattern match finished with a line ending.
908    
909    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
910        matched in multiline mode, the following line was shown as part of the
911        match. This seems wrong, so I have changed it.
912    
913    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
914        the check for further matches in the same line (so they could be coloured)
915        to overrun the end of the current line. If another match was found, it was
916        incorrectly shown (and then shown again when found in the next line).
917    
918    7.  If pcregrep was compiled under Windows, there was a reference to the
919        function pcregrep_exit() before it was defined. I am assuming this was
920        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
921        reported by a user. I've moved the definition above the reference.
922    
923    
924    Version 8.11 10-Dec-2010
925    ------------------------
926    
927    1.  (*THEN) was not working properly if there were untried alternatives prior
928        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
929        backtracked to try for "b" instead of moving to the next alternative branch
930        at the same level (in this case, to look for "c"). The Perl documentation
931        is clear that when (*THEN) is backtracked onto, it goes to the "next
932        alternative in the innermost enclosing group".
933    
934    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
935        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
936        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
937        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
938        (*THEN).
939    
940    3.  If \s appeared in a character class, it removed the VT character from
941        the class, even if it had been included by some previous item, for example
942        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
943        of \s, but is part of the POSIX "space" class.)
944    
945    4.  A partial match never returns an empty string (because you can always
946        match an empty string at the end of the subject); however the checking for
947        an empty string was starting at the "start of match" point. This has been
948        changed to the "earliest inspected character" point, because the returned
949        data for a partial match starts at this character. This means that, for
950        example, /(?<=abc)def/ gives a partial match for the subject "abc"
951        (previously it gave "no match").
952    
953    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
954        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
955        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
956        has an implication that the given string is incomplete (because a partial
957        match is preferred over a full match). For this reason, these items now
958        give a partial match in this situation. [Aside: previously, the one case
959        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
960        match rather than a full match, which was wrong by the old rules, but is
961        now correct.]
962    
963    6.  There was a bug in the handling of #-introduced comments, recognized when
964        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
965        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
966        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
967        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
968        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
969        places in pcre_compile().
970    
971    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
972        comments when looking ahead for named forward references to subpatterns,
973        the only newline sequence it recognized was NL. It now handles newlines
974        according to the set newline convention.
975    
976    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
977        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
978        cater for a lack of strerror(). These oversights have been fixed.
979    
980    9.  Added --match-limit and --recursion-limit to pcregrep.
981    
982    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
983    
984    11. When the -o option was used, pcregrep was setting a return code of 1, even
985        when matches were found, and --line-buffered was not being honoured.
986    
987    12. Added an optional parentheses number to the -o and --only-matching options
988        of pcregrep.
989    
990    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
991        can match an empty string. The code to do it in pcretest and pcredemo
992        needed fixing:
993    
994        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
995            only one byte after an empty string match just before CRLF (this case
996            just got forgotten; "any" and "anycrlf" were OK).
997    
998        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
999            mode when an empty string match preceded an ASCII character followed by
1000            a non-ASCII character. (The code for advancing by one character rather
1001            than one byte was nonsense.)
1002    
1003        (c) The pcredemo.c sample program did not have any code at all to handle
1004            the cases when CRLF is a valid newline sequence.
1005    
1006    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1007        as a starting offset was within the subject string. There is now a new
1008        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1009        negative or greater than the length of the string. In order to test this,
1010        pcretest is extended to allow the setting of negative starting offsets.
1011    
1012    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1013        starting offset points to the beginning of a UTF-8 character was
1014        unnecessarily clumsy. I tidied it up.
1015    
1016    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1017        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1018    
1019    17. Nobody had reported that the --include_dir option, which was added in
1020        release 7.7 should have been called --include-dir (hyphen, not underscore)
1021        for compatibility with GNU grep. I have changed it to --include-dir, but
1022        left --include_dir as an undocumented synonym, and the same for
1023        --exclude-dir, though that is not available in GNU grep, at least as of
1024        release 2.5.4.
1025    
1026    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1027        characters from a string of bytes) have been redefined so as not to use
1028        loops, in order to improve performance in some environments. At the same
1029        time, I abstracted some of the common code into auxiliary macros to save
1030        repetition (this should not affect the compiled code).
1031    
1032    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1033        compile-time error is now given if \c is not followed by an ASCII
1034        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1035        different, and any byte value is allowed.)
1036    
1037    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1038        START_OPTIMIZE option, which is now allowed at compile time - but just
1039        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1040        to pcregrep and other applications that have no direct access to PCRE
1041        options. The new /Y option in pcretest sets this option when calling
1042        pcre_compile().
1043    
1044    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1045        back references. Groups containing recursive back references were forced to
1046        be atomic by that change, but in the case of named groups, the amount of
1047        memory required was incorrectly computed, leading to "Failed: internal
1048        error: code overflow". This has been fixed.
1049    
1050    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1051        pcretest.c, to avoid build problems in some Borland environments.
1052    
1053    
1054    Version 8.10 25-Jun-2010
1055    ------------------------
1056    
1057    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1058        THEN.
1059    
1060    2.  (*ACCEPT) was not working when inside an atomic group.
1061    
1062    3.  Inside a character class, \B is treated as a literal by default, but
1063        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1064        causes the error). The code is unchanged, but I tidied the documentation.
1065    
1066    4.  Inside a character class, PCRE always treated \R and \X as literals,
1067        whereas Perl faults them if its -w option is set. I have changed PCRE so
1068        that it faults them when PCRE_EXTRA is set.
1069    
1070    5.  Added support for \N, which always matches any character other than
1071        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1072    
1073    6.  When compiling pcregrep with newer versions of gcc which may have
1074        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1075        declared with attribute warn_unused_result" were given. Just casting the
1076        result to (void) does not stop the warnings; a more elaborate fudge is
1077        needed. I've used a macro to implement this.
1078    
1079    7.  Minor change to pcretest.c to avoid a compiler warning.
1080    
1081    8.  Added four artifical Unicode properties to help with an option to make
1082        \s etc use properties (see next item). The new properties are: Xan
1083        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1084    
1085    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1086        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1087        this option. Modified pcretest to add /W to test this facility. Added
1088        REG_UCP to make it available via the POSIX interface.
1089    
1090    10. Added --line-buffered to pcregrep.
1091    
1092    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1093        studied, and the match started with a letter with a code point greater than
1094        127 whose first byte was different to the first byte of the other case of
1095        the letter, the other case of this starting letter was not recognized
1096        (#976).
1097    
1098    12. If a pattern that was studied started with a repeated Unicode property
1099        test, for example, \p{Nd}+, there was the theoretical possibility of
1100        setting up an incorrect bitmap of starting bytes, but fortunately it could
1101        not have actually happened in practice until change 8 above was made (it
1102        added property types that matched character-matching opcodes).
1103    
1104    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1105        possible starting bytes for non-anchored patterns.
1106    
1107    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1108        \R, and also a number of cases that involve Unicode properties, both
1109        explicit and implicit when PCRE_UCP is set.
1110    
1111    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1112        input, it could crash or give wrong results if characters with values
1113        greater than 0xc0 were present in the subject string. (Detail: it assumed
1114        UTF-8 input when processing these items.)
1115    
1116    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1117        size_t is 64-bit (#991).
1118    
1119    17. Added a check for running out of memory when PCRE is compiled with
1120        --disable-stack-for-recursion (#990).
1121    
1122    18. If the last data line in a file for pcretest does not have a newline on
1123        the end, a newline was missing in the output.
1124    
1125    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1126        less than 128) in its various bitmaps. However, there is a facility for
1127        generating tables according to the current locale when PCRE is compiled. It
1128        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1129        characters, are recognized by isspace() and therefore were getting set in
1130        these tables, and indeed these tables seem to approximate to ISO 8859. This
1131        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1132        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1133        which of course cannot start UTF-8 characters. I have changed the code so
1134        that only real ASCII characters (less than 128) and the correct starting
1135        bytes for UTF-8 encodings are set for characters greater than 127 when in
1136        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1137        altogether.)
1138    
1139    20. Added the /T option to pcretest so as to be able to run tests with non-
1140        standard character tables, thus making it possible to include the tests
1141        used for 19 above in the standard set of tests.
1142    
1143    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1144        reference to a subpattern the other side of a comment that contains an
1145        opening parenthesis caused either an internal compiling error, or a
1146        reference to the wrong subpattern.
1147    
1148    
1149    Version 8.02 19-Mar-2010
1150    ------------------------
1151    
1152    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1153    
1154    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1155        configured.
1156    
1157    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1158        original author of that file, following a query about its status.
1159    
1160    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1161        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1162    
1163    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1164        quantifier applied to a forward-referencing subroutine call, could compile
1165        incorrect code or give the error "internal error: previously-checked
1166        referenced subpattern not found".
1167    
1168    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1169        variables to point to external functions. For these systems, therefore,
1170        pcre_malloc etc. are now initialized to local functions that call the
1171        relevant global functions.
1172    
1173    7.  There were two entries missing in the vectors called coptable and poptable
1174        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1175        I've fixed the data, and added a kludgy way of testing at compile time that
1176        the lengths are correct (equal to the number of opcodes).
1177    
1178    8.  Following on from 7, I added a similar kludge to check the length of the
1179        eint vector in pcreposix.c.
1180    
1181    9.  Error texts for pcre_compile() are held as one long string to avoid too
1182        much relocation at load time. To find a text, the string is searched,
1183        counting zeros. There was no check for running off the end of the string,
1184        which could happen if a new error number was added without updating the
1185        string.
1186    
1187    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1188    
1189    11. \K was not working if it appeared in an atomic group or in a group that
1190        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1191        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1192        the assertion is positive, but not if it is negative.
1193    
1194    12. Change 11 fortuitously reduced the size of the stack frame used in the
1195        "match()" function of pcre_exec.c by one pointer. Forthcoming
1196        implementation of support for (*MARK) will need an extra pointer on the
1197        stack; I have reserved it now, so that the stack frame size does not
1198        decrease.
1199    
1200    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1201        item in branch that calls a recursion is a subroutine call - as in the
1202        second branch in the above example - was incorrectly given the compile-
1203        time error "recursive call could loop indefinitely" because pcre_compile()
1204        was not correctly checking the subroutine for matching a non-empty string.
1205    
1206    14. The checks for overrunning compiling workspace could trigger after an
1207        overrun had occurred. This is a "should never occur" error, but it can be
1208        triggered by pathological patterns such as hundreds of nested parentheses.
1209        The checks now trigger 100 bytes before the end of the workspace.
1210    
1211    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1212    
1213    
1214    Version 8.01 19-Jan-2010
1215    ------------------------
1216    
1217    1.  If a pattern contained a conditional subpattern with only one branch (in
1218        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1219        computed the wrong minimum data length (which is of course zero for such
1220        subpatterns). This could cause incorrect "no match" results.
1221    
1222    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1223        the pattern is reset in the first branch, pcre_compile() failed with
1224        "internal error: code overflow at offset...". This happened only when
1225        the reset was to the original external option setting. (An optimization
1226        abstracts leading options settings into an external setting, which was the
1227        cause of this.)
1228    
1229    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1230        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1231        assertion pattern did not match (meaning that the assertion was true), it
1232        was incorrectly treated as false if the SKIP had been reached during the
1233        matching. This also applied to assertions used as conditions.
1234    
1235    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1236        assertion subpattern, including such a pattern used as a condition,
1237        unpredictable results occurred, instead of the error return
1238        PCRE_ERROR_DFA_UITEM.
1239    
1240    5.  The C++ GlobalReplace function was not working like Perl for the special
1241        situation when an empty string is matched. It now does the fancy magic
1242        stuff that is necessary.
1243    
1244    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1245        removed. (These were left over from very, very early versions of PCRE.)
1246    
1247    7.  Some cosmetic changes to the code to make life easier when compiling it
1248        as part of something else:
1249    
1250        (a) Change DEBUG to PCRE_DEBUG.
1251    
1252        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1253            called "current" as "current_branch", to prevent a collision with the
1254            Linux macro when compiled as a kernel module.
1255    
1256        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1257            prevent a collision with the Linux macro when compiled as a kernel
1258            module.
1259    
1260    8.  In pcre_compile() there are some checks for integer overflows that used to
1261        cast potentially large values to (double). This has been changed to that
1262        when building, a check for int64_t is made, and if it is found, it is used
1263        instead, thus avoiding the use of floating point arithmetic. (There is no
1264        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1265        double.
1266    
1267    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1268        2005 (difference between two addresses compared to an unsigned value).
1269    
1270    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1271        custom one, because of the following reported problem in Windows:
1272    
1273          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1274              under Win32.
1275          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1276              therefore missing the function definition.
1277          - The compiler thus generates a "C" signature for the test function.
1278          - The linker fails to find the "C" function.
1279          - PCRE fails to configure if asked to do so against libbz2.
1280    
1281    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1282        messages were output:
1283    
1284          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1285          rerunning libtoolize, to keep the correct libtool macros in-tree.
1286          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1287    
1288        I have done both of these things.
1289    
1290    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1291        most of the time, it *can* run out if it is given a pattern that contains a
1292        runaway infinite recursion. I updated the discussion in the pcrestack man
1293        page.
1294    
1295    13. Now that we have gone to the x.xx style of version numbers, the minor
1296        version may start with zero. Using 08 or 09 is a bad idea because users
1297        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1298        interpreted as invalid octal numbers. I've updated the previous comment in
1299        configure.ac, and also added a check that gives an error if 08 or 09 are
1300        used.
1301    
1302    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1303        causing partial matching to fail when the end of the subject matched \W
1304        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1305    
1306    15. There were some discrepancies between the declarations in pcre_internal.h
1307        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1308        their definitions. The declarations used "const uschar *" and the
1309        definitions used USPTR. Even though USPTR is normally defined as "const
1310        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1311        reported that: "This difference in casting confuses some C++ compilers, for
1312        example, SunCC recognizes above declarations as different functions and
1313        generates broken code for hbpcre." I have changed the declarations to use
1314        USPTR.
1315    
1316    16. GNU libtool is named differently on some systems. The autogen.sh script now
1317        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1318        (FreeBSD).
1319    
1320    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1321        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1322        comment: "Figure out how to create a longlong from a string: strtoll and
1323        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1324        instance, but it only takes 2 args instead of 3!"
1325    
1326    18. A subtle bug concerned with back references has been fixed by a change of
1327        specification, with a corresponding code fix. A pattern such as
1328        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1329        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1330        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1331        same bug. Such groups have to be quantified to be useful, or contained
1332        inside another quantified group. (If there's no repetition, the reference
1333        can never match.) The problem arises because, having left the group and
1334        moved on to the rest of the pattern, a later failure that backtracks into
1335        the group uses the captured value from the final iteration of the group
1336        rather than the correct earlier one. I have fixed this in PCRE by forcing
1337        any group that contains a reference to itself to be an atomic group; that
1338        is, there cannot be any backtracking into it once it has completed. This is
1339        similar to recursive and subroutine calls.
1340    
1341    
1342    Version 8.00 19-Oct-09
1343    ----------------------
1344    
1345    1.  The table for translating pcre_compile() error codes into POSIX error codes
1346        was out-of-date, and there was no check on the pcre_compile() error code
1347        being within the table. This could lead to an OK return being given in
1348        error.
1349    
1350    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1351        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1352        in a Windows environment.
1353    
1354    3.  The pcregrep --count option prints the count for each file even when it is
1355        zero, as does GNU grep. However, pcregrep was also printing all files when
1356        --files-with-matches was added. Now, when both options are given, it prints
1357        counts only for those files that have at least one match. (GNU grep just
1358        prints the file name in this circumstance, but including the count seems
1359        more useful - otherwise, why use --count?) Also ensured that the
1360        combination -clh just lists non-zero counts, with no names.
1361    
1362    4.  The long form of the pcregrep -F option was incorrectly implemented as
1363        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1364        but it seems right to fix it, and I didn't think it was worth preserving
1365        the old behaviour.
1366    
1367    5.  The command line items --regex=pattern and --regexp=pattern were not
1368        recognized by pcregrep, which required --regex pattern or --regexp pattern
1369        (with a space rather than an '='). The man page documented the '=' forms,
1370        which are compatible with GNU grep; these now work.
1371    
1372    6.  No libpcreposix.pc file was created for pkg-config; there was just
1373        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1374    
1375    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1376        when UCP support is not needed, by modifying the Python script that
1377        generates it from Unicode data files. This should not matter if the module
1378        is correctly used as a library, but I received one complaint about 50K of
1379        unwanted data. My guess is that the person linked everything into his
1380        program rather than using a library. Anyway, it does no harm.
1381    
1382    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1383        was a minimum greater than 1 for a wide character in a possessive
1384        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1385        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1386        character. Chaos in the form of incorrect output or a compiling loop could
1387        result.
1388    
1389    9.  The restrictions on what a pattern can contain when partial matching is
1390        requested for pcre_exec() have been removed. All patterns can now be
1391        partially matched by this function. In addition, if there are at least two
1392        slots in the offset vector, the offset of the earliest inspected character
1393        for the match and the offset of the end of the subject are set in them when
1394        PCRE_ERROR_PARTIAL is returned.
1395    
1396    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1397        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1398        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1399        and may be more useful for multi-segment matching.
1400    
1401    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1402        used to be given if ever the end of the subject was reached; now it is
1403        given only if matching could not proceed because another character was
1404        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1405        string "Z", which now yields "no match" instead of "partial match". In the
1406        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1407        final character ended with (*FAIL).
1408    
1409    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1410        if the pattern had a "must contain" character that was already found in the
1411        earlier partial match, unless partial matching was again requested. For
1412        example, with the pattern /dog.(body)?/, the "must contain" character is
1413        "g". If the first part-match was for the string "dog", restarting with
1414        "sbody" failed. This bug has been fixed.
1415    
1416    13. The string returned by pcre_dfa_exec() after a partial match has been
1417        changed so that it starts at the first inspected character rather than the
1418        first character of the match. This makes a difference only if the pattern
1419        starts with a lookbehind assertion or \b or \B (\K is not supported by
1420        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1421        matching functions compatible, and I think it's the right thing to do.
1422    
1423    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1424        so that the demonstration program is easily available in environments where
1425        PCRE has not been installed from source.
1426    
1427    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1428        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1429        library.
1430    
1431    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1432        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1433        is not the first non-POSIX option to be added. Clearly some people find
1434        these options useful.
1435    
1436    17. If a caller to the POSIX matching function regexec() passes a non-zero
1437        value for nmatch with a NULL value for pmatch, the value of
1438        nmatch is forced to zero.
1439    
1440    18. RunGrepTest did not have a test for the availability of the -u option of
1441        the diff command, as RunTest does. It now checks in the same way as
1442        RunTest, and also checks for the -b option.
1443    
1444    19. If an odd number of negated classes containing just a single character
1445        interposed, within parentheses, between a forward reference to a named
1446        subpattern and the definition of the subpattern, compilation crashed with
1447        an internal error, complaining that it could not find the referenced
1448        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1449        [The bug was that it was starting one character too far in when skipping
1450        over the character class, thus treating the ] as data rather than
1451        terminating the class. This meant it could skip too much.]
1452    
1453    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1454        /g option in pcretest when the pattern contains \K, which makes it possible
1455        to have an empty string match not at the start, even when the pattern is
1456        anchored. Updated pcretest and pcredemo to use this option.
1457    
1458    21. If the maximum number of capturing subpatterns in a recursion was greater
1459        than the maximum at the outer level, the higher number was returned, but
1460        with unset values at the outer level. The correct (outer level) value is
1461        now given.
1462    
1463    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1464        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1465        make it do so. The string so far is captured, making this feature
1466        compatible with Perl.
1467    
1468    23. The tests have been re-organized, adding tests 11 and 12, to make it
1469        possible to check the Perl 5.10 features against Perl 5.10.
1470    
1471    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1472        pattern matches a fixed length string. PCRE did not allow this; now it
1473        does. Neither allows recursion.
1474    
1475    25. I finally figured out how to implement a request to provide the minimum
1476        length of subject string that was needed in order to match a given pattern.
1477        (It was back references and recursion that I had previously got hung up
1478        on.) This code has now been added to pcre_study(); it finds a lower bound
1479        to the length of subject needed. It is not necessarily the greatest lower
1480        bound, but using it to avoid searching strings that are too short does give
1481        some useful speed-ups. The value is available to calling programs via
1482        pcre_fullinfo().
1483    
1484    26. While implementing 25, I discovered to my embarrassment that pcretest had
1485        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1486        study optimizations had never been tested with that matching function.
1487        Oops. What is worse, even when it was passed study data, there was a bug in
1488        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1489        were also very few tests of studied patterns with pcre_dfa_exec().
1490    
1491    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1492        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1493        on the other side of the coin, they are no longer allowed to have different
1494        names, because these cannot be distinguished in PCRE, and this has caused
1495        confusion. (This is a difference from Perl.)
1496    
1497    28. When duplicate subpattern names are present (necessarily with different
1498        numbers, as required by 27 above), and a test is made by name in a
1499        conditional pattern, either for a subpattern having been matched, or for
1500        recursion in such a pattern, all the associated numbered subpatterns are
1501        tested, and the overall condition is true if the condition is true for any
1502        one of them. This is the way Perl works, and is also more like the way
1503        testing by number works.
1504    
1505    
1506    Version 7.9 11-Apr-09
1507    ---------------------
1508    
1509    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1510        (pcretest), all targets were linked against these libraries. This included
1511        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1512        libraries. This caused unwanted dependencies to be created. This problem
1513        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1514        pcretest is linked with readline.
1515    
1516    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1517        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1518        moved outside it again, because FALSE and TRUE are already defined in AIX,
1519        but BOOL is not.
1520    
1521    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1522        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1523    
1524    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1525        hyphens) following file names and line numbers when outputting matching
1526        lines. This is not true; no spaces are inserted. I have also clarified the
1527        wording for the --colour (or --color) option.
1528    
1529    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1530        was not coloured; this is different to GNU grep, so I have changed it to be
1531        the same.
1532    
1533    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1534        each matching line was coloured. Now it goes on to look for further matches
1535        of any of the test patterns, which is the same behaviour as GNU grep.
1536    
1537    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1538        doesn't make sense to accept an empty string match in pcregrep, so I have
1539        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1540        seems to be how GNU grep behaves.
1541    
1542    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1543        start or after a newline", because the conditional assertion was not being
1544        correctly handled. The rule now is that both the assertion and what follows
1545        in the first alternative must satisfy the test.
1546    
1547    9.  If auto-callout was enabled in a pattern with a conditional group whose
1548        condition was an assertion, PCRE could crash during matching, both with
1549        pcre_exec() and pcre_dfa_exec().
1550    
1551    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1552        used for matching.
1553    
1554    11. Unicode property support in character classes was not working for
1555        characters (bytes) greater than 127 when not in UTF-8 mode.
1556    
1557    12. Added the -M command line option to pcretest.
1558    
1559    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1560    
1561    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1562    
1563    16. Added comments and documentation about mis-use of no_arg in the C++
1564        wrapper.
1565    
1566    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1567        from Martin Jerabek that uses macro names for all relevant character and
1568        string constants.
1569    
1570    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1571        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1572        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1573        these, but not everybody uses configure.
1574    
1575    19. A conditional group that had only one branch was not being correctly
1576        recognized as an item that could match an empty string. This meant that an
1577        enclosing group might also not be so recognized, causing infinite looping
1578        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1579        with the subject "ab", where knowledge that the repeated group can match
1580        nothing is needed in order to break the loop.
1581    
1582    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1583        exec(), but without supplying a callout function, matching went wrong.
1584    
1585    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1586        leak if the size of the offset vector was greater than 30. When the vector
1587        is smaller, the saved offsets during recursion go onto a local stack
1588        vector, but for larger vectors malloc() is used. It was failing to free
1589        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1590        error, in fact).
1591    
1592    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1593        heapframe that is used only when UTF-8 support is enabled. This caused no
1594        problem, but was untidy.
1595    
1596    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1597        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1598        included within another project.
1599    
1600    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1601        slightly modified by me:
1602    
1603          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1604              not building pcregrep.
1605    
1606          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1607              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1608    
1609    25. Forward references, both numeric and by name, in patterns that made use of
1610        duplicate group numbers, could behave incorrectly or give incorrect errors,
1611        because when scanning forward to find the reference group, PCRE was not
1612        taking into account the duplicate group numbers. A pattern such as
1613        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1614    
1615    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1616        the feature of a custom pointer more persuasive (as requested by a user).
1617    
1618    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1619        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1620        pre-defined.
1621    
1622    28. Added support for (*UTF8) at the start of a pattern.
1623    
1624    29. Arrange for flags added by the "release type" setting in CMake to be shown
1625        in the configuration summary.
1626    
1627    
1628    Version 7.8 05-Sep-08
1629    ---------------------
1630    
1631    1.  Replaced UCP searching code with optimized version as implemented for Ad
1632        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1633        stage table and inline lookup instead of a function, giving speed ups of 2
1634        to 5 times on some simple patterns that I tested. Permission was given to
1635        distribute the MultiStage2.py script that generates the tables (it's not in
1636        the tarball, but is in the Subversion repository).
1637    
1638    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1639        scripts.
1640    
1641    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1642        a group with a zero qualifier. The result of the study could be incorrect,
1643        or the function might crash, depending on the pattern.
1644    
1645    4.  Caseless matching was not working for non-ASCII characters in back
1646        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1647        It now works when Unicode Property Support is available.
1648    
1649    5.  In pcretest, an escape such as \x{de} in the data was always generating
1650        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1651        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1652        truncation.
1653    
1654    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1655    
1656    7.  Added two (int) casts to pcregrep when printing the difference of two
1657        pointers, in case they are 64-bit values.
1658    
1659    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1660        test 2 if it fails.
1661    
1662    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1663        and a #define of that name to empty if it is not externally set. This is to
1664        allow users of MSVC to set it if necessary.
1665    
1666    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1667        the convenience functions in the pcre_get.c source file.
1668    
1669    11. An option change at the start of a pattern that had top-level alternatives
1670        could cause overwriting and/or a crash. This command provoked a crash in
1671        some environments:
1672    
1673          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1674    
1675        This potential security problem was recorded as CVE-2008-2371.
1676    
1677    12. For a pattern where the match had to start at the beginning or immediately
1678        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1679        pcre_dfa_exec() could read past the end of the passed subject if there was
1680        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1681        pcretest so that it places the subject at the end of its malloc-ed buffer.
1682    
1683    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1684        exec() might read past the end of the data buffer in UTF-8 mode.
1685    
1686    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1687        the data contained the byte 0x85 as part of a UTF-8 character within its
1688        first line. This applied both to normal and DFA matching.
1689    
1690    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1691        /^[^d]*?$/8 failed to match "abc".
1692    
1693    16. Added a missing copyright notice to pcrecpp_internal.h.
1694    
1695    17. Make it more clear in the documentation that values returned from
1696        pcre_exec() in ovector are byte offsets, not character counts.
1697    
1698    18. Tidied a few places to stop certain compilers from issuing warnings.
1699    
1700    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1701        supplied by Stefan Weber. I made a further small update for 7.8 because
1702        there is a change of source arrangements: the pcre_searchfuncs.c module is
1703        replaced by pcre_ucd.c.
1704    
1705    
1706    Version 7.7 07-May-08
1707    ---------------------
1708    
1709    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1710        a string to a long long, pretend we don't even have a long long." This is
1711        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1712    
1713    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1714        pre-7.6 versions, which defined a global no_arg variable instead of putting
1715        it in the RE class. (See also #8 below.)
1716    
1717    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1718        Lopes.
1719    
1720    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1721    
1722        (1) The include/exclude patterns were being applied to the whole pathnames
1723            of files, instead of just to the final components.
1724    
1725        (2) If there was more than one level of directory, the subdirectories were
1726            skipped unless they satisfied the include/exclude conditions. This is
1727            inconsistent with GNU grep (and could even be seen as contrary to the
1728            pcregrep specification - which I improved to make it absolutely clear).
1729            The action now is always to scan all levels of directory, and just
1730            apply the include/exclude patterns to regular files.
1731    
1732    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1733        --exclude_dir in the tests to avoid scanning .svn directories.
1734    
1735    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1736        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1737        doesn't support NULs in patterns.
1738    
1739    7.  Added some missing "const"s to declarations of static tables in
1740        pcre_compile.c and pcre_dfa_exec.c.
1741    
1742    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1743        caused by fix #2  above. (Subsequently also a second patch to fix the
1744        first patch. And a third patch - this was a messy problem.)
1745    
1746    9.  Applied Craig's patch to remove the use of push_back().
1747    
1748    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1749        matching function regexec().
1750    
1751    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1752        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1753        references. PCRE supports relative numbers with this syntax (I don't think
1754        Oniguruma does).
1755    
1756    12. Previously, a group with a zero repeat such as (...){0} was completely
1757        omitted from the compiled regex. However, this means that if the group
1758        was called as a subroutine from elsewhere in the pattern, things went wrong
1759        (an internal error was given). Such groups are now left in the compiled
1760        pattern, with a new opcode that causes them to be skipped at execution
1761        time.
1762    
1763    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1764        to the way PCRE behaves:
1765    
1766        (a) A lone ] character is dis-allowed (Perl treats it as data).
1767    
1768        (b) A back reference to an unmatched subpattern matches an empty string
1769            (Perl fails the current match path).
1770    
1771        (c) A data ] in a character class must be notated as \] because if the
1772            first data character in a class is ], it defines an empty class. (In
1773            Perl it is not possible to have an empty class.) The empty class []
1774            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1775            The negative empty class [^] matches any one character, independently
1776            of the DOTALL setting.
1777    
1778    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1779        non-existent subpattern following a character class starting with ']' and
1780        containing () gave an internal compiling error instead of "reference to
1781        non-existent subpattern". Fortunately, when the pattern did exist, the
1782        compiled code was correct. (When scanning forwards to check for the
1783        existencd of the subpattern, it was treating the data ']' as terminating
1784        the class, so got the count wrong. When actually compiling, the reference
1785        was subsequently set up correctly.)
1786    
1787    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1788        it was being rejected as not supported by pcre_dfa_exec(), even though
1789        other assertions are supported. I have made pcre_dfa_exec() support
1790        (*FAIL).
1791    
1792    16. The implementation of 13c above involved the invention of a new opcode,
1793        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1794        cannot be changed at match time, I realized I could make a small
1795        improvement to matching performance by compiling OP_ALLANY instead of
1796        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1797        on the OP_ANY path.
1798    
1799    17. Compiling pcretest on Windows with readline support failed without the
1800        following two fixes: (1) Make the unistd.h include conditional on
1801        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1802    
1803    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1804        ncurses library to be included for pcretest when ReadLine support is
1805        requested, but also to allow for it to be overridden. This patch came from
1806        Daniel Bergström.
1807    
1808    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1809        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1810        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1811        spotting this.
1812    
1813    
1814    Version 7.6 28-Jan-08
1815    ---------------------
1816    
1817    1.  A character class containing a very large number of characters with
1818        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1819        overflow.
1820    
1821    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1822        HAVE_LONG_LONG is not defined.
1823    
1824    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1825        bring them up to date and include new features. This patch includes:
1826    
1827        - Fixed PH's badly added libz and libbz2 support.
1828        - Fixed a problem with static linking.
1829        - Added pcredemo. [But later removed - see 7 below.]
1830        - Fixed dftables problem and added an option.
1831        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1832            HAVE_LONG_LONG.
1833        - Added readline support for pcretest.
1834        - Added an listing of the option settings after cmake has run.
1835    
1836    4.  A user submitted a patch to Makefile that makes it easy to create
1837        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1838        Makefile.am that cause it to include this special target, without
1839        affecting anything else. Note that the same mingw target plus all
1840        the other distribution libraries and programs are now supported
1841        when configuring with CMake (see 6 below) instead of with
1842        Configure/Make.
1843    
1844    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1845        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1846        exported in the Windows port". It has not yet been confirmed that the patch
1847        solves the problem, but it does no harm.
1848    
1849    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1850        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1851        with CMake, and also correct the comment about stack recursion.
1852    
1853    7.  Remove the automatic building of pcredemo from the ./configure system and
1854        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1855        of a program that users should build themselves after PCRE is installed, so
1856        building it automatically is not really right. What is more, it gave
1857        trouble in some build environments.
1858    
1859    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1860    
1861    
1862    Version 7.5 10-Jan-08
1863    ---------------------
1864    
1865    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1866        values in parens when parsing an RE using the C++ wrapper."
1867    
1868    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1869        Characters greater than 255 were excluded from the class instead of being
1870        included.
1871    
1872    3.  The same bug as (2) above applied to negated POSIX classes such as
1873        [:^space:].
1874    
1875    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1876        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1877        I have changed it.
1878    
1879    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1880        first named subpattern) and a construct such as (?&a) would reference the
1881        first named subpattern whose name started with "a" (in other words, the
1882        length check was missing). Both these problems are fixed. "Subpattern name
1883        expected" is now given for (?&) (a zero-length name), and this patch also
1884        makes it give the same error for \k'' (previously it complained that that
1885        was a reference to a non-existent subpattern).
1886    
1887    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1888        this is right because (?- can be followed by option settings as well as by
1889        digits. I have, however, made the messages clearer.
1890    
1891    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1892        than the number used in the conditional) now cause a compile-time error.
1893        This is actually not compatible with Perl, which accepts such patterns, but
1894        treats the conditional as always being FALSE (as PCRE used to), but it
1895        seems to me that giving a diagnostic is better.
1896    
1897    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1898        and messages.
1899    
1900    9.  Fix two occurrences of "backslash" in comments that should have been
1901        "backspace".
1902    
1903    10. Remove two redundant lines of code that can never be obeyed (their function
1904        was moved elsewhere).
1905    
1906    11. The program that makes PCRE's Unicode character property table had a bug
1907        which caused it to generate incorrect table entries for sequences of
1908        characters that have the same character type, but are in different scripts.
1909        It amalgamated them into a single range, with the script of the first of
1910        them. In other words, some characters were in the wrong script. There were
1911        thirteen such cases, affecting characters in the following ranges:
1912    
1913          U+002b0 - U+002c1
1914          U+0060c - U+0060d
1915          U+0061e - U+00612
1916          U+0064b - U+0065e
1917          U+0074d - U+0076d
1918          U+01800 - U+01805
1919          U+01d00 - U+01d77
1920          U+01d9b - U+01dbf
1921          U+0200b - U+0200f
1922          U+030fc - U+030fe
1923          U+03260 - U+0327f
1924          U+0fb46 - U+0fbb1
1925          U+10450 - U+1049d
1926    
1927    12. The -o option (show only the matching part of a line) for pcregrep was not
1928        compatible with GNU grep in that, if there was more than one match in a
1929        line, it showed only the first of them. It now behaves in the same way as
1930        GNU grep.
1931    
1932    13. If the -o and -v options were combined for pcregrep, it printed a blank
1933        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1934        does the same. The return code can be used to tell if there were any
1935        non-matching lines.
1936    
1937    14. Added --file-offsets and --line-offsets to pcregrep.
1938    
1939    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1940        infinitely looping recursion. The bug was that positive lookaheads were not
1941        being skipped when checking for a possible empty match (negative lookaheads
1942        and both kinds of lookbehind were skipped).
1943    
1944    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1945        inclusion of <windows.h> to before rather than after the definition of
1946        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1947    
1948    17. Specifying a possessive quantifier with a specific limit for a Unicode
1949        character property caused pcre_compile() to compile bad code, which led at
1950        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1951        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1952        caused the error; without that there was no problem.
1953    
1954    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1955    
1956    19. Added --enable-pcretest-libreadline.
1957    
1958    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1959        RE::GlobalReplace(). As a result, the number of replacements returned was
1960        double what it should be. I removed one of the increments, but Craig sent a
1961        later patch that removed the other one (the right fix) and added unit tests
1962        that check the return values (which was not done before).
1963    
1964    21. Several CMake things:
1965    
1966        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1967            the names libpcre and libpcreposix, not just pcre and pcreposix.
1968    
1969        (2) The above change means that pcretest and pcregrep are now correctly
1970            linked with the newly-built libraries, not previously installed ones.
1971    
1972        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1973    
1974    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1975        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1976        UTF-8 newline character). The key issue is that the pattern starts .*;
1977        this means that the match must be either at the beginning, or after a
1978        newline. The bug was in the code for advancing after a failed match and
1979        checking that the new position followed a newline. It was not taking
1980        account of UTF-8 characters correctly.
1981    
1982    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1983        character classes. PCRE was not treating the sequence [:...:] as a
1984        character class unless the ... were all letters. Perl, however, seems to
1985        allow any characters between [: and :], though of course it rejects as
1986        unknown any "names" that contain non-letters, because all the known class
1987        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1988        for example, whereas PCRE did not - it did not recognize a POSIX character
1989        class. This seemed a bit dangerous, so the code has been changed to be
1990        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1991        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1992        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1993        Perl does, and where it didn't before.
1994    
1995    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1996        Windows environments %n is disabled by default.
1997    
1998    
1999    Version 7.4 21-Sep-07
2000    ---------------------
2001    
2002    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2003        means that a class such as [\s] counted as "explicit reference to CR or
2004        LF". That isn't really right - the whole point of the change was to try to
2005        help when there was an actual mention of one of the two characters. So now
2006        the change happens only if \r or \n (or a literal CR or LF) character is
2007        encountered.
2008    
2009    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2010        of both had grown to the point where there were only 3 bits left.
2011        Fortunately, there was spare space in the data structure, and so I have
2012        moved the internal flags into a new 16-bit field to free up more option
2013        bits.
2014    
2015    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2016        but did not set the internal JCHANGED flag - either of these is enough to
2017        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2018        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2019        start sets both bits.
2020    
2021    4.  Added options (at build time, compile time, exec time) to change \R from
2022        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2023    
2024    5.  doc/pcresyntax.html was missing from the distribution.
2025    
2026    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2027        compatibility, even though it is no longer used.
2028    
2029    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2030        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2031        windows.h file is present (where different names are used). [This was
2032        reversed later after testing - see 16 below.]
2033    
2034    8.  Changed all #include <config.h> to #include "config.h". There were also
2035        some further <pcre.h> cases that I changed to "pcre.h".
2036    
2037    9.  When pcregrep was used with the --colour option, it missed the line ending
2038        sequence off the lines that it output.
2039    
2040    10. It was pointed out to me that arrays of string pointers cause lots of
2041        relocations when a shared library is dynamically loaded. A technique of
2042        using a single long string with a table of offsets can drastically reduce
2043        these. I have refactored PCRE in four places to do this. The result is
2044        dramatic:
2045    
2046          Originally:                          290
2047          After changing UCP table:            187
2048          After changing error message table:   43
2049          After changing table of "verbs"       36
2050          After changing table of Posix names   22
2051    
2052        Thanks to the folks working on Gregex for glib for this insight.
2053    
2054    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2055        unicode-properties was also set.
2056    
2057    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2058    
2059    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2060        checked only for CRLF.
2061    
2062    14. Added casts to pcretest.c to avoid compiler warnings.
2063    
2064    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2065    
2066    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2067        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2068        entirely. This removes changes made in 7 above.
2069    
2070    17. The CMake files have been updated, and there is now more information about
2071        building with CMake in the NON-UNIX-USE document.
2072    
2073    
2074    Version 7.3 28-Aug-07
2075    ---------------------
2076    
2077     1. In the rejigging of the build system that eventually resulted in 7.1, the
2078        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2079        brackets there is not right, since it causes compilers to look for an
2080        installed pcre.h, not the version that is in the source that is being
2081        compiled (which of course may be different). I have changed it back to:
2082    
2083          #include "pcre.h"
2084    
2085        I have a vague recollection that the change was concerned with compiling in
2086        different directories, but in the new build system, that is taken care of
2087        by the VPATH setting the Makefile.
2088    
2089     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2090        when the subject happened to end in the byte 0x85 (e.g. if the last
2091        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2092        characters but of course it shouldn't be taken as a newline when it is part
2093        of another character. The bug was that, for an unlimited repeat of . in
2094        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2095        characters when looking for a newline.
2096    
2097     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2098    
2099     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2100        in debug output.
2101    
2102     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2103        long printing in the pcrecpp unittest when running under MinGW.
2104    
2105     6. ESC_K was left out of the EBCDIC table.
2106    
2107     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2108        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2109        limit also applies to "virtual nesting" when a pattern is recursive, and in
2110        this case 1000 isn't so big. I have been able to remove this limit at the
2111        expense of backing off one optimization in certain circumstances. Normally,
2112        when pcre_exec() would call its internal match() function recursively and
2113        immediately return the result unconditionally, it uses a "tail recursion"
2114        feature to save stack. However, when a subpattern that can match an empty
2115        string has an unlimited repetition quantifier, it no longer makes this
2116        optimization. That gives it a stack frame in which to save the data for
2117        checking that an empty string has been matched. Previously this was taken
2118        from the 1000-entry workspace that had been reserved. So now there is no
2119        explicit limit, but more stack is used.
2120    
2121     8. Applied Daniel's patches to solve problems with the import/export magic
2122        syntax that is required for Windows, and which was going wrong for the
2123        pcreposix and pcrecpp parts of the library. These were overlooked when this
2124        problem was solved for the main library.
2125    
2126     9. There were some crude static tests to avoid integer overflow when computing
2127        the size of patterns that contain repeated groups with explicit upper
2128        limits. As the maximum quantifier is 65535, the maximum group length was
2129        set at 30,000 so that the product of these two numbers did not overflow a
2130        32-bit integer. However, it turns out that people want to use groups that
2131        are longer than 30,000 bytes (though not repeat them that many times).
2132        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2133        made it possible to implement the integer overflow checks in a much more
2134        dynamic way, which I have now done. The artificial limitation on group
2135        length has been removed - we now have only the limit on the total length of
2136        the compiled pattern, which depends on the LINK_SIZE setting.
2137    
2138    10. Fixed a bug in the documentation for get/copy named substring when
2139        duplicate names are permitted. If none of the named substrings are set, the
2140        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2141        empty string.
2142    
2143    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2144        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2145        because the ] is interpreted as the first data character and the
2146        terminating ] is not found. PCRE has been made compatible with Perl in this
2147        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2148        cause memory overwriting.
2149    
2150    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2151        string has been matched (to stop an infinite loop). It was not recognizing
2152        a conditional subpattern that could match an empty string if that
2153        subpattern was within another subpattern. For example, it looped when
2154        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2155        condition was not nested. This bug has been fixed.
2156    
2157    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2158        past the start of the subject in the presence of bytes with the top bit
2159        set, for example "\x8aBCD".
2160    
2161    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2162        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2163    
2164    14. Optimized (?!) to (*FAIL).
2165    
2166    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2167        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2168        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2169        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2170        does: it's just the validity check that is more restrictive.
2171    
2172    16. Inserted checks for integer overflows during escape sequence (backslash)
2173        processing, and also fixed erroneous offset values for syntax errors during
2174        backslash processing.
2175    
2176    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2177        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2178    
2179    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2180        caused an overrun.
2181    
2182    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2183        something other than just ASCII characters) inside a group that had an
2184        unlimited repeat caused a loop at compile time (while checking to see
2185        whether the group could match an empty string).
2186    
2187    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2188        [\P{Any}] did so. (Error in the code for printing property names.)
2189    
2190    21. An orphan \E inside a character class could cause a crash.
2191    
2192    22. A repeated capturing bracket such as (A)? could cause a wild memory
2193        reference during compilation.
2194    
2195    23. There are several functions in pcre_compile() that scan along a compiled
2196        expression for various reasons (e.g. to see if it's fixed length for look
2197        behind). There were bugs in these functions when a repeated \p or \P was
2198        present in the pattern. These operators have additional parameters compared
2199        with \d, etc, and these were not being taken into account when moving along
2200        the compiled data. Specifically:
2201    
2202        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2203            length.
2204    
2205        (b) An item such as \pL+ within a repeated group could cause crashes or
2206            loops.
2207    
2208        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2209            "reference to non-existent subpattern" error.
2210    
2211        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2212    
2213    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2214        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2215    
2216    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2217    
2218    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2219        character were causing crashes (broken optimization).
2220    
2221    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2222        \p or \P) caused a compile-time loop.
2223    
2224    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2225        break. For example, the unstudied pattern [\r\n]A does not match the string
2226        "\r\nA" because change 7.0/46 below moves the current point on by two
2227        characters after failing to match at the start. However, the pattern \nA
2228        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2229        the same is true. There doesn't seem any very clean way out of this, but
2230        what I have chosen to do makes the common cases work: PCRE now takes note
2231        of whether there can be an explicit match for \r or \n anywhere in the
2232        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2233        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2234        pattern has explicit CR or LF references.
2235    
2236    29. Added (*CR) etc for changing newline setting at start of pattern.
2237    
2238    
2239    Version 7.2 19-Jun-07
2240    ---------------------
2241    
2242     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2243        which is apparently normally available under Windows.
2244    
2245     2. Re-jig the pcregrep tests with different newline settings in an attempt
2246        to make them independent of the local environment's newline setting.
2247    
2248     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2249    
2250     4. Some of the "internals" tests were previously cut out when the link size
2251        was not 2, because the output contained actual offsets. The recent new
2252        "Z" feature of pcretest means that these can be cut out, making the tests
2253        usable with all link sizes.
2254    
2255     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2256        stack recursion. This gives a massive performance boost under BSD, but just
2257        a small improvement under Linux. However, it saves one field in the frame
2258        in all cases.
2259    
2260     6. Added more features from the forthcoming Perl 5.10:
2261    
2262        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2263            recursion call. It refers to the nth most recently opened parentheses.
2264    
2265        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2266            to be opened parentheses.
2267    
2268        (c) Conditions that refer to capturing parentheses can be specified
2269            relatively, for example, (?(-2)... or (?(+3)...
2270    
2271        (d) \K resets the start of the current match so that everything before
2272            is not part of it.
2273    
2274        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2275    
2276        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2277            reference syntax.
2278    
2279        (g) (?| introduces a group in which the numbering of parentheses in each
2280            alternative starts with the same number.
2281    
2282        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2283    
2284     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2285        PCRE_INFO_JCHANGED.
2286    
2287     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2288        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2289        for detecting groups that can match an empty string.
2290    
2291     9. A pattern with a very large number of alternatives (more than several
2292        hundred) was running out of internal workspace during the pre-compile
2293        phase, where pcre_compile() figures out how much memory will be needed. A
2294        bit of new cunning has reduced the workspace needed for groups with
2295        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2296        workspace instead of running out of the 4096 that are available.
2297    
2298    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2299    
2300    11. Applied patch from Google to remove an optimization that didn't quite work.
2301        The report of the bug said:
2302    
2303          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2304          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2305          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2306    
2307    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2308        it matched the wrong number of bytes.
2309    
2310    
2311    Version 7.1 24-Apr-07
2312  ---------------------  ---------------------
2313    
2314   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
# Line 33  Version 7.1 12-Mar-07 Line 2340  Version 7.1 12-Mar-07
2340   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2341      Weber: (1) pcre_internal.h was missing some function renames; (2) updated      Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2342      makevp.bat for the current PCRE, using the additional files      makevp.bat for the current PCRE, using the additional files
2343      makevp-c.txt, makevp-l.txt, and pcregexp.pas.      makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2344    
2345   6. A Windows user reported a minor discrepancy with test 2, which turned out   6. A Windows user reported a minor discrepancy with test 2, which turned out
2346      to be caused by a trailing space on an input line that had got lost in his      to be caused by a trailing space on an input line that had got lost in his
# Line 115  Version 7.1 12-Mar-07 Line 2422  Version 7.1 12-Mar-07
2422    
2423  14. If erroroffset was passed as NULL to pcre_compile, it provoked a  14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2424      segmentation fault instead of returning the appropriate error message.      segmentation fault instead of returning the appropriate error message.
2425    
2426  15. In multiline mode when the newline sequence was set to "any", the pattern  15. In multiline mode when the newline sequence was set to "any", the pattern
2427      ^$ would give a match between the \r and \n of a subject such as "A\r\nB".      ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2428      This doesn't seem right; it now treats the CRLF combination as the line      This doesn't seem right; it now treats the CRLF combination as the line
2429      ending, and so does not match in that case. It's only a pattern such as ^$      ending, and so does not match in that case. It's only a pattern such as ^$
2430      that would hit this one: something like ^ABC$ would have failed after \r      that would hit this one: something like ^ABC$ would have failed after \r
2431      and then tried again after \r\n.      and then tried again after \r\n.
2432    
2433  16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"  16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2434      in an attempt to make files that differ only in their line terminators      in an attempt to make files that differ only in their line terminators
2435      compare equal. This works on Linux.      compare equal. This works on Linux.
2436    
2437    17. Under certain error circumstances pcregrep might try to free random memory
2438        as it exited. This is now fixed, thanks to valgrind.
2439    
2440    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2441        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2442        was because its rules for how to advance for /g after matching an empty
2443        string at the end of a line did not allow for this case. They now check for
2444        it specially.
2445    
2446    20. pcretest is supposed to handle patterns and data of any length, by
2447        extending its buffers when necessary. It was getting this wrong when the
2448        buffer for a data line had to be extended.
2449    
2450    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2451        CRLF as a newline sequence.
2452    
2453    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2454        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2455        I have nevertheless tidied it up.
2456    
2457    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2458    
2459    24. Added a man page for pcre-config.
2460    
2461    
2462  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06

Legend:
Removed from v.131  
changed lines
  Added in v.1239

  ViewVC Help
Powered by ViewVC 1.1.5