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

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 1249 by zherczeg, Mon Feb 18 09:55:43 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    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    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33        and later.
34    
35    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36    
37    9.  Optimizing fast_forward_start_bits in JIT.
38    
39    10. Adding experimental support for callouts in JIT, and fixing some
40        issues revealed during this work. Namely:
41    
42        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43    
44        (b) Minimum length was not checked before the matching is started.
45    
46    11. The value of capture_last that is passed to callouts was incorrect in some
47        cases when there was a capture on one path that was subsequently abandoned
48        after a backtrack. Also, the capture_last value is now reset after a
49        recursion, since all captures are also reset in this case.
50    
51    12. The interpreter no longer returns the "too many substrings" error in the
52        case when an overflowing capture is in a branch that is subsequently
53        abandoned after a backtrack.
54    
55    13. In the pathological case when an offset vector of size 2 is used, pcretest
56        now prints out the matched string after a yield of 0 or 1.
57    
58    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59        Only supported by the JIT compiler at the moment.
60    
61    
62    Version 8.32 30-November-2012
63    -----------------------------
64    
65    1.  Improved JIT compiler optimizations for first character search and single
66        character iterators.
67    
68    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
69        Patch by Daniel Richard G.
70    
71    3.  Single character iterator optimizations in the JIT compiler.
72    
73    4.  Improved JIT compiler optimizations for character ranges.
74    
75    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
76        Reported by Giuseppe D'Angelo.
77    
78    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
79        of a line, was being set incorrectly in cases where .* appeared inside
80        atomic brackets at the start of a pattern, or where there was a subsequent
81        *PRUNE or *SKIP.
82    
83    7.  Improved instruction cache flush for POWER/PowerPC.
84        Patch by Daniel Richard G.
85    
86    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
87        grep:
88    
89        (a) There is now no limit to the number of patterns to be matched.
90    
91        (b) An error is given if a pattern is too long.
92    
93        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
94            are now supported.
95    
96        (d) --exclude-from and --include-from (multiple use) have been added.
97    
98        (e) Exclusions and inclusions now apply to all files and directories, not
99            just to those obtained from scanning a directory recursively.
100    
101        (f) Multiple uses of -f and --file-list are now supported.
102    
103        (g) In a Windows environment, the default for -d has been changed from
104            "read" (the GNU grep default) to "skip", because otherwise the presence
105            of a directory in the file list provokes an error.
106    
107        (h) The documentation has been revised and clarified in places.
108    
109    9.  Improve the matching speed of capturing brackets.
110    
111    10. Changed the meaning of \X so that it now matches a Unicode extended
112        grapheme cluster.
113    
114    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
115        out POSIX threads when JIT support is configured.
116    
117    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
118    
119    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
120        structure could go wrong in environments where size_t is not the same size
121        as int.
122    
123    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
124        set.
125    
126    15. The EBCDIC support had decayed; later updates to the code had included
127        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
128        general tidy up of EBCDIC-related issues, and the documentation was also
129        not quite right. There is now a test that can be run on ASCII systems to
130        check some of the EBCDIC-related things (but is it not a full test).
131    
132    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
133        in a small tidy to the code.
134    
135    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
136    
137    18. If the --only-matching (-o) option in pcregrep is specified multiple
138        times, each one causes appropriate output. For example, -o1 -o2 outputs the
139        substrings matched by the 1st and 2nd capturing parentheses. A separating
140        string can be specified by --om-separator (default empty).
141    
142    19. Improving the first n character searches.
143    
144    20. Turn case lists for horizontal and vertical white space into macros so that
145        they are defined only once.
146    
147    21. This set of changes together give more compatible Unicode case-folding
148        behaviour for characters that have more than one other case when UCP
149        support is available.
150    
151        (a) The Unicode property table now has offsets into a new table of sets of
152            three or more characters that are case-equivalent. The MultiStage2.py
153            script that generates these tables (the pcre_ucd.c file) now scans
154            CaseFolding.txt instead of UnicodeData.txt for character case
155            information.
156    
157        (b) The code for adding characters or ranges of characters to a character
158            class has been abstracted into a generalized function that also handles
159            case-independence. In UTF-mode with UCP support, this uses the new data
160            to handle characters with more than one other case.
161    
162        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
163            whose other case is greater than 256 are now correctly matched
164            caselessly. Previously, the high codepoint matched the low one, but not
165            vice versa.
166    
167        (d) The processing of \h, \H, \v, and \ in character classes now makes use
168            of the new class addition function, using character lists defined as
169            macros alongside the case definitions of 20 above.
170    
171        (e) Caseless back references now work with characters that have more than
172            one other case.
173    
174        (f) General caseless matching of characters with more than one other case
175            is supported.
176    
177    22. Unicode character properties were updated from Unicode 6.2.0
178    
179    23. Improved CMake support under Windows. Patch by Daniel Richard G.
180    
181    24. Add support for 32-bit character strings, and UTF-32
182    
183    25. Major JIT compiler update (code refactoring and bugfixing).
184        Experimental Sparc 32 support is added.
185    
186    26. Applied a modified version of Daniel Richard G's patch to create
187        pcre.h.generic and config.h.generic by "make" instead of in the
188        PrepareRelease script.
189    
190    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
191        pcre_compile.c when checking for a zero character.
192    
193    28. Introducing a native interface for JIT. Through this interface, the compiled
194        machine code can be directly executed. The purpose of this interface is to
195        provide fast pattern matching, so several sanity checks are not performed.
196        However, feature tests are still performed. The new interface provides
197        1.4x speedup compared to the old one.
198    
199    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
200        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
201        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
202    
203    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
204        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
205        the "old" RFC 2279). Instead, it ended up passing a negative length to
206        pcre_exec().
207    
208    31. Add support for GCC's visibility feature to hide internal functions.
209    
210    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
211        "unknown -C option" after outputting 0 or 1.
212    
213    33. There is now support for generating a code coverage report for the test
214        suite in environments where gcc is the compiler and lcov is installed. This
215        is mainly for the benefit of the developers.
216    
217    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
218        unaddressable using valgrind annotations, allowing valgrind to detect
219        invalid memory accesses. This is mainly for the benefit of the developers.
220    
221    25. (*UTF) can now be used to start a pattern in any of the three libraries.
222    
223    26. Give configure error if --enable-cpp but no C++ compiler found.
224    
225    
226    Version 8.31 06-July-2012
227    -------------------------
228    
229    1.  Fixing a wrong JIT test case and some compiler warnings.
230    
231    2.  Removed a bashism from the RunTest script.
232    
233    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
234        to unsigned type, result still unsigned" that was given by an MS compiler
235        on encountering the code "-sizeof(xxx)".
236    
237    4.  Partial matching support is added to the JIT compiler.
238    
239    5.  Fixed several bugs concerned with partial matching of items that consist
240        of more than one character:
241    
242        (a) /^(..)\1/ did not partially match "aba" because checking references was
243            done on an "all or nothing" basis. This also applied to repeated
244            references.
245    
246        (b) \R did not give a hard partial match if \r was found at the end of the
247            subject.
248    
249        (c) \X did not give a hard partial match after matching one or more
250            characters at the end of the subject.
251    
252        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
253            a partial match for the string "\r".
254    
255        (e) When newline was set to CRLF, the metacharacter "." did not recognize
256            a partial match for a CR character at the end of the subject string.
257    
258    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
259        running pcretest, the text "(JIT)" added to the output whenever JIT is
260        actually used to run the match.
261    
262    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
263        or /S+[+] with a digit between 1 and 7.
264    
265    8.  OP_NOT now supports any UTF character not just single-byte ones.
266    
267    9.  (*MARK) control verb is now supported by the JIT compiler.
268    
269    10. The command "./RunTest list" lists the available tests without actually
270        running any of them. (Because I keep forgetting what they all are.)
271    
272    11. Add PCRE_INFO_MAXLOOKBEHIND.
273    
274    12. Applied a (slightly modified) user-supplied patch that improves performance
275        when the heap is used for recursion (compiled with --disable-stack-for-
276        recursion). Instead of malloc and free for each heap frame each time a
277        logical recursion happens, frames are retained on a chain and re-used where
278        possible. This sometimes gives as much as 30% improvement.
279    
280    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
281        call.
282    
283    14. As documented, (*COMMIT) is now confined to within a positive assertion.
284    
285    15. It is now possible to link pcretest with libedit as an alternative to
286        libreadline.
287    
288    16. (*COMMIT) control verb is now supported by the JIT compiler.
289    
290    17. The Unicode data tables have been updated to Unicode 6.1.0.
291    
292    18. Added --file-list option to pcregrep.
293    
294    19. Added binary file support to pcregrep, including the -a, --binary-files,
295        -I, and --text options.
296    
297    20. The madvise function is renamed for posix_madvise for QNX compatibility
298        reasons. Fixed by Giuseppe D'Angelo.
299    
300    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
301    
302    22. Changed the option for creating symbolic links for 16-bit man pages from
303        -s to -sf so that re-installing does not cause issues.
304    
305    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
306    
307    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
308        matches in certain environments (the workspace was not being correctly
309        retained). Also added to pcre_dfa_exec() a simple plausibility check on
310        some of the workspace data at the beginning of a restart.
311    
312    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
313        was not doing so when it should - probably a typo introduced by SVN 528
314        (change 8.10/14).
315    
316    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
317        \w+ when the character tables indicated that \x{c4} was a word character.
318        There were several related cases, all because the tests for doing a table
319        lookup were testing for characters less than 127 instead of 255.
320    
321    27. If a pattern contains capturing parentheses that are not used in a match,
322        their slots in the ovector are set to -1. For those that are higher than
323        any matched groups, this happens at the end of processing. In the case when
324        there were back references that the ovector was too small to contain
325        (causing temporary malloc'd memory to be used during matching), and the
326        highest capturing number was not used, memory off the end of the ovector
327        was incorrectly being set to -1. (It was using the size of the temporary
328        memory instead of the true size.)
329    
330    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
331        ovector size, it uses memory at the end of the block that it has got.
332    
333    29. Check for an overlong MARK name and give an error at compile time. The
334        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
335    
336    30. JIT compiler update.
337    
338    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
339        Rill for the patch.
340    
341    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
342    
343    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
344    
345    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
346        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
347        was enabled.
348    
349    35. Improve JIT code generation for greedy plus quantifier.
350    
351    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
352        1 to "aa" instead of to an empty string. The bug affected repeated groups
353        that could potentially match an empty string.
354    
355    37. Optimizing single character iterators in JIT.
356    
357    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
358        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
359        codepoints that are too big for the mode are faulted, and in a UTF mode,
360        disallowed codepoints are also faulted.
361    
362    39. If PCRE was compiled with UTF support, in three places in the DFA
363        matcher there was code that should only have been obeyed in UTF mode, but
364        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
365        processing when bytes with values greater than 127 were present. In 16-bit
366        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
367        both cases the values are those that cannot be the first data item in a UTF
368        character. The three items that might have provoked this were recursions,
369        possessively repeated groups, and atomic groups.
370    
371    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
372        and pcregrep, because some OS require shared objects to be explicitly
373        passed to ld, causing the link step to fail if they are not.
374    
375    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
376        mode, patterns that started with \h* or \R* might be incorrectly matched.
377    
378    
379    Version 8.30 04-February-2012
380    -----------------------------
381    
382    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
383        name is defined in ctype.h.
384    
385    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
386        only in quite long subpatterns.
387    
388    3.  Removed the function pcre_info(), which has been obsolete and deprecated
389        since it was replaced by pcre_fullinfo() in February 2000.
390    
391    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
392        match a (*MARK), and the match failed at the start of the subject, a
393        reference to memory before the start of the subject could occur. This bug
394        was introduced by fix 17 of release 8.21.
395    
396    5.  A reference to an unset group with zero minimum repetition was giving
397        totally wrong answers (in non-JavaScript-compatibility mode). For example,
398        /(another)?(\1?)test/ matched against "hello world test". This bug was
399        introduced in release 8.13.
400    
401    6.  Add support for 16-bit character strings (a large amount of work involving
402        many changes and refactorings).
403    
404    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
405        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
406        from a file.
407    
408    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
409        rounding is not applied in this particular case).
410    
411    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
412        if they appear, or are escaped, in patterns.
413    
414    10. Get rid of a number of -Wunused-but-set-variable warnings.
415    
416    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
417        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
418        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
419        also returns the mark "x". This bug applied to capturing parentheses,
420        non-capturing parentheses, and atomic parentheses. It also applied to some
421        assertions.
422    
423    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
424        information out of configure.ac instead of relying on pcre.h.generic, which
425        is not stored in the repository.
426    
427    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
428        -lreadline.
429    
430    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
431    
432    15. Applied Graycode's patch to put the top-level frame on the stack rather
433        than the heap when not using the stack for recursion. This gives a
434        performance improvement in many cases when recursion is not deep.
435    
436    16. Experimental code added to "pcretest -C" to output the stack frame size.
437    
438    
439    Version 8.21 12-Dec-2011
440    ------------------------
441    
442    1.  Updating the JIT compiler.
443    
444    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
445        are added as well.
446    
447    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
448        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
449        calling _pcre_jit_exec. Some extra comments are added.
450    
451    4.  (*MARK) settings inside atomic groups that do not contain any capturing
452        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
453        was introduced by change 18 for 8.20.
454    
455    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
456        ECMA-262 standard.
457    
458    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
459        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
460        This bug was probably introduced by change 9 of 8.13.
461    
462    7.  While fixing 6 above, I noticed that a number of other items were being
463        incorrectly rejected as "not fixed length". This arose partly because newer
464        opcodes had not been added to the fixed-length checking code. I have (a)
465        corrected the bug and added tests for these items, and (b) arranged for an
466        error to occur if an unknown opcode is encountered while checking for fixed
467        length instead of just assuming "not fixed length". The items that were
468        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
469        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
470        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
471    
472    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
473        being incorrectly compiled and would have given unpredicatble results.
474    
475    9.  A possessively repeated subpattern with minimum repeat count greater than
476        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
477        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
478        the first (A) could occur when it should not.
479    
480    10. Add a cast and remove a redundant test from the code.
481    
482    11. JIT should use pcre_malloc/pcre_free for allocation.
483    
484    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
485        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
486        is anything other than /usr, -L is still shown).
487    
488    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
489    
490    14. Perl does not support \N without a following name in a [] class; PCRE now
491        also gives an error.
492    
493    15. If a forward reference was repeated with an upper limit of around 2000,
494        it caused the error "internal error: overran compiling workspace". The
495        maximum number of forward references (including repeats) was limited by the
496        internal workspace, and dependent on the LINK_SIZE. The code has been
497        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
498        the default depends on LINK_SIZE. There is a new upper limit (for safety)
499        of around 200,000 forward references. While doing this, I also speeded up
500        the filling in of repeated forward references.
501    
502    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
503        incorrectly expecting the subject to contain another "a" after the start.
504    
505    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
506        in the match, the SKIP should be ignored. This was not happening; instead
507        the SKIP was being treated as NOMATCH. For patterns such as
508        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
509        tested.
510    
511    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
512        now much more compatible with Perl, in particular in cases where the result
513        is a non-match for a non-anchored pattern. For example, if
514        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
515        "m", where previously it did not return a name. A side effect of this
516        change is that for partial matches, the last encountered mark name is
517        returned, as for non matches. A number of tests that were previously not
518        Perl-compatible have been moved into the Perl-compatible test files. The
519        refactoring has had the pleasing side effect of removing one argument from
520        the match() function, thus reducing its stack requirements.
521    
522    19. If the /S+ option was used in pcretest to study a pattern using JIT,
523        subsequent uses of /S (without +) incorrectly behaved like /S+.
524    
525    21. Retrieve executable code size support for the JIT compiler and fixing
526        some warnings.
527    
528    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
529        not work when the shorter character appeared right at the end of the
530        subject string.
531    
532    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
533        systems.
534    
535    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
536        output it when the /M option is used in pcretest.
537    
538    25. The CheckMan script was not being included in the distribution. Also, added
539        an explicit "perl" to run Perl scripts from the PrepareRelease script
540        because this is reportedly needed in Windows.
541    
542    26. If study data was being save in a file and studying had not found a set of
543        "starts with" bytes for the pattern, the data written to the file (though
544        never used) was taken from uninitialized memory and so caused valgrind to
545        complain.
546    
547    27. Updated RunTest.bat as provided by Sheri Pierce.
548    
549    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
550    
551    29. Computation of memory usage for the table of capturing group names was
552        giving an unnecessarily large value.
553    
554    
555    Version 8.20 21-Oct-2011
556    ------------------------
557    
558    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
559        a POSIX class. After further experiments with Perl, which convinced me that
560        Perl has bugs and confusions, a closing square bracket is no longer allowed
561        in a POSIX name. This bug also affected patterns with classes that started
562        with full stops.
563    
564    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
565        captured substring, but while checking the failing first alternative,
566        substring 1 is temporarily captured. If the output vector supplied to
567        pcre_exec() was not big enough for this capture, the yield of the function
568        was still zero ("insufficient space for captured substrings"). This cannot
569        be totally fixed without adding another stack variable, which seems a lot
570        of expense for a edge case. However, I have improved the situation in cases
571        such as /(a)(b)x|abc/ matched against "abc", where the return code
572        indicates that fewer than the maximum number of slots in the ovector have
573        been set.
574    
575    3.  Related to (2) above: when there are more back references in a pattern than
576        slots in the output vector, pcre_exec() uses temporary memory during
577        matching, and copies in the captures as far as possible afterwards. It was
578        using the entire output vector, but this conflicts with the specification
579        that only 2/3 is used for passing back captured substrings. Now it uses
580        only the first 2/3, for compatibility. This is, of course, another edge
581        case.
582    
583    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
584        main code base, and can be used by building with --enable-jit. When this is
585        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
586        runtime --no-jit option is given.
587    
588    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
589        ovector, the return from the function was zero, implying that there were
590        other matches that did not fit. The correct "exactly full" value is now
591        returned.
592    
593    6.  If a subpattern that was called recursively or as a subroutine contained
594        (*PRUNE) or any other control that caused it to give a non-standard return,
595        invalid errors such as "Error -26 (nested recursion at the same subject
596        position)" or even infinite loops could occur.
597    
598    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
599        computing the minimum length on reaching *ACCEPT, and so ended up with the
600        wrong value of 1 rather than 0. Further investigation indicates that
601        computing a minimum subject length in the presence of *ACCEPT is difficult
602        (think back references, subroutine calls), and so I have changed the code
603        so that no minimum is registered for a pattern that contains *ACCEPT.
604    
605    8.  If (*THEN) was present in the first (true) branch of a conditional group,
606        it was not handled as intended. [But see 16 below.]
607    
608    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
609        Sheri Pierce.
610    
611    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
612        the first byte in a match must be "a".
613    
614    11. Change 17 for 8.13 increased the recursion depth for patterns like
615        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
616        pattern contains any instances of (*THEN). If it does not, the old
617        optimizations are restored. It would be nice to do this on a per-group
618        basis, but at the moment that is not feasible.
619    
620    12. In some environments, the output of pcretest -C is CRLF terminated. This
621        broke RunTest's code that checks for the link size. A single white space
622        character after the value is now allowed for.
623    
624    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
625        For "fr", it uses the Windows-specific input and output files.
626    
627    14. If (*THEN) appeared in a group that was called recursively or as a
628        subroutine, it did not work as intended. [But see next item.]
629    
630    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
631        pattern fragments (but not containing any | characters). If A and B are
632        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
633        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
634        D. In other words, Perl considers parentheses that do not contain any |
635        characters to be part of a surrounding alternative, whereas PCRE was
636        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
637        differently. PCRE now behaves in the same way as Perl, except in the case
638        of subroutine/recursion calls such as (?1) which have in any case always
639        been different (but PCRE had them first :-).
640    
641    16. Related to 15 above: Perl does not treat the | in a conditional group as
642        creating alternatives. Such a group is treated in the same way as an
643        ordinary group without any | characters when processing (*THEN). PCRE has
644        been changed to match Perl's behaviour.
645    
646    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
647        RunGrepTest script failed.
648    
649    18. Change 22 for version 13 caused atomic groups to use more stack. This is
650        inevitable for groups that contain captures, but it can lead to a lot of
651        stack use in large patterns. The old behaviour has been restored for atomic
652        groups that do not contain any capturing parentheses.
653    
654    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
655        suppress the check for a minimum subject length at run time. (If it was
656        given to pcre_exec() or pcre_dfa_exec() it did work.)
657    
658    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
659        fail to work when decoding hex characters in data strings in EBCDIC
660        environments.
661    
662    21. It appears that in at least one Mac OS environment, the isxdigit() function
663        is implemented as a macro that evaluates to its argument more than once,
664        contravening the C 90 Standard (I haven't checked a later standard). There
665        was an instance in pcretest which caused it to go wrong when processing
666        \x{...} escapes in subject strings. The has been rewritten to avoid using
667        things like p++ in the argument of isxdigit().
668    
669    
670    Version 8.13 16-Aug-2011
671    ------------------------
672    
673    1.  The Unicode data tables have been updated to Unicode 6.0.0.
674    
675    2.  Two minor typos in pcre_internal.h have been fixed.
676    
677    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
678        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
679        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
680    
681    4.  There were a number of related bugs in the code for matching backrefences
682        caselessly in UTF-8 mode when codes for the characters concerned were
683        different numbers of bytes. For example, U+023A and U+2C65 are an upper
684        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
685        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
686        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
687        2-byte code at the end of the subject (it thought there wasn't enough data
688        left).
689    
690    5.  Comprehensive information about what went wrong is now returned by
691        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
692        as the output vector has at least 2 elements. The offset of the start of
693        the failing character and a reason code are placed in the vector.
694    
695    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
696        now returned is for the first byte of the failing character, instead of the
697        last byte inspected. This is an incompatible change, but I hope it is small
698        enough not to be a problem. It makes the returned offset consistent with
699        pcre_exec() and pcre_dfa_exec().
700    
701    7.  pcretest now gives a text phrase as well as the error number when
702        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
703        failure, the offset and reason code are output.
704    
705    8.  When \R was used with a maximizing quantifier it failed to skip backwards
706        over a \r\n pair if the subsequent match failed. Instead, it just skipped
707        back over a single character (\n). This seems wrong (because it treated the
708        two characters as a single entity when going forwards), conflicts with the
709        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
710        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
711        has been changed.
712    
713    9.  Some internal refactoring has changed the processing so that the handling
714        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
715        time (the PCRE_DOTALL option was changed this way some time ago: version
716        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
717        which was always a bit of a fudge. It also means that there is one less
718        argument for the match() function, which reduces its stack requirements
719        slightly. This change also fixes an incompatibility with Perl: the pattern
720        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
721    
722    10. More internal refactoring has drastically reduced the number of recursive
723        calls to match() for possessively repeated groups such as (abc)++ when
724        using pcre_exec().
725    
726    11. While implementing 10, a number of bugs in the handling of groups were
727        discovered and fixed:
728    
729        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
730        (a|)*(?1) gave a compile-time internal error.
731        ((a|)+)+  did not notice that the outer group could match an empty string.
732        (^a|^)+   was not marked as anchored.
733        (.*a|.*)+ was not marked as matching at start or after a newline.
734    
735    12. Yet more internal refactoring has removed another argument from the match()
736        function. Special calls to this function are now indicated by setting a
737        value in a variable in the "match data" data block.
738    
739    13. Be more explicit in pcre_study() instead of relying on "default" for
740        opcodes that mean there is no starting character; this means that when new
741        ones are added and accidentally left out of pcre_study(), testing should
742        pick them up.
743    
744    14. The -s option of pcretest has been documented for ages as being an old
745        synonym of -m (show memory usage). I have changed it to mean "force study
746        for every regex", that is, assume /S for every regex. This is similar to -i
747        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
748        using it. It makes it easier to run collections of tests with and without
749        study enabled, and thereby test pcre_study() more easily. All the standard
750        tests are now run with and without -s (but some patterns can be marked as
751        "never study" - see 20 below).
752    
753    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
754        restoration of the capturing data to the outer values was not happening
755        correctly.
756    
757    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
758        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
759        pattern had matched an empty string, and so incorrectly returned a no
760        match.
761    
762    17. There was optimizing code for the last branch of non-capturing parentheses,
763        and also for the obeyed branch of a conditional subexpression, which used
764        tail recursion to cut down on stack usage. Unfortunately, now that there is
765        the possibility of (*THEN) occurring in these branches, tail recursion is
766        no longer possible because the return has to be checked for (*THEN). These
767        two optimizations have therefore been removed. [But see 8.20/11 above.]
768    
769    18. If a pattern containing \R was studied, it was assumed that \R always
770        matched two bytes, thus causing the minimum subject length to be
771        incorrectly computed because \R can also match just one byte.
772    
773    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
774        was incorrectly computed.
775    
776    20. If /S is present twice on a test pattern in pcretest input, it now
777        *disables* studying, thereby overriding the use of -s on the command line
778        (see 14 above). This is necessary for one or two tests to keep the output
779        identical in both cases.
780    
781    21. When (*ACCEPT) was used in an assertion that matched an empty string and
782        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
783    
784    22. When an atomic group that contained a capturing parenthesis was
785        successfully matched, but the branch in which it appeared failed, the
786        capturing was not being forgotten if a higher numbered group was later
787        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
788        group 1 to "a", when in fact it should be unset. This applied to multi-
789        branched capturing and non-capturing groups, repeated or not, and also to
790        positive assertions (capturing in negative assertions does not happen
791        in PCRE) and also to nested atomic groups.
792    
793    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
794        subject after a captured substring, to make it easier to tell which of a
795        number of identical substrings has been captured.
796    
797    24. The way atomic groups are processed by pcre_exec() has been changed so that
798        if they are repeated, backtracking one repetition now resets captured
799        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
800        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
801        "aaa". Previously, it would have been "a". As part of this code
802        refactoring, the way recursive calls are handled has also been changed.
803    
804    25. If an assertion condition captured any substrings, they were not passed
805        back unless some other capturing happened later. For example, if
806        (?(?=(a))a) was matched against "a", no capturing was returned.
807    
808    26. When studying a pattern that contained subroutine calls or assertions,
809        the code for finding the minimum length of a possible match was handling
810        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
811        group 1 called group 2 while simultaneously a separate group 2 called group
812        1). A stack overflow occurred in this case. I have fixed this by limiting
813        the recursion depth to 10.
814    
815    27. Updated RunTest.bat in the distribution to the version supplied by Tom
816        Fortmann. This supports explicit test numbers on the command line, and has
817        argument validation and error reporting.
818    
819    28. An instance of \X with an unlimited repeat could fail if at any point the
820        first character it looked at was a mark character.
821    
822    29. Some minor code refactoring concerning Unicode properties and scripts
823        should reduce the stack requirement of match() slightly.
824    
825    30. Added the '=' option to pcretest to check the setting of unused capturing
826        slots at the end of the pattern, which are documented as being -1, but are
827        not included in the return count.
828    
829    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
830        compiled something random. Now it gives a compile-time error (as does
831        Perl).
832    
833    32. A *MARK encountered during the processing of a positive assertion is now
834        recorded and passed back (compatible with Perl).
835    
836    33. If --only-matching or --colour was set on a pcregrep call whose pattern
837        had alternative anchored branches, the search for a second match in a line
838        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
839        matched the line "0102" twice. The same bug affected patterns that started
840        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
841        twice.
842    
843    34. Previously, PCRE did not allow quantification of assertions. However, Perl
844        does, and because of capturing effects, quantifying parenthesized
845        assertions may at times be useful. Quantifiers are now allowed for
846        parenthesized assertions.
847    
848    35. A minor code tidy in pcre_compile() when checking options for \R usage.
849    
850    36. \g was being checked for fancy things in a character class, when it should
851        just be a literal "g".
852    
853    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
854        appearance of a nested POSIX class supersedes an apparent external class.
855        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
856        unescaped square brackets may also appear as part of class names. For
857        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
858        more like Perl. (But see 8.20/1 above.)
859    
860    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
861        was because it thought it was \N{name}, which is not supported).
862    
863    39. Add minix to OS list not supporting the -S option in pcretest.
864    
865    40. PCRE tries to detect cases of infinite recursion at compile time, but it
866        cannot analyze patterns in sufficient detail to catch mutual recursions
867        such as ((?1))((?2)). There is now a runtime test that gives an error if a
868        subgroup is called recursively as a subpattern for a second time at the
869        same position in the subject string. In previous releases this might have
870        been caught by the recursion limit, or it might have run out of stack.
871    
872    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
873        happen only once. PCRE was, however incorrectly giving a compile time error
874        "recursive call could loop indefinitely" because it cannot analyze the
875        pattern in sufficient detail. The compile time test no longer happens when
876        PCRE is compiling a conditional subpattern, but actual runaway loops are
877        now caught at runtime (see 40 above).
878    
879    42. It seems that Perl allows any characters other than a closing parenthesis
880        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
881        has been changed to be the same.
882    
883    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
884        as not to get warnings when autogen.sh is called. Also changed
885        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
886    
887    44. To help people who use pcregrep to scan files containing exceedingly long
888        lines, the following changes have been made:
889    
890        (a) The default value of the buffer size parameter has been increased from
891            8K to 20K. (The actual buffer used is three times this size.)
892    
893        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
894            PCRE is built.
895    
896        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
897            to be set at run time.
898    
899        (d) Numerical values in pcregrep options can be followed by K or M, for
900            example --buffer-size=50K.
901    
902        (e) If a line being scanned overflows pcregrep's buffer, an error is now
903            given and the return code is set to 2.
904    
905    45. Add a pointer to the latest mark to the callout data block.
906    
907    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
908        partial match of an empty string instead of no match. This was specific to
909        the use of ".".
910    
911    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
912        complete match instead of a partial match. This bug was dependent on both
913        the PCRE_UTF8 and PCRE_DOTALL options being set.
914    
915    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
916        starting byte set, because \b was not being ignored.
917    
918    
919    Version 8.12 15-Jan-2011
920    ------------------------
921    
922    1.  Fixed some typos in the markup of the man pages, and wrote a script that
923        checks for such things as part of the documentation building process.
924    
925    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
926        --match-limit and --recursion-limit options (added for 8.11). In
927        particular, this made one of the standard tests fail. (The integer value
928        went into the wrong half of a long int.)
929    
930    3.  If the --colour option was given to pcregrep with -v (invert match), it
931        did strange things, either producing crazy output, or crashing. It should,
932        of course, ignore a request for colour when reporting lines that do not
933        match.
934    
935    4.  Another pcregrep bug caused similar problems if --colour was specified with
936        -M (multiline) and the pattern match finished with a line ending.
937    
938    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
939        matched in multiline mode, the following line was shown as part of the
940        match. This seems wrong, so I have changed it.
941    
942    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
943        the check for further matches in the same line (so they could be coloured)
944        to overrun the end of the current line. If another match was found, it was
945        incorrectly shown (and then shown again when found in the next line).
946    
947    7.  If pcregrep was compiled under Windows, there was a reference to the
948        function pcregrep_exit() before it was defined. I am assuming this was
949        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
950        reported by a user. I've moved the definition above the reference.
951    
952    
953    Version 8.11 10-Dec-2010
954    ------------------------
955    
956    1.  (*THEN) was not working properly if there were untried alternatives prior
957        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
958        backtracked to try for "b" instead of moving to the next alternative branch
959        at the same level (in this case, to look for "c"). The Perl documentation
960        is clear that when (*THEN) is backtracked onto, it goes to the "next
961        alternative in the innermost enclosing group".
962    
963    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
964        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
965        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
966        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
967        (*THEN).
968    
969    3.  If \s appeared in a character class, it removed the VT character from
970        the class, even if it had been included by some previous item, for example
971        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
972        of \s, but is part of the POSIX "space" class.)
973    
974    4.  A partial match never returns an empty string (because you can always
975        match an empty string at the end of the subject); however the checking for
976        an empty string was starting at the "start of match" point. This has been
977        changed to the "earliest inspected character" point, because the returned
978        data for a partial match starts at this character. This means that, for
979        example, /(?<=abc)def/ gives a partial match for the subject "abc"
980        (previously it gave "no match").
981    
982    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
983        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
984        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
985        has an implication that the given string is incomplete (because a partial
986        match is preferred over a full match). For this reason, these items now
987        give a partial match in this situation. [Aside: previously, the one case
988        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
989        match rather than a full match, which was wrong by the old rules, but is
990        now correct.]
991    
992    6.  There was a bug in the handling of #-introduced comments, recognized when
993        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
994        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
995        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
996        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
997        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
998        places in pcre_compile().
999    
1000    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1001        comments when looking ahead for named forward references to subpatterns,
1002        the only newline sequence it recognized was NL. It now handles newlines
1003        according to the set newline convention.
1004    
1005    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1006        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1007        cater for a lack of strerror(). These oversights have been fixed.
1008    
1009    9.  Added --match-limit and --recursion-limit to pcregrep.
1010    
1011    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1012    
1013    11. When the -o option was used, pcregrep was setting a return code of 1, even
1014        when matches were found, and --line-buffered was not being honoured.
1015    
1016    12. Added an optional parentheses number to the -o and --only-matching options
1017        of pcregrep.
1018    
1019    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1020        can match an empty string. The code to do it in pcretest and pcredemo
1021        needed fixing:
1022    
1023        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1024            only one byte after an empty string match just before CRLF (this case
1025            just got forgotten; "any" and "anycrlf" were OK).
1026    
1027        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1028            mode when an empty string match preceded an ASCII character followed by
1029            a non-ASCII character. (The code for advancing by one character rather
1030            than one byte was nonsense.)
1031    
1032        (c) The pcredemo.c sample program did not have any code at all to handle
1033            the cases when CRLF is a valid newline sequence.
1034    
1035    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1036        as a starting offset was within the subject string. There is now a new
1037        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1038        negative or greater than the length of the string. In order to test this,
1039        pcretest is extended to allow the setting of negative starting offsets.
1040    
1041    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1042        starting offset points to the beginning of a UTF-8 character was
1043        unnecessarily clumsy. I tidied it up.
1044    
1045    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1046        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1047    
1048    17. Nobody had reported that the --include_dir option, which was added in
1049        release 7.7 should have been called --include-dir (hyphen, not underscore)
1050        for compatibility with GNU grep. I have changed it to --include-dir, but
1051        left --include_dir as an undocumented synonym, and the same for
1052        --exclude-dir, though that is not available in GNU grep, at least as of
1053        release 2.5.4.
1054    
1055    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1056        characters from a string of bytes) have been redefined so as not to use
1057        loops, in order to improve performance in some environments. At the same
1058        time, I abstracted some of the common code into auxiliary macros to save
1059        repetition (this should not affect the compiled code).
1060    
1061    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1062        compile-time error is now given if \c is not followed by an ASCII
1063        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1064        different, and any byte value is allowed.)
1065    
1066    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1067        START_OPTIMIZE option, which is now allowed at compile time - but just
1068        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1069        to pcregrep and other applications that have no direct access to PCRE
1070        options. The new /Y option in pcretest sets this option when calling
1071        pcre_compile().
1072    
1073    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1074        back references. Groups containing recursive back references were forced to
1075        be atomic by that change, but in the case of named groups, the amount of
1076        memory required was incorrectly computed, leading to "Failed: internal
1077        error: code overflow". This has been fixed.
1078    
1079    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1080        pcretest.c, to avoid build problems in some Borland environments.
1081    
1082    
1083    Version 8.10 25-Jun-2010
1084    ------------------------
1085    
1086    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1087        THEN.
1088    
1089    2.  (*ACCEPT) was not working when inside an atomic group.
1090    
1091    3.  Inside a character class, \B is treated as a literal by default, but
1092        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1093        causes the error). The code is unchanged, but I tidied the documentation.
1094    
1095    4.  Inside a character class, PCRE always treated \R and \X as literals,
1096        whereas Perl faults them if its -w option is set. I have changed PCRE so
1097        that it faults them when PCRE_EXTRA is set.
1098    
1099    5.  Added support for \N, which always matches any character other than
1100        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1101    
1102    6.  When compiling pcregrep with newer versions of gcc which may have
1103        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1104        declared with attribute warn_unused_result" were given. Just casting the
1105        result to (void) does not stop the warnings; a more elaborate fudge is
1106        needed. I've used a macro to implement this.
1107    
1108    7.  Minor change to pcretest.c to avoid a compiler warning.
1109    
1110    8.  Added four artifical Unicode properties to help with an option to make
1111        \s etc use properties (see next item). The new properties are: Xan
1112        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1113    
1114    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1115        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1116        this option. Modified pcretest to add /W to test this facility. Added
1117        REG_UCP to make it available via the POSIX interface.
1118    
1119    10. Added --line-buffered to pcregrep.
1120    
1121    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1122        studied, and the match started with a letter with a code point greater than
1123        127 whose first byte was different to the first byte of the other case of
1124        the letter, the other case of this starting letter was not recognized
1125        (#976).
1126    
1127    12. If a pattern that was studied started with a repeated Unicode property
1128        test, for example, \p{Nd}+, there was the theoretical possibility of
1129        setting up an incorrect bitmap of starting bytes, but fortunately it could
1130        not have actually happened in practice until change 8 above was made (it
1131        added property types that matched character-matching opcodes).
1132    
1133    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1134        possible starting bytes for non-anchored patterns.
1135    
1136    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1137        \R, and also a number of cases that involve Unicode properties, both
1138        explicit and implicit when PCRE_UCP is set.
1139    
1140    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1141        input, it could crash or give wrong results if characters with values
1142        greater than 0xc0 were present in the subject string. (Detail: it assumed
1143        UTF-8 input when processing these items.)
1144    
1145    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1146        size_t is 64-bit (#991).
1147    
1148    17. Added a check for running out of memory when PCRE is compiled with
1149        --disable-stack-for-recursion (#990).
1150    
1151    18. If the last data line in a file for pcretest does not have a newline on
1152        the end, a newline was missing in the output.
1153    
1154    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1155        less than 128) in its various bitmaps. However, there is a facility for
1156        generating tables according to the current locale when PCRE is compiled. It
1157        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1158        characters, are recognized by isspace() and therefore were getting set in
1159        these tables, and indeed these tables seem to approximate to ISO 8859. This
1160        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1161        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1162        which of course cannot start UTF-8 characters. I have changed the code so
1163        that only real ASCII characters (less than 128) and the correct starting
1164        bytes for UTF-8 encodings are set for characters greater than 127 when in
1165        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1166        altogether.)
1167    
1168    20. Added the /T option to pcretest so as to be able to run tests with non-
1169        standard character tables, thus making it possible to include the tests
1170        used for 19 above in the standard set of tests.
1171    
1172    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1173        reference to a subpattern the other side of a comment that contains an
1174        opening parenthesis caused either an internal compiling error, or a
1175        reference to the wrong subpattern.
1176    
1177    
1178    Version 8.02 19-Mar-2010
1179    ------------------------
1180    
1181    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1182    
1183    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1184        configured.
1185    
1186    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1187        original author of that file, following a query about its status.
1188    
1189    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1190        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1191    
1192    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1193        quantifier applied to a forward-referencing subroutine call, could compile
1194        incorrect code or give the error "internal error: previously-checked
1195        referenced subpattern not found".
1196    
1197    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1198        variables to point to external functions. For these systems, therefore,
1199        pcre_malloc etc. are now initialized to local functions that call the
1200        relevant global functions.
1201    
1202    7.  There were two entries missing in the vectors called coptable and poptable
1203        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1204        I've fixed the data, and added a kludgy way of testing at compile time that
1205        the lengths are correct (equal to the number of opcodes).
1206    
1207    8.  Following on from 7, I added a similar kludge to check the length of the
1208        eint vector in pcreposix.c.
1209    
1210    9.  Error texts for pcre_compile() are held as one long string to avoid too
1211        much relocation at load time. To find a text, the string is searched,
1212        counting zeros. There was no check for running off the end of the string,
1213        which could happen if a new error number was added without updating the
1214        string.
1215    
1216    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1217    
1218    11. \K was not working if it appeared in an atomic group or in a group that
1219        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1220        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1221        the assertion is positive, but not if it is negative.
1222    
1223    12. Change 11 fortuitously reduced the size of the stack frame used in the
1224        "match()" function of pcre_exec.c by one pointer. Forthcoming
1225        implementation of support for (*MARK) will need an extra pointer on the
1226        stack; I have reserved it now, so that the stack frame size does not
1227        decrease.
1228    
1229    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1230        item in branch that calls a recursion is a subroutine call - as in the
1231        second branch in the above example - was incorrectly given the compile-
1232        time error "recursive call could loop indefinitely" because pcre_compile()
1233        was not correctly checking the subroutine for matching a non-empty string.
1234    
1235    14. The checks for overrunning compiling workspace could trigger after an
1236        overrun had occurred. This is a "should never occur" error, but it can be
1237        triggered by pathological patterns such as hundreds of nested parentheses.
1238        The checks now trigger 100 bytes before the end of the workspace.
1239    
1240    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1241    
1242    
1243    Version 8.01 19-Jan-2010
1244    ------------------------
1245    
1246    1.  If a pattern contained a conditional subpattern with only one branch (in
1247        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1248        computed the wrong minimum data length (which is of course zero for such
1249        subpatterns). This could cause incorrect "no match" results.
1250    
1251    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1252        the pattern is reset in the first branch, pcre_compile() failed with
1253        "internal error: code overflow at offset...". This happened only when
1254        the reset was to the original external option setting. (An optimization
1255        abstracts leading options settings into an external setting, which was the
1256        cause of this.)
1257    
1258    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1259        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1260        assertion pattern did not match (meaning that the assertion was true), it
1261        was incorrectly treated as false if the SKIP had been reached during the
1262        matching. This also applied to assertions used as conditions.
1263    
1264    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1265        assertion subpattern, including such a pattern used as a condition,
1266        unpredictable results occurred, instead of the error return
1267        PCRE_ERROR_DFA_UITEM.
1268    
1269    5.  The C++ GlobalReplace function was not working like Perl for the special
1270        situation when an empty string is matched. It now does the fancy magic
1271        stuff that is necessary.
1272    
1273    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1274        removed. (These were left over from very, very early versions of PCRE.)
1275    
1276    7.  Some cosmetic changes to the code to make life easier when compiling it
1277        as part of something else:
1278    
1279        (a) Change DEBUG to PCRE_DEBUG.
1280    
1281        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1282            called "current" as "current_branch", to prevent a collision with the
1283            Linux macro when compiled as a kernel module.
1284    
1285        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1286            prevent a collision with the Linux macro when compiled as a kernel
1287            module.
1288    
1289    8.  In pcre_compile() there are some checks for integer overflows that used to
1290        cast potentially large values to (double). This has been changed to that
1291        when building, a check for int64_t is made, and if it is found, it is used
1292        instead, thus avoiding the use of floating point arithmetic. (There is no
1293        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1294        double.
1295    
1296    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1297        2005 (difference between two addresses compared to an unsigned value).
1298    
1299    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1300        custom one, because of the following reported problem in Windows:
1301    
1302          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1303              under Win32.
1304          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1305              therefore missing the function definition.
1306          - The compiler thus generates a "C" signature for the test function.
1307          - The linker fails to find the "C" function.
1308          - PCRE fails to configure if asked to do so against libbz2.
1309    
1310    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1311        messages were output:
1312    
1313          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1314          rerunning libtoolize, to keep the correct libtool macros in-tree.
1315          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1316    
1317        I have done both of these things.
1318    
1319    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1320        most of the time, it *can* run out if it is given a pattern that contains a
1321        runaway infinite recursion. I updated the discussion in the pcrestack man
1322        page.
1323    
1324    13. Now that we have gone to the x.xx style of version numbers, the minor
1325        version may start with zero. Using 08 or 09 is a bad idea because users
1326        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1327        interpreted as invalid octal numbers. I've updated the previous comment in
1328        configure.ac, and also added a check that gives an error if 08 or 09 are
1329        used.
1330    
1331    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1332        causing partial matching to fail when the end of the subject matched \W
1333        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1334    
1335    15. There were some discrepancies between the declarations in pcre_internal.h
1336        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1337        their definitions. The declarations used "const uschar *" and the
1338        definitions used USPTR. Even though USPTR is normally defined as "const
1339        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1340        reported that: "This difference in casting confuses some C++ compilers, for
1341        example, SunCC recognizes above declarations as different functions and
1342        generates broken code for hbpcre." I have changed the declarations to use
1343        USPTR.
1344    
1345    16. GNU libtool is named differently on some systems. The autogen.sh script now
1346        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1347        (FreeBSD).
1348    
1349    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1350        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1351        comment: "Figure out how to create a longlong from a string: strtoll and
1352        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1353        instance, but it only takes 2 args instead of 3!"
1354    
1355    18. A subtle bug concerned with back references has been fixed by a change of
1356        specification, with a corresponding code fix. A pattern such as
1357        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1358        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1359        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1360        same bug. Such groups have to be quantified to be useful, or contained
1361        inside another quantified group. (If there's no repetition, the reference
1362        can never match.) The problem arises because, having left the group and
1363        moved on to the rest of the pattern, a later failure that backtracks into
1364        the group uses the captured value from the final iteration of the group
1365        rather than the correct earlier one. I have fixed this in PCRE by forcing
1366        any group that contains a reference to itself to be an atomic group; that
1367        is, there cannot be any backtracking into it once it has completed. This is
1368        similar to recursive and subroutine calls.
1369    
1370    
1371    Version 8.00 19-Oct-09
1372    ----------------------
1373    
1374    1.  The table for translating pcre_compile() error codes into POSIX error codes
1375        was out-of-date, and there was no check on the pcre_compile() error code
1376        being within the table. This could lead to an OK return being given in
1377        error.
1378    
1379    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1380        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1381        in a Windows environment.
1382    
1383    3.  The pcregrep --count option prints the count for each file even when it is
1384        zero, as does GNU grep. However, pcregrep was also printing all files when
1385        --files-with-matches was added. Now, when both options are given, it prints
1386        counts only for those files that have at least one match. (GNU grep just
1387        prints the file name in this circumstance, but including the count seems
1388        more useful - otherwise, why use --count?) Also ensured that the
1389        combination -clh just lists non-zero counts, with no names.
1390    
1391    4.  The long form of the pcregrep -F option was incorrectly implemented as
1392        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1393        but it seems right to fix it, and I didn't think it was worth preserving
1394        the old behaviour.
1395    
1396    5.  The command line items --regex=pattern and --regexp=pattern were not
1397        recognized by pcregrep, which required --regex pattern or --regexp pattern
1398        (with a space rather than an '='). The man page documented the '=' forms,
1399        which are compatible with GNU grep; these now work.
1400    
1401    6.  No libpcreposix.pc file was created for pkg-config; there was just
1402        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1403    
1404    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1405        when UCP support is not needed, by modifying the Python script that
1406        generates it from Unicode data files. This should not matter if the module
1407        is correctly used as a library, but I received one complaint about 50K of
1408        unwanted data. My guess is that the person linked everything into his
1409        program rather than using a library. Anyway, it does no harm.
1410    
1411    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1412        was a minimum greater than 1 for a wide character in a possessive
1413        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1414        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1415        character. Chaos in the form of incorrect output or a compiling loop could
1416        result.
1417    
1418    9.  The restrictions on what a pattern can contain when partial matching is
1419        requested for pcre_exec() have been removed. All patterns can now be
1420        partially matched by this function. In addition, if there are at least two
1421        slots in the offset vector, the offset of the earliest inspected character
1422        for the match and the offset of the end of the subject are set in them when
1423        PCRE_ERROR_PARTIAL is returned.
1424    
1425    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1426        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1427        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1428        and may be more useful for multi-segment matching.
1429    
1430    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1431        used to be given if ever the end of the subject was reached; now it is
1432        given only if matching could not proceed because another character was
1433        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1434        string "Z", which now yields "no match" instead of "partial match". In the
1435        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1436        final character ended with (*FAIL).
1437    
1438    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1439        if the pattern had a "must contain" character that was already found in the
1440        earlier partial match, unless partial matching was again requested. For
1441        example, with the pattern /dog.(body)?/, the "must contain" character is
1442        "g". If the first part-match was for the string "dog", restarting with
1443        "sbody" failed. This bug has been fixed.
1444    
1445    13. The string returned by pcre_dfa_exec() after a partial match has been
1446        changed so that it starts at the first inspected character rather than the
1447        first character of the match. This makes a difference only if the pattern
1448        starts with a lookbehind assertion or \b or \B (\K is not supported by
1449        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1450        matching functions compatible, and I think it's the right thing to do.
1451    
1452    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1453        so that the demonstration program is easily available in environments where
1454        PCRE has not been installed from source.
1455    
1456    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1457        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1458        library.
1459    
1460    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1461        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1462        is not the first non-POSIX option to be added. Clearly some people find
1463        these options useful.
1464    
1465    17. If a caller to the POSIX matching function regexec() passes a non-zero
1466        value for nmatch with a NULL value for pmatch, the value of
1467        nmatch is forced to zero.
1468    
1469    18. RunGrepTest did not have a test for the availability of the -u option of
1470        the diff command, as RunTest does. It now checks in the same way as
1471        RunTest, and also checks for the -b option.
1472    
1473    19. If an odd number of negated classes containing just a single character
1474        interposed, within parentheses, between a forward reference to a named
1475        subpattern and the definition of the subpattern, compilation crashed with
1476        an internal error, complaining that it could not find the referenced
1477        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1478        [The bug was that it was starting one character too far in when skipping
1479        over the character class, thus treating the ] as data rather than
1480        terminating the class. This meant it could skip too much.]
1481    
1482    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1483        /g option in pcretest when the pattern contains \K, which makes it possible
1484        to have an empty string match not at the start, even when the pattern is
1485        anchored. Updated pcretest and pcredemo to use this option.
1486    
1487    21. If the maximum number of capturing subpatterns in a recursion was greater
1488        than the maximum at the outer level, the higher number was returned, but
1489        with unset values at the outer level. The correct (outer level) value is
1490        now given.
1491    
1492    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1493        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1494        make it do so. The string so far is captured, making this feature
1495        compatible with Perl.
1496    
1497    23. The tests have been re-organized, adding tests 11 and 12, to make it
1498        possible to check the Perl 5.10 features against Perl 5.10.
1499    
1500    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1501        pattern matches a fixed length string. PCRE did not allow this; now it
1502        does. Neither allows recursion.
1503    
1504    25. I finally figured out how to implement a request to provide the minimum
1505        length of subject string that was needed in order to match a given pattern.
1506        (It was back references and recursion that I had previously got hung up
1507        on.) This code has now been added to pcre_study(); it finds a lower bound
1508        to the length of subject needed. It is not necessarily the greatest lower
1509        bound, but using it to avoid searching strings that are too short does give
1510        some useful speed-ups. The value is available to calling programs via
1511        pcre_fullinfo().
1512    
1513    26. While implementing 25, I discovered to my embarrassment that pcretest had
1514        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1515        study optimizations had never been tested with that matching function.
1516        Oops. What is worse, even when it was passed study data, there was a bug in
1517        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1518        were also very few tests of studied patterns with pcre_dfa_exec().
1519    
1520    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1521        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1522        on the other side of the coin, they are no longer allowed to have different
1523        names, because these cannot be distinguished in PCRE, and this has caused
1524        confusion. (This is a difference from Perl.)
1525    
1526    28. When duplicate subpattern names are present (necessarily with different
1527        numbers, as required by 27 above), and a test is made by name in a
1528        conditional pattern, either for a subpattern having been matched, or for
1529        recursion in such a pattern, all the associated numbered subpatterns are
1530        tested, and the overall condition is true if the condition is true for any
1531        one of them. This is the way Perl works, and is also more like the way
1532        testing by number works.
1533    
1534    
1535    Version 7.9 11-Apr-09
1536    ---------------------
1537    
1538    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1539        (pcretest), all targets were linked against these libraries. This included
1540        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1541        libraries. This caused unwanted dependencies to be created. This problem
1542        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1543        pcretest is linked with readline.
1544    
1545    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1546        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1547        moved outside it again, because FALSE and TRUE are already defined in AIX,
1548        but BOOL is not.
1549    
1550    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1551        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1552    
1553    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1554        hyphens) following file names and line numbers when outputting matching
1555        lines. This is not true; no spaces are inserted. I have also clarified the
1556        wording for the --colour (or --color) option.
1557    
1558    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1559        was not coloured; this is different to GNU grep, so I have changed it to be
1560        the same.
1561    
1562    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1563        each matching line was coloured. Now it goes on to look for further matches
1564        of any of the test patterns, which is the same behaviour as GNU grep.
1565    
1566    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1567        doesn't make sense to accept an empty string match in pcregrep, so I have
1568        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1569        seems to be how GNU grep behaves.
1570    
1571    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1572        start or after a newline", because the conditional assertion was not being
1573        correctly handled. The rule now is that both the assertion and what follows
1574        in the first alternative must satisfy the test.
1575    
1576    9.  If auto-callout was enabled in a pattern with a conditional group whose
1577        condition was an assertion, PCRE could crash during matching, both with
1578        pcre_exec() and pcre_dfa_exec().
1579    
1580    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1581        used for matching.
1582    
1583    11. Unicode property support in character classes was not working for
1584        characters (bytes) greater than 127 when not in UTF-8 mode.
1585    
1586    12. Added the -M command line option to pcretest.
1587    
1588    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1589    
1590    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1591    
1592    16. Added comments and documentation about mis-use of no_arg in the C++
1593        wrapper.
1594    
1595    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1596        from Martin Jerabek that uses macro names for all relevant character and
1597        string constants.
1598    
1599    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1600        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1601        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1602        these, but not everybody uses configure.
1603    
1604    19. A conditional group that had only one branch was not being correctly
1605        recognized as an item that could match an empty string. This meant that an
1606        enclosing group might also not be so recognized, causing infinite looping
1607        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1608        with the subject "ab", where knowledge that the repeated group can match
1609        nothing is needed in order to break the loop.
1610    
1611    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1612        exec(), but without supplying a callout function, matching went wrong.
1613    
1614    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1615        leak if the size of the offset vector was greater than 30. When the vector
1616        is smaller, the saved offsets during recursion go onto a local stack
1617        vector, but for larger vectors malloc() is used. It was failing to free
1618        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1619        error, in fact).
1620    
1621    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1622        heapframe that is used only when UTF-8 support is enabled. This caused no
1623        problem, but was untidy.
1624    
1625    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1626        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1627        included within another project.
1628    
1629    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1630        slightly modified by me:
1631    
1632          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1633              not building pcregrep.
1634    
1635          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1636              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1637    
1638    25. Forward references, both numeric and by name, in patterns that made use of
1639        duplicate group numbers, could behave incorrectly or give incorrect errors,
1640        because when scanning forward to find the reference group, PCRE was not
1641        taking into account the duplicate group numbers. A pattern such as
1642        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1643    
1644    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1645        the feature of a custom pointer more persuasive (as requested by a user).
1646    
1647    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1648        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1649        pre-defined.
1650    
1651    28. Added support for (*UTF8) at the start of a pattern.
1652    
1653    29. Arrange for flags added by the "release type" setting in CMake to be shown
1654        in the configuration summary.
1655    
1656    
1657    Version 7.8 05-Sep-08
1658    ---------------------
1659    
1660    1.  Replaced UCP searching code with optimized version as implemented for Ad
1661        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1662        stage table and inline lookup instead of a function, giving speed ups of 2
1663        to 5 times on some simple patterns that I tested. Permission was given to
1664        distribute the MultiStage2.py script that generates the tables (it's not in
1665        the tarball, but is in the Subversion repository).
1666    
1667    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1668        scripts.
1669    
1670    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1671        a group with a zero qualifier. The result of the study could be incorrect,
1672        or the function might crash, depending on the pattern.
1673    
1674    4.  Caseless matching was not working for non-ASCII characters in back
1675        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1676        It now works when Unicode Property Support is available.
1677    
1678    5.  In pcretest, an escape such as \x{de} in the data was always generating
1679        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1680        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1681        truncation.
1682    
1683    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1684    
1685    7.  Added two (int) casts to pcregrep when printing the difference of two
1686        pointers, in case they are 64-bit values.
1687    
1688    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1689        test 2 if it fails.
1690    
1691    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1692        and a #define of that name to empty if it is not externally set. This is to
1693        allow users of MSVC to set it if necessary.
1694    
1695    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1696        the convenience functions in the pcre_get.c source file.
1697    
1698    11. An option change at the start of a pattern that had top-level alternatives
1699        could cause overwriting and/or a crash. This command provoked a crash in
1700        some environments:
1701    
1702          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1703    
1704        This potential security problem was recorded as CVE-2008-2371.
1705    
1706    12. For a pattern where the match had to start at the beginning or immediately
1707        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1708        pcre_dfa_exec() could read past the end of the passed subject if there was
1709        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1710        pcretest so that it places the subject at the end of its malloc-ed buffer.
1711    
1712    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1713        exec() might read past the end of the data buffer in UTF-8 mode.
1714    
1715    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1716        the data contained the byte 0x85 as part of a UTF-8 character within its
1717        first line. This applied both to normal and DFA matching.
1718    
1719    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1720        /^[^d]*?$/8 failed to match "abc".
1721    
1722    16. Added a missing copyright notice to pcrecpp_internal.h.
1723    
1724    17. Make it more clear in the documentation that values returned from
1725        pcre_exec() in ovector are byte offsets, not character counts.
1726    
1727    18. Tidied a few places to stop certain compilers from issuing warnings.
1728    
1729    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1730        supplied by Stefan Weber. I made a further small update for 7.8 because
1731        there is a change of source arrangements: the pcre_searchfuncs.c module is
1732        replaced by pcre_ucd.c.
1733    
1734    
1735    Version 7.7 07-May-08
1736    ---------------------
1737    
1738    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1739        a string to a long long, pretend we don't even have a long long." This is
1740        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1741    
1742    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1743        pre-7.6 versions, which defined a global no_arg variable instead of putting
1744        it in the RE class. (See also #8 below.)
1745    
1746    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1747        Lopes.
1748    
1749    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1750    
1751        (1) The include/exclude patterns were being applied to the whole pathnames
1752            of files, instead of just to the final components.
1753    
1754        (2) If there was more than one level of directory, the subdirectories were
1755            skipped unless they satisfied the include/exclude conditions. This is
1756            inconsistent with GNU grep (and could even be seen as contrary to the
1757            pcregrep specification - which I improved to make it absolutely clear).
1758            The action now is always to scan all levels of directory, and just
1759            apply the include/exclude patterns to regular files.
1760    
1761    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1762        --exclude_dir in the tests to avoid scanning .svn directories.
1763    
1764    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1765        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1766        doesn't support NULs in patterns.
1767    
1768    7.  Added some missing "const"s to declarations of static tables in
1769        pcre_compile.c and pcre_dfa_exec.c.
1770    
1771    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1772        caused by fix #2  above. (Subsequently also a second patch to fix the
1773        first patch. And a third patch - this was a messy problem.)
1774    
1775    9.  Applied Craig's patch to remove the use of push_back().
1776    
1777    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1778        matching function regexec().
1779    
1780    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1781        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1782        references. PCRE supports relative numbers with this syntax (I don't think
1783        Oniguruma does).
1784    
1785    12. Previously, a group with a zero repeat such as (...){0} was completely
1786        omitted from the compiled regex. However, this means that if the group
1787        was called as a subroutine from elsewhere in the pattern, things went wrong
1788        (an internal error was given). Such groups are now left in the compiled
1789        pattern, with a new opcode that causes them to be skipped at execution
1790        time.
1791    
1792    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1793        to the way PCRE behaves:
1794    
1795        (a) A lone ] character is dis-allowed (Perl treats it as data).
1796    
1797        (b) A back reference to an unmatched subpattern matches an empty string
1798            (Perl fails the current match path).
1799    
1800        (c) A data ] in a character class must be notated as \] because if the
1801            first data character in a class is ], it defines an empty class. (In
1802            Perl it is not possible to have an empty class.) The empty class []
1803            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1804            The negative empty class [^] matches any one character, independently
1805            of the DOTALL setting.
1806    
1807    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1808        non-existent subpattern following a character class starting with ']' and
1809        containing () gave an internal compiling error instead of "reference to
1810        non-existent subpattern". Fortunately, when the pattern did exist, the
1811        compiled code was correct. (When scanning forwards to check for the
1812        existencd of the subpattern, it was treating the data ']' as terminating
1813        the class, so got the count wrong. When actually compiling, the reference
1814        was subsequently set up correctly.)
1815    
1816    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1817        it was being rejected as not supported by pcre_dfa_exec(), even though
1818        other assertions are supported. I have made pcre_dfa_exec() support
1819        (*FAIL).
1820    
1821    16. The implementation of 13c above involved the invention of a new opcode,
1822        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1823        cannot be changed at match time, I realized I could make a small
1824        improvement to matching performance by compiling OP_ALLANY instead of
1825        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1826        on the OP_ANY path.
1827    
1828    17. Compiling pcretest on Windows with readline support failed without the
1829        following two fixes: (1) Make the unistd.h include conditional on
1830        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1831    
1832    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1833        ncurses library to be included for pcretest when ReadLine support is
1834        requested, but also to allow for it to be overridden. This patch came from
1835        Daniel Bergström.
1836    
1837    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1838        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1839        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1840        spotting this.
1841    
1842    
1843    Version 7.6 28-Jan-08
1844    ---------------------
1845    
1846    1.  A character class containing a very large number of characters with
1847        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1848        overflow.
1849    
1850    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1851        HAVE_LONG_LONG is not defined.
1852    
1853    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1854        bring them up to date and include new features. This patch includes:
1855    
1856        - Fixed PH's badly added libz and libbz2 support.
1857        - Fixed a problem with static linking.
1858        - Added pcredemo. [But later removed - see 7 below.]
1859        - Fixed dftables problem and added an option.
1860        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1861            HAVE_LONG_LONG.
1862        - Added readline support for pcretest.
1863        - Added an listing of the option settings after cmake has run.
1864    
1865    4.  A user submitted a patch to Makefile that makes it easy to create
1866        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1867        Makefile.am that cause it to include this special target, without
1868        affecting anything else. Note that the same mingw target plus all
1869        the other distribution libraries and programs are now supported
1870        when configuring with CMake (see 6 below) instead of with
1871        Configure/Make.
1872    
1873    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1874        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1875        exported in the Windows port". It has not yet been confirmed that the patch
1876        solves the problem, but it does no harm.
1877    
1878    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1879        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1880        with CMake, and also correct the comment about stack recursion.
1881    
1882    7.  Remove the automatic building of pcredemo from the ./configure system and
1883        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1884        of a program that users should build themselves after PCRE is installed, so
1885        building it automatically is not really right. What is more, it gave
1886        trouble in some build environments.
1887    
1888    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1889    
1890    
1891    Version 7.5 10-Jan-08
1892    ---------------------
1893    
1894    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1895        values in parens when parsing an RE using the C++ wrapper."
1896    
1897    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1898        Characters greater than 255 were excluded from the class instead of being
1899        included.
1900    
1901    3.  The same bug as (2) above applied to negated POSIX classes such as
1902        [:^space:].
1903    
1904    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1905        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1906        I have changed it.
1907    
1908    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1909        first named subpattern) and a construct such as (?&a) would reference the
1910        first named subpattern whose name started with "a" (in other words, the
1911        length check was missing). Both these problems are fixed. "Subpattern name
1912        expected" is now given for (?&) (a zero-length name), and this patch also
1913        makes it give the same error for \k'' (previously it complained that that
1914        was a reference to a non-existent subpattern).
1915    
1916    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1917        this is right because (?- can be followed by option settings as well as by
1918        digits. I have, however, made the messages clearer.
1919    
1920    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1921        than the number used in the conditional) now cause a compile-time error.
1922        This is actually not compatible with Perl, which accepts such patterns, but
1923        treats the conditional as always being FALSE (as PCRE used to), but it
1924        seems to me that giving a diagnostic is better.
1925    
1926    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1927        and messages.
1928    
1929    9.  Fix two occurrences of "backslash" in comments that should have been
1930        "backspace".
1931    
1932    10. Remove two redundant lines of code that can never be obeyed (their function
1933        was moved elsewhere).
1934    
1935    11. The program that makes PCRE's Unicode character property table had a bug
1936        which caused it to generate incorrect table entries for sequences of
1937        characters that have the same character type, but are in different scripts.
1938        It amalgamated them into a single range, with the script of the first of
1939        them. In other words, some characters were in the wrong script. There were
1940        thirteen such cases, affecting characters in the following ranges:
1941    
1942          U+002b0 - U+002c1
1943          U+0060c - U+0060d
1944          U+0061e - U+00612
1945          U+0064b - U+0065e
1946          U+0074d - U+0076d
1947          U+01800 - U+01805
1948          U+01d00 - U+01d77
1949          U+01d9b - U+01dbf
1950          U+0200b - U+0200f
1951          U+030fc - U+030fe
1952          U+03260 - U+0327f
1953          U+0fb46 - U+0fbb1
1954          U+10450 - U+1049d
1955    
1956    12. The -o option (show only the matching part of a line) for pcregrep was not
1957        compatible with GNU grep in that, if there was more than one match in a
1958        line, it showed only the first of them. It now behaves in the same way as
1959        GNU grep.
1960    
1961    13. If the -o and -v options were combined for pcregrep, it printed a blank
1962        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1963        does the same. The return code can be used to tell if there were any
1964        non-matching lines.
1965    
1966    14. Added --file-offsets and --line-offsets to pcregrep.
1967    
1968    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1969        infinitely looping recursion. The bug was that positive lookaheads were not
1970        being skipped when checking for a possible empty match (negative lookaheads
1971        and both kinds of lookbehind were skipped).
1972    
1973    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1974        inclusion of <windows.h> to before rather than after the definition of
1975        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1976    
1977    17. Specifying a possessive quantifier with a specific limit for a Unicode
1978        character property caused pcre_compile() to compile bad code, which led at
1979        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1980        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1981        caused the error; without that there was no problem.
1982    
1983    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1984    
1985    19. Added --enable-pcretest-libreadline.
1986    
1987    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1988        RE::GlobalReplace(). As a result, the number of replacements returned was
1989        double what it should be. I removed one of the increments, but Craig sent a
1990        later patch that removed the other one (the right fix) and added unit tests
1991        that check the return values (which was not done before).
1992    
1993    21. Several CMake things:
1994    
1995        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1996            the names libpcre and libpcreposix, not just pcre and pcreposix.
1997    
1998        (2) The above change means that pcretest and pcregrep are now correctly
1999            linked with the newly-built libraries, not previously installed ones.
2000    
2001        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2002    
2003    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2004        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2005        UTF-8 newline character). The key issue is that the pattern starts .*;
2006        this means that the match must be either at the beginning, or after a
2007        newline. The bug was in the code for advancing after a failed match and
2008        checking that the new position followed a newline. It was not taking
2009        account of UTF-8 characters correctly.
2010    
2011    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2012        character classes. PCRE was not treating the sequence [:...:] as a
2013        character class unless the ... were all letters. Perl, however, seems to
2014        allow any characters between [: and :], though of course it rejects as
2015        unknown any "names" that contain non-letters, because all the known class
2016        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2017        for example, whereas PCRE did not - it did not recognize a POSIX character
2018        class. This seemed a bit dangerous, so the code has been changed to be
2019        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2020        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2021        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2022        Perl does, and where it didn't before.
2023    
2024    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2025        Windows environments %n is disabled by default.
2026    
2027    
2028    Version 7.4 21-Sep-07
2029    ---------------------
2030    
2031    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2032        means that a class such as [\s] counted as "explicit reference to CR or
2033        LF". That isn't really right - the whole point of the change was to try to
2034        help when there was an actual mention of one of the two characters. So now
2035        the change happens only if \r or \n (or a literal CR or LF) character is
2036        encountered.
2037    
2038    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2039        of both had grown to the point where there were only 3 bits left.
2040        Fortunately, there was spare space in the data structure, and so I have
2041        moved the internal flags into a new 16-bit field to free up more option
2042        bits.
2043    
2044    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2045        but did not set the internal JCHANGED flag - either of these is enough to
2046        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2047        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2048        start sets both bits.
2049    
2050    4.  Added options (at build time, compile time, exec time) to change \R from
2051        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2052    
2053    5.  doc/pcresyntax.html was missing from the distribution.
2054    
2055    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2056        compatibility, even though it is no longer used.
2057    
2058    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2059        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2060        windows.h file is present (where different names are used). [This was
2061        reversed later after testing - see 16 below.]
2062    
2063    8.  Changed all #include <config.h> to #include "config.h". There were also
2064        some further <pcre.h> cases that I changed to "pcre.h".
2065    
2066    9.  When pcregrep was used with the --colour option, it missed the line ending
2067        sequence off the lines that it output.
2068    
2069    10. It was pointed out to me that arrays of string pointers cause lots of
2070        relocations when a shared library is dynamically loaded. A technique of
2071        using a single long string with a table of offsets can drastically reduce
2072        these. I have refactored PCRE in four places to do this. The result is
2073        dramatic:
2074    
2075          Originally:                          290
2076          After changing UCP table:            187
2077          After changing error message table:   43
2078          After changing table of "verbs"       36
2079          After changing table of Posix names   22
2080    
2081        Thanks to the folks working on Gregex for glib for this insight.
2082    
2083    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2084        unicode-properties was also set.
2085    
2086    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2087    
2088    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2089        checked only for CRLF.
2090    
2091    14. Added casts to pcretest.c to avoid compiler warnings.
2092    
2093    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2094    
2095    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2096        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2097        entirely. This removes changes made in 7 above.
2098    
2099    17. The CMake files have been updated, and there is now more information about
2100        building with CMake in the NON-UNIX-USE document.
2101    
2102    
2103    Version 7.3 28-Aug-07
2104    ---------------------
2105    
2106     1. In the rejigging of the build system that eventually resulted in 7.1, the
2107        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2108        brackets there is not right, since it causes compilers to look for an
2109        installed pcre.h, not the version that is in the source that is being
2110        compiled (which of course may be different). I have changed it back to:
2111    
2112          #include "pcre.h"
2113    
2114        I have a vague recollection that the change was concerned with compiling in
2115        different directories, but in the new build system, that is taken care of
2116        by the VPATH setting the Makefile.
2117    
2118     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2119        when the subject happened to end in the byte 0x85 (e.g. if the last
2120        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2121        characters but of course it shouldn't be taken as a newline when it is part
2122        of another character. The bug was that, for an unlimited repeat of . in
2123        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2124        characters when looking for a newline.
2125    
2126     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2127    
2128     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2129        in debug output.
2130    
2131     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2132        long printing in the pcrecpp unittest when running under MinGW.
2133    
2134     6. ESC_K was left out of the EBCDIC table.
2135    
2136     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2137        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2138        limit also applies to "virtual nesting" when a pattern is recursive, and in
2139        this case 1000 isn't so big. I have been able to remove this limit at the
2140        expense of backing off one optimization in certain circumstances. Normally,
2141        when pcre_exec() would call its internal match() function recursively and
2142        immediately return the result unconditionally, it uses a "tail recursion"
2143        feature to save stack. However, when a subpattern that can match an empty
2144        string has an unlimited repetition quantifier, it no longer makes this
2145        optimization. That gives it a stack frame in which to save the data for
2146        checking that an empty string has been matched. Previously this was taken
2147        from the 1000-entry workspace that had been reserved. So now there is no
2148        explicit limit, but more stack is used.
2149    
2150     8. Applied Daniel's patches to solve problems with the import/export magic
2151        syntax that is required for Windows, and which was going wrong for the
2152        pcreposix and pcrecpp parts of the library. These were overlooked when this
2153        problem was solved for the main library.
2154    
2155     9. There were some crude static tests to avoid integer overflow when computing
2156        the size of patterns that contain repeated groups with explicit upper
2157        limits. As the maximum quantifier is 65535, the maximum group length was
2158        set at 30,000 so that the product of these two numbers did not overflow a
2159        32-bit integer. However, it turns out that people want to use groups that
2160        are longer than 30,000 bytes (though not repeat them that many times).
2161        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2162        made it possible to implement the integer overflow checks in a much more
2163        dynamic way, which I have now done. The artificial limitation on group
2164        length has been removed - we now have only the limit on the total length of
2165        the compiled pattern, which depends on the LINK_SIZE setting.
2166    
2167    10. Fixed a bug in the documentation for get/copy named substring when
2168        duplicate names are permitted. If none of the named substrings are set, the
2169        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2170        empty string.
2171    
2172    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2173        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2174        because the ] is interpreted as the first data character and the
2175        terminating ] is not found. PCRE has been made compatible with Perl in this
2176        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2177        cause memory overwriting.
2178    
2179    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2180        string has been matched (to stop an infinite loop). It was not recognizing
2181        a conditional subpattern that could match an empty string if that
2182        subpattern was within another subpattern. For example, it looped when
2183        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2184        condition was not nested. This bug has been fixed.
2185    
2186    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2187        past the start of the subject in the presence of bytes with the top bit
2188        set, for example "\x8aBCD".
2189    
2190    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2191        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2192    
2193    14. Optimized (?!) to (*FAIL).
2194    
2195    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2196        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2197        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2198        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2199        does: it's just the validity check that is more restrictive.
2200    
2201    16. Inserted checks for integer overflows during escape sequence (backslash)
2202        processing, and also fixed erroneous offset values for syntax errors during
2203        backslash processing.
2204    
2205    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2206        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2207    
2208    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2209        caused an overrun.
2210    
2211    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2212        something other than just ASCII characters) inside a group that had an
2213        unlimited repeat caused a loop at compile time (while checking to see
2214        whether the group could match an empty string).
2215    
2216    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2217        [\P{Any}] did so. (Error in the code for printing property names.)
2218    
2219    21. An orphan \E inside a character class could cause a crash.
2220    
2221    22. A repeated capturing bracket such as (A)? could cause a wild memory
2222        reference during compilation.
2223    
2224    23. There are several functions in pcre_compile() that scan along a compiled
2225        expression for various reasons (e.g. to see if it's fixed length for look
2226        behind). There were bugs in these functions when a repeated \p or \P was
2227        present in the pattern. These operators have additional parameters compared
2228        with \d, etc, and these were not being taken into account when moving along
2229        the compiled data. Specifically:
2230    
2231        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2232            length.
2233    
2234        (b) An item such as \pL+ within a repeated group could cause crashes or
2235            loops.
2236    
2237        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2238            "reference to non-existent subpattern" error.
2239    
2240        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2241    
2242    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2243        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2244    
2245    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2246    
2247    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2248        character were causing crashes (broken optimization).
2249    
2250    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2251        \p or \P) caused a compile-time loop.
2252    
2253    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2254        break. For example, the unstudied pattern [\r\n]A does not match the string
2255        "\r\nA" because change 7.0/46 below moves the current point on by two
2256        characters after failing to match at the start. However, the pattern \nA
2257        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2258        the same is true. There doesn't seem any very clean way out of this, but
2259        what I have chosen to do makes the common cases work: PCRE now takes note
2260        of whether there can be an explicit match for \r or \n anywhere in the
2261        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2262        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2263        pattern has explicit CR or LF references.
2264    
2265    29. Added (*CR) etc for changing newline setting at start of pattern.
2266    
2267    
2268    Version 7.2 19-Jun-07
2269    ---------------------
2270    
2271     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2272        which is apparently normally available under Windows.
2273    
2274     2. Re-jig the pcregrep tests with different newline settings in an attempt
2275        to make them independent of the local environment's newline setting.
2276    
2277     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2278    
2279     4. Some of the "internals" tests were previously cut out when the link size
2280        was not 2, because the output contained actual offsets. The recent new
2281        "Z" feature of pcretest means that these can be cut out, making the tests
2282        usable with all link sizes.
2283    
2284     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2285        stack recursion. This gives a massive performance boost under BSD, but just
2286        a small improvement under Linux. However, it saves one field in the frame
2287        in all cases.
2288    
2289     6. Added more features from the forthcoming Perl 5.10:
2290    
2291        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2292            recursion call. It refers to the nth most recently opened parentheses.
2293    
2294        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2295            to be opened parentheses.
2296    
2297        (c) Conditions that refer to capturing parentheses can be specified
2298            relatively, for example, (?(-2)... or (?(+3)...
2299    
2300        (d) \K resets the start of the current match so that everything before
2301            is not part of it.
2302    
2303        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2304    
2305        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2306            reference syntax.
2307    
2308        (g) (?| introduces a group in which the numbering of parentheses in each
2309            alternative starts with the same number.
2310    
2311        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2312    
2313     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2314        PCRE_INFO_JCHANGED.
2315    
2316     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2317        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2318        for detecting groups that can match an empty string.
2319    
2320     9. A pattern with a very large number of alternatives (more than several
2321        hundred) was running out of internal workspace during the pre-compile
2322        phase, where pcre_compile() figures out how much memory will be needed. A
2323        bit of new cunning has reduced the workspace needed for groups with
2324        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2325        workspace instead of running out of the 4096 that are available.
2326    
2327    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2328    
2329    11. Applied patch from Google to remove an optimization that didn't quite work.
2330        The report of the bug said:
2331    
2332          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2333          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2334          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2335    
2336    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2337        it matched the wrong number of bytes.
2338    
2339    
2340    Version 7.1 24-Apr-07
2341    ---------------------
2342    
2343     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2344        that is more "standard", making use of automake and other Autotools. There
2345        is some re-arrangement of the files and adjustment of comments consequent
2346        on this.
2347    
2348     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2349        for recursive directory scanning broke on some systems because the files
2350        are not scanned in any specific order and on different systems the order
2351        was different. A call to "sort" has been inserted into RunGrepTest for the
2352        approprate test as a short-term fix. In the longer term there may be an
2353        alternative.
2354    
2355     3. I had an email from Eric Raymond about problems translating some of PCRE's
2356        man pages to HTML (despite the fact that I distribute HTML pages, some
2357        people do their own conversions for various reasons). The problems
2358        concerned the use of low-level troff macros .br and .in. I have therefore
2359        removed all such uses from the man pages (some were redundant, some could
2360        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2361        HTML has been updated to handle .nf/.fi and to complain if it encounters
2362        .br or .in.
2363    
2364     4. Updated comments in configure.ac that get placed in config.h.in and also
2365        arranged for config.h to be included in the distribution, with the name
2366        config.h.generic, for the benefit of those who have to compile without
2367        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2368    
2369     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2370        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2371        makevp.bat for the current PCRE, using the additional files
2372        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2373    
2374     6. A Windows user reported a minor discrepancy with test 2, which turned out
2375        to be caused by a trailing space on an input line that had got lost in his
2376        copy. The trailing space was an accident, so I've just removed it.
2377    
2378     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2379        that is needed.
2380    
2381     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2382        as "const" (a) because they are and (b) because it helps the PHP
2383        maintainers who have recently made a script to detect big data structures
2384        in the php code that should be moved to the .rodata section. I remembered
2385        to update Builducptable as well, so it won't revert if ucptable.h is ever
2386        re-created.
2387    
2388     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2389        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2390        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2391        support is not required. This saves 1.5-2K of code, which is important in
2392        some applications.
2393    
2394        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2395        so as not to refer to the tables, even though these functions will never be
2396        called when UTF-8 support is disabled. Otherwise there are problems with a
2397        shared library.
2398    
2399    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2400    
2401        (a) It was defining its arguments as char * instead of void *.
2402    
2403        (b) It was assuming that all moves were upwards in memory; this was true
2404            a long time ago when I wrote it, but is no longer the case.
2405    
2406        The emulated memove() is provided for those environments that have neither
2407        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2408        is clearly not the case, as these two bugs were recently reported.
2409    
2410    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2411        and Detrail to create the HTML documentation, the .txt form of the man
2412        pages, and it removes trailing spaces from listed files. It also creates
2413        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2414        case, it wraps all the #defines with #ifndefs. This script should be run
2415        before "make dist".
2416    
2417    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2418        with Unicode property support.
2419    
2420        (a) For a maximizing quantifier, if the two different cases of the
2421            character were of different lengths in their UTF-8 codings (there are
2422            some cases like this - I found 11), and the matching function had to
2423            back up over a mixture of the two cases, it incorrectly assumed they
2424            were both the same length.
2425    
2426        (b) When PCRE was configured to use the heap rather than the stack for
2427            recursion during matching, it was not correctly preserving the data for
2428            the other case of a UTF-8 character when checking ahead for a match
2429            while processing a minimizing repeat. If the check also involved
2430            matching a wide character, but failed, corruption could cause an
2431            erroneous result when trying to check for a repeat of the original
2432            character.
2433    
2434    13. Some tidying changes to the testing mechanism:
2435    
2436        (a) The RunTest script now detects the internal link size and whether there
2437            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2438            values substituted by "configure". (The RunGrepTest script already did
2439            this for UTF-8.) The configure.ac script no longer substitutes the
2440            relevant variables.
2441    
2442        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2443            with length and offset values. This means that the output is different
2444            for different internal link sizes. Test 2 is skipped for link sizes
2445            other than 2 because of this, bypassing the problem. Unfortunately,
2446            there was also a test in test 3 (the locale tests) that used /B and
2447            failed for link sizes other than 2. Rather than cut the whole test out,
2448            I have added a new /Z option to pcretest that replaces the length and
2449            offset values with spaces. This is now used to make test 3 independent
2450            of link size. (Test 2 will be tidied up later.)
2451    
2452    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2453        segmentation fault instead of returning the appropriate error message.
2454    
2455    15. In multiline mode when the newline sequence was set to "any", the pattern
2456        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2457        This doesn't seem right; it now treats the CRLF combination as the line
2458        ending, and so does not match in that case. It's only a pattern such as ^$
2459        that would hit this one: something like ^ABC$ would have failed after \r
2460        and then tried again after \r\n.
2461    
2462    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2463        in an attempt to make files that differ only in their line terminators
2464        compare equal. This works on Linux.
2465    
2466    17. Under certain error circumstances pcregrep might try to free random memory
2467        as it exited. This is now fixed, thanks to valgrind.
2468    
2469    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2470        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2471        was because its rules for how to advance for /g after matching an empty
2472        string at the end of a line did not allow for this case. They now check for
2473        it specially.
2474    
2475    20. pcretest is supposed to handle patterns and data of any length, by
2476        extending its buffers when necessary. It was getting this wrong when the
2477        buffer for a data line had to be extended.
2478    
2479    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2480        CRLF as a newline sequence.
2481    
2482    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2483        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2484        I have nevertheless tidied it up.
2485    
2486    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2487    
2488    24. Added a man page for pcre-config.
2489    
2490    
2491    Version 7.0 19-Dec-06
2492    ---------------------
2493    
2494     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2495        moving to gcc 4.1.1.
2496    
2497     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2498        sys/time.h, which is documented as needed for this function. It doesn't
2499        seem to matter on Linux, but it showed up on some releases of OS X.
2500    
2501     3. It seems that there are systems where bytes whose values are greater than
2502        127 match isprint() in the "C" locale. The "C" locale should be the
2503        default when a C program starts up. In most systems, only ASCII printing
2504        characters match isprint(). This difference caused the output from pcretest
2505        to vary, making some of the tests fail. I have changed pcretest so that:
2506    
2507        (a) When it is outputting text in the compiled version of a pattern, bytes
2508            other than 32-126 are always shown as hex escapes.
2509    
2510        (b) When it is outputting text that is a matched part of a subject string,
2511            it does the same, unless a different locale has been set for the match
2512            (using the /L modifier). In this case, it uses isprint() to decide.
2513    
2514     4. Fixed a major bug that caused incorrect computation of the amount of memory
2515        required for a compiled pattern when options that changed within the
2516        pattern affected the logic of the preliminary scan that determines the
2517        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2518        that the computed length was too small. The symptoms of this bug were
2519        either the PCRE error "internal error: code overflow" from pcre_compile(),
2520        or a glibc crash with a message such as "pcretest: free(): invalid next
2521        size (fast)". Examples of patterns that provoked this bug (shown in
2522        pcretest format) are:
2523    
2524          /(?-x: )/x
2525          /(?x)(?-x: \s*#\s*)/
2526          /((?i)[\x{c0}])/8
2527          /(?i:[\x{c0}])/8
2528    
2529        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2530        is now done differently.
2531    
2532     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2533        wrapper classes; (b) implement a new function in the C++ scanner that is
2534        more efficient than the old way of doing things because it avoids levels of
2535        recursion in the regex matching; (c) add a paragraph to the documentation
2536        for the FullMatch() function.
2537    
2538     6. The escape sequence \n was being treated as whatever was defined as
2539        "newline". Not only was this contrary to the documentation, which states
2540        that \n is character 10 (hex 0A), but it also went horribly wrong when
2541        "newline" was defined as CRLF. This has been fixed.
2542    
2543     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2544        was being set to -1 for the "end of line" case (supposedly a value that no
2545        character can have). Though this value is never used (the check for end of
2546        line is "zero bytes in current character"), it caused compiler complaints.
2547        I've changed it to 0xffffffff.
2548    
2549     8. In pcre_version.c, the version string was being built by a sequence of
2550        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2551        string (as it is for production releases) called a macro with an empty
2552        argument. The C standard says the result of this is undefined. The gcc
2553        compiler treats it as an empty string (which was what was wanted) but it is
2554        reported that Visual C gives an error. The source has been hacked around to
2555        avoid this problem.
2556    
2557     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2558        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2559        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2560        of them did).
2561    
2562    10. Originally, pcretest opened its input and output without "b"; then I was
2563        told that "b" was needed in some environments, so it was added for release
2564        5.0 to both the input and output. (It makes no difference on Unix-like
2565        systems.) Later I was told that it is wrong for the input on Windows. I've
2566        now abstracted the modes into two macros, to make it easier to fiddle with
2567        them, and removed "b" from the input mode under Windows.
2568    
2569    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2570    
2571    12. Added -help and --help to pcretest as an official way of being reminded
2572        of the options.
2573    
2574    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2575        and pcrecpp.cc because they annoy compilers at high warning levels.
2576    
2577    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2578    
2579    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2580        scripts are not C programs :-) and which was not noticed because it works
2581        on Linux.
2582    
2583    16. pcretest is supposed to handle any length of pattern and data line (as one
2584        line or as a continued sequence of lines) by extending its input buffer if
2585        necessary. This feature was broken for very long pattern lines, leading to
2586        a string of junk being passed to pcre_compile() if the pattern was longer
2587        than about 50K.
2588    
2589    17. I have done a major re-factoring of the way pcre_compile() computes the
2590        amount of memory needed for a compiled pattern. Previously, there was code
2591        that made a preliminary scan of the pattern in order to do this. That was
2592        OK when PCRE was new, but as the facilities have expanded, it has become
2593        harder and harder to keep it in step with the real compile phase, and there
2594        have been a number of bugs (see for example, 4 above). I have now found a
2595        cunning way of running the real compile function in a "fake" mode that
2596        enables it to compute how much memory it would need, while actually only
2597        ever using a few hundred bytes of working memory and without too many
2598        tests of the mode. This should make future maintenance and development
2599        easier. A side effect of this work is that the limit of 200 on the nesting
2600        depth of parentheses has been removed (though this was never a serious
2601        limitation, I suspect). However, there is a downside: pcre_compile() now
2602        runs more slowly than before (30% or more, depending on the pattern). I
2603        hope this isn't a big issue. There is no effect on runtime performance.
2604    
2605    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2606        newline (only possible for the last line of a file) and it was a
2607        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2608    
2609    19. Added additional timing features to pcretest. (1) The -tm option now times
2610        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2611        separate command line item, by a number that specifies the number of
2612        repeats to use when timing. The default is 50000; this gives better
2613        precision, but takes uncomfortably long for very large patterns.
2614    
2615    20. Extended pcre_study() to be more clever in cases where a branch of a
2616        subpattern has no definite first character. For example, (a*|b*)[cd] would
2617        previously give no result from pcre_study(). Now it recognizes that the
2618        first character must be a, b, c, or d.
2619    
2620    21. There was an incorrect error "recursive call could loop indefinitely" if
2621        a subpattern (or the entire pattern) that was being tested for matching an
2622        empty string contained only one non-empty item after a nested subpattern.
2623        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2624        incorrectly, because the \d was being skipped in the check.
2625    
2626    22. The pcretest program now has a new pattern option /B and a command line
2627        option -b, which is equivalent to adding /B to every pattern. This causes
2628        it to show the compiled bytecode, without the additional information that
2629        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2630        is the same as /B/I).
2631    
2632    23. A new optimization is now able automatically to treat some sequences such
2633        as a*b as a*+b. More specifically, if something simple (such as a character
2634        or a simple class like \d) has an unlimited quantifier, and is followed by
2635        something that cannot possibly match the quantified thing, the quantifier
2636        is automatically "possessified".
2637    
2638    24. A recursive reference to a subpattern whose number was greater than 39
2639        went wrong under certain circumstances in UTF-8 mode. This bug could also
2640        have affected the operation of pcre_study().
2641    
2642    25. Realized that a little bit of performance could be had by replacing
2643        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2644    
2645    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2646    
2647    27. Possessive quantifiers such as a++ were previously implemented by turning
2648        them into atomic groups such as ($>a+). Now they have their own opcodes,
2649        which improves performance. This includes the automatically created ones
2650        from 23 above.
2651    
2652    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2653        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2654        the first matched character to be a colon. This applied both to named and
2655        numbered groups.
2656    
2657    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2658    
2659    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2660        building PCRE on Windows easier, so I have included it in the distribution.
2661    
2662    31. There is now a check in pcretest against a ridiculously large number being
2663        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2664        loop, the loop is abandoned.
2665    
2666    32. Forward references to subpatterns in conditions such as (?(2)...) where
2667        subpattern 2 is defined later cause pcre_compile() to search forwards in
2668        the pattern for the relevant set of parentheses. This search went wrong
2669        when there were unescaped parentheses in a character class, parentheses
2670        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2671    
2672    33. "Subroutine" calls and backreferences were previously restricted to
2673        referencing subpatterns earlier in the regex. This restriction has now
2674        been removed.
2675    
2676    34. Added a number of extra features that are going to be in Perl 5.10. On the
2677        whole, these are just syntactic alternatives for features that PCRE had
2678        previously implemented using the Python syntax or my own invention. The
2679        other formats are all retained for compatibility.
2680    
2681        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2682            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2683            also .NET compatible.
2684    
2685        (b) A recursion or subroutine call to a named group can now be defined as
2686            (?&name) as well as (?P>name).
2687    
2688        (c) A backreference to a named group can now be defined as \k<name> or
2689            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2690            5.10, are also .NET compatible.
2691    
2692        (d) A conditional reference to a named group can now use the syntax
2693            (?(<name>) or (?('name') as well as (?(name).
2694    
2695        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2696            groups (named and numbered) that are never evaluated inline, but can be
2697            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2698            is always false. There may be only one alternative in such a group.
2699    
2700        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2701            as the simple (?(R). The condition is true only if the most recent
2702            recursion is that of the given number or name. It does not search out
2703            through the entire recursion stack.
2704    
2705        (g) The escape \gN or \g{N} has been added, where N is a positive or
2706            negative number, specifying an absolute or relative reference.
2707    
2708    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2709        some "unreachable code" warnings.
2710    
2711    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2712        things, this adds five new scripts.
2713    
2714    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2715        There were also incompatibilities regarding the handling of \Q..\E inside
2716        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2717        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2718    
2719    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2720        matches an empty string, and forcibly breaks the loop. There were bugs in
2721        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2722        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2723        separate and independent bugs (that affected different cases) have been
2724        fixed.
2725    
2726    39. Refactored the code to abolish the use of different opcodes for small
2727        capturing bracket numbers. This is a tidy that I avoided doing when I
2728        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2729        The new approach is not only tidier, it makes it possible to reduce the
2730        memory needed to fix the previous bug (38).
2731    
2732    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2733        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2734        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2735        mode.
2736    
2737    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2738        report.
2739    
2740    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2741        copy construction and assignment in the C++ wrapper.
2742    
2743    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2744        couple of bugs that could have given wrong results in the "--newline=crlf"
2745        case.
2746    
2747    44. Added a number of casts and did some reorganization of signed/unsigned int
2748        variables following suggestions from Dair Grant. Also renamed the variable
2749        "this" as "item" because it is a C++ keyword.
2750    
2751    45. Arranged for dftables to add
2752    
2753          #include "pcre_internal.h"
2754    
2755        to pcre_chartables.c because without it, gcc 4.x may remove the array
2756        definition from the final binary if PCRE is built into a static library and
2757        dead code stripping is activated.
2758    
2759    46. For an unanchored pattern, if a match attempt fails at the start of a
2760        newline sequence, and the newline setting is CRLF or ANY, and the next two
2761        characters are CRLF, advance by two characters instead of one.
2762    
2763    
2764    Version 6.7 04-Jul-06
2765    ---------------------
2766    
2767     1. In order to handle tests when input lines are enormously long, pcretest has
2768        been re-factored so that it automatically extends its buffers when
2769        necessary. The code is crude, but this _is_ just a test program. The
2770        default size has been increased from 32K to 50K.
2771    
2772     2. The code in pcre_study() was using the value of the re argument before
2773        testing it for NULL. (Of course, in any sensible call of the function, it
2774        won't be NULL.)
2775    
2776     3. The memmove() emulation function in pcre_internal.h, which is used on
2777        systems that lack both memmove() and bcopy() - that is, hardly ever -
2778        was missing a "static" storage class specifier.
2779    
2780     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2781        containing an extended class (one that cannot be represented by a bitmap
2782        because it contains high-valued characters or Unicode property items, e.g.
2783        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2784        pattern, but PCRE should not loop if you do not (it no longer does).
2785        [Detail: two cases were found: (a) a repeated subpattern containing an
2786        extended class; (b) a recursive reference to a subpattern that followed a
2787        previous extended class. It wasn't skipping over the extended class
2788        correctly when UTF-8 mode was not set.]
2789    
2790     5. A negated single-character class was not being recognized as fixed-length
2791        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2792        compile error "lookbehind assertion is not fixed length".
2793    
2794     6. The RunPerlTest auxiliary script was showing an unexpected difference
2795        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2796        write a Perl script that can interpret lines of an input file either as
2797        byte characters or as UTF-8, which is what "perltest" was being required to
2798        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2799        can't do is switch easily at run time between having the "use utf8;" pragma
2800        or not. In the end, I fudged it by using the RunPerlTest script to insert
2801        "use utf8;" explicitly for the UTF-8 tests.
2802    
2803     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2804        the end of the subject string, contrary to the documentation and to what
2805        Perl does. This was true of both matching functions. Now it matches only at
2806        the start of the subject and immediately after *internal* newlines.
2807    
2808     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2809        a pointer to an int instead of a pointer to an unsigned long int. This
2810        caused problems on 64-bit systems.
2811    
2812     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2813        instance of the 'standard' template library not being so standard".
2814    
2815    10. There was no check on the number of named subpatterns nor the maximum
2816        length of a subpattern name. The product of these values is used to compute
2817        the size of the memory block for a compiled pattern. By supplying a very
2818        long subpattern name and a large number of named subpatterns, the size
2819        computation could be caused to overflow. This is now prevented by limiting
2820        the length of names to 32 characters, and the number of named subpatterns
2821        to 10,000.
2822    
2823    11. Subpatterns that are repeated with specific counts have to be replicated in
2824        the compiled pattern. The size of memory for this was computed from the
2825        length of the subpattern and the repeat count. The latter is limited to
2826        65535, but there was no limit on the former, meaning that integer overflow
2827        could in principle occur. The compiled length of a repeated subpattern is
2828        now limited to 30,000 bytes in order to prevent this.
2829    
2830    12. Added the optional facility to have named substrings with the same name.
2831    
2832    13. Added the ability to use a named substring as a condition, using the
2833        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2834        are numbers (not recommended). Forward references are permitted.
2835    
2836    14. Added forward references in named backreferences (if you see what I mean).
2837    
2838    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2839        pattern could run off the end of the subject. For example, the pattern
2840        "(?s)(.{1,5})"8 did this with the subject "ab".
2841    
2842    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2843        PCRE_CASELESS was set when matching characters that were quantified with ?
2844        or *.
2845    
2846    17. A character class other than a single negated character that had a minimum
2847        but no maximum quantifier - for example [ab]{6,} - was not handled
2848        correctly by pce_dfa_exec(). It would match only one character.
2849    
2850    18. A valid (though odd) pattern that looked like a POSIX character
2851        class but used an invalid character after [ (for example [[,abc,]]) caused
2852        pcre_compile() to give the error "Failed: internal error: code overflow" or
2853        in some cases to crash with a glibc free() error. This could even happen if
2854        the pattern terminated after [[ but there just happened to be a sequence of
2855        letters, a binary zero, and a closing ] in the memory that followed.
2856    
2857    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2858        over the years. Originally (before any Unicode support), just the bottom 8
2859        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2860        output from "man perlunicode" includes this:
2861    
2862          The regular expression compiler produces polymorphic opcodes.  That
2863          is, the pattern adapts to the data and automatically switches to
2864          the Unicode character scheme when presented with Unicode data--or
2865          instead uses a traditional byte scheme when presented with byte
2866          data.
2867    
2868        Sadly, a wide octal escape does not cause a switch, and in a string with
2869        no other multibyte characters, these octal escapes are treated as before.
2870        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2871        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2872        Unicode string.
2873    
2874        I have not perpetrated such confusion in PCRE. Up till now, it took just
2875        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2876        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2877        translate to the appropriate multibyte character.
2878    
2879    29. Applied some refactoring to reduce the number of warnings from Microsoft
2880        and Borland compilers. This has included removing the fudge introduced
2881        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2882        a warning about an unused variable.
2883    
2884    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2885        characters since release 4.0, because Perl (from release 5.004) does not.
2886        [Or at least, is documented not to: some releases seem to be in conflict
2887        with the documentation.] However, when a pattern was studied with
2888        pcre_study() and all its branches started with \s, PCRE still included VT
2889        as a possible starting character. Of course, this did no harm; it just
2890        caused an unnecessary match attempt.
2891    
2892    22. Removed a now-redundant internal flag bit that recorded the fact that case
2893        dependency changed within the pattern. This was once needed for "required
2894        byte" processing, but is no longer used. This recovers a now-scarce options
2895        bit. Also moved the least significant internal flag bit to the most-
2896        significant bit of the word, which was not previously used (hangover from
2897        the days when it was an int rather than a uint) to free up another bit for
2898        the future.
2899    
2900    23. Added support for CRLF line endings as well as CR and LF. As well as the
2901        default being selectable at build time, it can now be changed at runtime
2902        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2903        specify that it is scanning data with non-default line endings.
2904    
2905    24. Changed the definition of CXXLINK to make it agree with the definition of
2906        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2907    
2908    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2909        recursions. This makes a big different to stack usage for some patterns.
2910    
2911    26. If a subpattern containing a named recursion or subroutine reference such
2912        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2913        the space required for the compiled pattern went wrong and gave too small a
2914        value. Depending on the environment, this could lead to "Failed: internal
2915        error: code overflow at offset 49" or "glibc detected double free or
2916        corruption" errors.
2917    
2918    27. Applied patches from Google (a) to support the new newline modes and (b) to
2919        advance over multibyte UTF-8 characters in GlobalReplace.
2920    
2921    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2922        difference for some implementation of PCRE in some Windows version.
2923    
2924    29. Added some extra testing facilities to pcretest:
2925    
2926        \q<number>   in a data line sets the "match limit" value
2927        \Q<number>   in a data line sets the "match recursion limt" value
2928        -S <number>  sets the stack size, where <number> is in megabytes
2929    
2930        The -S option isn't available for Windows.
2931    
2932    
2933    Version 6.6 06-Feb-06
2934    ---------------------
2935    
2936     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2937        in pcreposix.h. I have copied the definition from pcre.h.
2938    
2939     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2940        because pcre.h is no longer a built file.
2941    
2942     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2943        not normally included in the compiled code.
2944    
2945    
2946    Version 6.5 01-Feb-06
2947    ---------------------
2948    
2949     1. When using the partial match feature with pcre_dfa_exec(), it was not
2950        anchoring the second and subsequent partial matches at the new starting
2951        point. This could lead to incorrect results. For example, with the pattern
2952        /1234/, partially matching against "123" and then "a4" gave a match.
2953    
2954     2. Changes to pcregrep:
2955    
2956        (a) All non-match returns from pcre_exec() were being treated as failures
2957            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2958            error message is output. Some extra information is given for the
2959            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2960            probably the only errors that are likely to be caused by users (by
2961            specifying a regex that has nested indefinite repeats, for instance).
2962            If there are more than 20 of these errors, pcregrep is abandoned.
2963    
2964        (b) A binary zero was treated as data while matching, but terminated the
2965            output line if it was written out. This has been fixed: binary zeroes
2966            are now no different to any other data bytes.
2967    
2968        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2969            used to set a locale for matching. The --locale=xxxx long option has
2970            been added (no short equivalent) to specify a locale explicitly on the
2971            pcregrep command, overriding the environment variables.
2972    
2973        (d) When -B was used with -n, some line numbers in the output were one less
2974            than they should have been.
2975    
2976        (e) Added the -o (--only-matching) option.
2977    
2978        (f) If -A or -C was used with -c (count only), some lines of context were
2979            accidentally printed for the final match.
2980    
2981        (g) Added the -H (--with-filename) option.
2982    
2983        (h) The combination of options -rh failed to suppress file names for files
2984            that were found from directory arguments.
2985    
2986        (i) Added the -D (--devices) and -d (--directories) options.
2987    
2988        (j) Added the -F (--fixed-strings) option.
2989    
2990        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2991    
2992        (l) Added the --colo(u)r option.
2993    
2994        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2995            is not present by default.
2996    
2997     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2998        items such as (?R) or (?1), when the recursion could match a number of
2999        alternatives. If it matched one of the alternatives, but subsequently,
3000        outside the recursion, there was a failure, the code tried to back up into
3001        the recursion. However, because of the way PCRE is implemented, this is not
3002        possible, and the result was an incorrect result from the match.
3003    
3004        In order to prevent this happening, the specification of recursion has
3005        been changed so that all such subpatterns are automatically treated as
3006        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3007    
3008     4. I had overlooked the fact that, in some locales, there are characters for
3009        which isalpha() is true but neither isupper() nor islower() are true. In
3010        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3011        and ordfeminine) are like this. This affected the treatment of \w and \W
3012        when they appeared in character classes, but not when they appeared outside
3013        a character class. The bit map for "word" characters is now created
3014        separately from the results of isalnum() instead of just taking it from the
3015        upper, lower, and digit maps. (Plus the underscore character, of course.)
3016    
3017     5. The above bug also affected the handling of POSIX character classes such as
3018        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3019        permanent tables. Instead, the bit maps for such a class were previously
3020        created as the appropriate unions of the upper, lower, and digit bitmaps.
3021        Now they are created by subtraction from the [[:word:]] class, which has
3022        its own bitmap.
3023    
3024     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3025        It is created by subtracting the vertical space characters (\x09, \x0a,
3026        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3027        subtraction was done in the overall bitmap for a character class, meaning
3028        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3029        be recognized. This bug has been fixed.
3030    
3031     7. Patches from the folks at Google:
3032    
3033          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3034          real life, but is still worth protecting against".
3035    
3036          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3037          regular expressions".
3038    
3039          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3040          have it.
3041    
3042          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3043          "configure" and the latter not, in order to fix a problem somebody had
3044          with compiling the Arg class on HP-UX.
3045    
3046          (e) Improve the error-handling of the C++ wrapper a little bit.
3047    
3048          (f) New tests for checking recursion limiting.
3049    
3050     8. The pcre_memmove() function, which is used only if the environment does not
3051        have a standard memmove() function (and is therefore rarely compiled),
3052        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3053        returning a result (though PCRE never actually uses the result).
3054    
3055     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3056        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3057        returned instead of calling malloc() with an overflowing number that would
3058        most likely cause subsequent chaos.
3059    
3060    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3061    
3062    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3063        with this option is matched, the nmatch and pmatch options of regexec() are
3064        ignored.
3065    
3066    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3067        provided in case anyone wants to the the POSIX interface with UTF-8
3068        strings.
3069    
3070    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3071        C++ linking (needed for some HP-UX environments).
3072    
3073    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3074        (unused parameter) and in the pcre_printint() function (omitted "default"
3075        switch label when the default is to do nothing).
3076    
3077    15. Added some code to make it possible, when PCRE is compiled as a C++
3078        library, to replace subject pointers for pcre_exec() with a smart pointer
3079        class, thus making it possible to process discontinuous strings.
3080    
3081    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3082        much the same function. They were added by different people who were trying
3083        to make PCRE easy to compile on non-Unix systems. It has been suggested
3084        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3085        for compiling on Windows systems. I have therefore replaced it with
3086        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3087        defaults to "extern" for C or "extern C" for C++, which works fine on
3088        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3089        SCOPE with something explicit in config.h. In addition:
3090    
3091        (a) pcreposix.h still had just "extern" instead of either of these macros;
3092            I have replaced it with PCRE_DATA_SCOPE.
3093    
3094        (b) Functions such as _pcre_xclass(), which are internal to the library,
3095            but external in the C sense, all had PCRE_EXPORT in their definitions.
3096            This is apparently wrong for the Windows case, so I have removed it.
3097            (It makes no difference on Unix-like systems.)
3098    
3099    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3100        of recursive calls to match(). This is different to MATCH_LIMIT because
3101        that limits the total number of calls to match(), not all of which increase
3102        the depth of recursion. Limiting the recursion depth limits the amount of
3103        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3104        when PCRE is compiled, and changed at run time. A patch from Google adds
3105        this functionality to the C++ interface.
3106    
3107    18. Changes to the handling of Unicode character properties:
3108    
3109        (a) Updated the table to Unicode 4.1.0.
3110    
3111        (b) Recognize characters that are not in the table as "Cn" (undefined).
3112    
3113        (c) I revised the way the table is implemented to a much improved format
3114            which includes recognition of ranges. It now supports the ranges that
3115            are defined in UnicodeData.txt, and it also amalgamates other
3116            characters into ranges. This has reduced the number of entries in the
3117            table from around 16,000 to around 3,000, thus reducing its size
3118            considerably. I realized I did not need to use a tree structure after
3119            all - a binary chop search is just as efficient. Having reduced the
3120            number of entries, I extended their size from 6 bytes to 8 bytes to
3121            allow for more data.
3122    
3123        (d) Added support for Unicode script names via properties such as \p{Han}.
3124    
3125    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3126        matching that character.
3127    
3128    20. When matching a repeated Unicode property with a minimum greater than zero,
3129        (for example \pL{2,}), PCRE could look past the end of the subject if it
3130        reached it while seeking the minimum number of characters. This could
3131        happen only if some of the characters were more than one byte long, because
3132        there is a check for at least the minimum number of bytes.
3133    
3134    21. Refactored the implementation of \p and \P so as to be more general, to
3135        allow for more different types of property in future. This has changed the
3136        compiled form incompatibly. Anybody with saved compiled patterns that use
3137        \p or \P will have to recompile them.
3138    
3139    22. Added "Any" and "L&" to the supported property types.
3140    
3141    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3142        but give a compile time error if the value is greater than 0xff.
3143    
3144    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3145        accidentally not being installed or uninstalled.
3146    
3147    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3148        made were to insert the current release number. This seemed silly, because
3149        it made things harder for people building PCRE on systems that don't run
3150        "configure". I have turned pcre.h into a distributed file, no longer built
3151        by "configure", with the version identification directly included. There is
3152        no longer a pcre.h.in file.
3153    
3154        However, this change necessitated a change to the pcre-config script as
3155        well. It is built from pcre-config.in, and one of the substitutions was the
3156        release number. I have updated configure.ac so that ./configure now finds
3157        the release number by grepping pcre.h.
3158    
3159    26. Added the ability to run the tests under valgrind.
3160    
3161    
3162    Version 6.4 05-Sep-05
3163    ---------------------
3164    
3165     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3166        "--" to be printed when multiple files were scanned, even when none of the
3167        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3168        consider it to be a bug, and have restored the previous behaviour.
3169    
3170     2. A couple of code tidies to get rid of compiler warnings.
3171    
3172     3. The pcretest program used to cheat by referring to symbols in the library
3173        whose names begin with _pcre_. These are internal symbols that are not
3174        really supposed to be visible externally, and in some environments it is
3175        possible to suppress them. The cheating is now confined to including
3176        certain files from the library's source, which is a bit cleaner.
3177    
3178     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3179        file's purpose clearer.
3180    
3181     5. Reorganized pcre_ucp_findchar().
3182    
3183    
3184    Version 6.3 15-Aug-05
3185    ---------------------
3186    
3187     1. The file libpcre.pc.in did not have general read permission in the tarball.
3188    
3189     2. There were some problems when building without C++ support:
3190    
3191        (a) If C++ support was not built, "make install" and "make test" still
3192            tried to test it.
3193    
3194        (b) There were problems when the value of CXX was explicitly set. Some
3195            changes have been made to try to fix these, and ...
3196    
3197        (c) --disable-cpp can now be used to explicitly disable C++ support.
3198    
3199        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3200            backslash in a target when C++ was disabled. This confuses some
3201            versions of "make", apparently. Using an intermediate variable solves
3202            this. (Same for CPP_LOBJ.)
3203    
3204     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3205        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3206        necessary on certain architectures.
3207    
3208     4. Added a setting of -export-symbols-regex to the link command to remove
3209        those symbols that are exported in the C sense, but actually are local
3210        within the library, and not documented. Their names all begin with
3211        "_pcre_". This is not a perfect job, because (a) we have to except some
3212        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3213        available (and never for static libraries). I have made a note to try to
3214        find a way round (a) in the future.
3215    
3216    
3217    Version 6.2 01-Aug-05
3218    ---------------------
3219    
3220     1. There was no test for integer overflow of quantifier values. A construction
3221        such as {1111111111111111} would give undefined results. What is worse, if
3222        a minimum quantifier for a parenthesized subpattern overflowed and became
3223        negative, the calculation of the memory size went wrong. This could have
3224        led to memory overwriting.
3225    
3226     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3227    
3228     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3229        operating environments where this matters.
3230    
3231     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3232        PCRE options from within the C++ wrapper.
3233    
3234     5. Named capturing subpatterns were not being correctly counted when a pattern
3235        was compiled. This caused two problems: (a) If there were more than 100
3236        such subpatterns, the calculation of the memory needed for the whole
3237        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3238        back references of the form \12, where the number was greater than 9, were
3239        not recognized as back references, even though there were sufficient
3240        previous subpatterns.
3241    
3242     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3243        versions of gcc, e.g. 2.95.4.
3244    
3245    
3246    Version 6.1 21-Jun-05
3247    ---------------------
3248    
3249     1. There was one reference to the variable "posix" in pcretest.c that was not
3250        surrounded by "#if !defined NOPOSIX".
3251    
3252     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3253        the cross-check on the old pcre_info() function, for the benefit of the
3254        cut-down version of PCRE that is currently imported into Exim.
3255    
3256     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3257        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3258        patterns that start (?i) but I don't think that matters. The use of (?i) is
3259        just an example; this all applies to the other options as well.
3260    
3261     4. Since libtool seems to echo the compile commands it is issuing, the output
3262        from "make" can be reduced a bit by putting "@" in front of each libtool
3263        compile command.
3264    
3265     5. Patch from the folks at Google for configure.in to be a bit more thorough
3266        in checking for a suitable C++ installation before trying to compile the
3267        C++ stuff. This should fix a reported problem when a compiler was present,
3268        but no suitable headers.
3269    
3270     6. The man pages all had just "PCRE" as their title. I have changed them to
3271        be the relevant file name. I have also arranged that these names are
3272        retained in the file doc/pcre.txt, which is a concatenation in text format
3273        of all the man pages except the little individual ones for each function.
3274    
3275     7. The NON-UNIX-USE file had not been updated for the different set of source
3276        files that come with release 6. I also added a few comments about the C++
3277        wrapper.
3278    
3279    
3280    Version 6.0 07-Jun-05
3281    ---------------------
3282    
3283     1. Some minor internal re-organization to help with my DFA experiments.
3284    
3285     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3286        didn't matter for the library itself when fully configured, but did matter
3287        when compiling without UCP support, or within Exim, where the ucp files are
3288        not imported.
3289    
3290     3. Refactoring of the library code to split up the various functions into
3291        different source modules. The addition of the new DFA matching code (see
3292        below) to a single monolithic source would have made it really too
3293        unwieldy, quite apart from causing all the code to be include in a
3294        statically linked application, when only some functions are used. This is
3295        relevant even without the DFA addition now that patterns can be compiled in
3296        one application and matched in another.
3297    
3298        The downside of splitting up is that there have to be some external
3299        functions and data tables that are used internally in different modules of
3300        the library but which are not part of the API. These have all had their
3301        names changed to start with "_pcre_" so that they are unlikely to clash
3302        with other external names.
3303    
3304     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3305        a different (DFA) algorithm. Although it is slower than the original
3306        function, it does have some advantages for certain types of matching
3307        problem.
3308    
3309     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3310        including restarting after a partial match.
3311    
3312     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3313        defined when compiling for Windows was sent to me. I have put it into the
3314        code, though I have no means of testing or verifying it.
3315    
3316     7. Added the pcre_refcount() auxiliary function.
3317    
3318     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3319        match before or at the first newline in the subject string. In pcretest,
3320        the /f option on a pattern can be used to set this.
3321    
3322     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3323        would behave wrongly. This has been present in PCRE since release 4.0.
3324    
3325    10. A number of changes to the pcregrep command:
3326    
3327        (a) Refactored how -x works; insert ^(...)$ instead of setting
3328            PCRE_ANCHORED and checking the length, in preparation for adding
3329            something similar for -w.
3330    
3331        (b) Added the -w (match as a word) option.
3332    
3333        (c) Refactored the way lines are read and buffered so as to have more
3334            than one at a time available.
3335    
3336        (d) Implemented a pcregrep test script.
3337    
3338        (e) Added the -M (multiline match) option. This allows patterns to match
3339            over several lines of the subject. The buffering ensures that at least
3340            8K, or the rest of the document (whichever is the shorter) is available
3341            for matching (and similarly the previous 8K for lookbehind assertions).
3342    
3343        (f) Changed the --help output so that it now says
3344    
3345              -w, --word-regex(p)
3346    
3347            instead of two lines, one with "regex" and the other with "regexp"
3348            because that confused at least one person since the short forms are the
3349            same. (This required a bit of code, as the output is generated
3350            automatically from a table. It wasn't just a text change.)
3351    
3352        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3353            option but starts with a hyphen. Could be a pattern or a path name
3354            starting with a hyphen, for instance.
3355    
3356        (h) "-" can be given as a file name to represent stdin.
3357    
3358        (i) When file names are being printed, "(standard input)" is used for
3359            the standard input, for compatibility with GNU grep. Previously
3360            "<stdin>" was used.
3361    
3362        (j) The option --label=xxx can be used to supply a name to be used for
3363            stdin when file names are being printed. There is no short form.
3364    
3365        (k) Re-factored the options decoding logic because we are going to add
3366            two more options that take data. Such options can now be given in four
3367            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3368    
3369        (l) Added the -A, -B, and -C options for requesting that lines of context
3370            around matches be printed.
3371    
3372        (m) Added the -L option to print the names of files that do not contain
3373            any matching lines, that is, the complement of -l.
3374    
3375        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3376            continue to scan other files.
3377    
3378        (o) The -s option was incorrectly implemented. For compatibility with other
3379            greps, it now suppresses the error message for a non-existent or non-
3380            accessible file (but not the return code). There is a new option called
3381            -q that suppresses the output of matching lines, which was what -s was
3382            previously doing.
3383    
3384        (p) Added --include and --exclude options to specify files for inclusion
3385            and exclusion when recursing.
3386    
3387    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3388        Hopefully, it now does.
3389    
3390    12. Missing cast in pcre_study().
3391    
3392    13. Added an "uninstall" target to the makefile.
3393    
3394    14. Replaced "extern" in the function prototypes in Makefile.in with
3395        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3396        world, but is set differently for Windows.
3397    
3398    15. Added a second compiling function called pcre_compile2(). The only
3399        difference is that it has an extra argument, which is a pointer to an
3400        integer error code. When there is a compile-time failure, this is set
3401        non-zero, in addition to the error test pointer being set to point to an
3402        error message. The new argument may be NULL if no error number is required
3403        (but then you may as well call pcre_compile(), which is now just a
3404        wrapper). This facility is provided because some applications need a
3405        numeric error indication, but it has also enabled me to tidy up the way
3406        compile-time errors are handled in the POSIX wrapper.
3407    
3408    16. Added VPATH=.libs to the makefile; this should help when building with one
3409        prefix path and installing with another. (Or so I'm told by someone who
3410        knows more about this stuff than I do.)
3411    
3412    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3413        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3414        match everything, including newlines. This is not POSIX-compatible, but
3415        somebody wanted the feature. From pcretest it can be activated by using
3416        both the P and the s flags.
3417    
3418    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3419    
3420    19. libpcre.pc was being incorrectly installed as executable.
3421    
3422    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3423        it now also looks for '\r' so that it will work unmodified on Windows.
3424    
3425    21. Added Google's contributed C++ wrapper to the distribution.
3426    
3427    22. Added some untidy missing memory free() calls in pcretest, to keep
3428        Electric Fence happy when testing.
3429    
3430    
3431    
3432    Version 5.0 13-Sep-04
3433    ---------------------
3434    
3435     1. Internal change: literal characters are no longer packed up into items
3436        containing multiple characters in a single byte-string. Each character
3437        is now matched using a separate opcode. However, there may be more than one
3438        byte in the character in UTF-8 mode.
3439    
3440     2. The pcre_callout_block structure has two new fields: pattern_position and
3441        next_item_length. These contain the offset in the pattern to the next match
3442        item, and its length, respectively.
3443    
3444     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3445        insertion of callouts before each pattern item. Added the /C option to
3446        pcretest to make use of this.
3447    
3448     4. On the advice of a Windows user, the lines
3449    
3450          #if defined(_WIN32) || defined(WIN32)
3451          _setmode( _fileno( stdout ), 0x8000 );
3452          #endif  /* defined(_WIN32) || defined(WIN32) */
3453    
3454        have been added to the source of pcretest. This apparently does useful
3455        magic in relation to line terminators.
3456    
3457     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3458        for the benefit of those environments where the "b" makes a difference.
3459    
3460     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3461        to know about it. I have put a hack into configure.in that adds in code
3462        to set GCC=yes if CC=icc. This seems to end up at a point in the
3463        generated configure script that is early enough to affect the setting of
3464        compiler options, which is what is needed, but I have no means of testing
3465        whether it really works. (The user who reported this had patched the
3466        generated configure script, which of course I cannot do.)
3467    
3468        LATER: After change 22 below (new libtool files), the configure script
3469        seems to know about icc (and also ecc). Therefore, I have commented out
3470        this hack in configure.in.
3471    
3472     7. Added support for pkg-config (2 patches were sent in).
3473    
3474     8. Negated POSIX character classes that used a combination of internal tables
3475        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3476        [[:^ascii]]. Typically, they would match almost any characters. The other
3477        POSIX classes were not broken in this way.
3478    
3479     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3480        to find the match, as PCRE was deluded into thinking that the match had to
3481        start at the start point or following a newline. The same bug applied to
3482        patterns with negative forward assertions or any backward assertions
3483        preceding ".*" at the start, unless the pattern required a fixed first
3484        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3485    
3486    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3487        starting at the last subject character, bytes beyond the end of the subject
3488        string were read.
3489    
3490    11. Renamed the variable "class" as "classbits" to make life easier for C++
3491        users. (Previously there was a macro definition, but it apparently wasn't
3492        enough.)
3493    
3494    12. Added the new field "tables" to the extra data so that tables can be passed
3495        in at exec time, or the internal tables can be re-selected. This allows
3496        a compiled regex to be saved and re-used at a later time by a different
3497        program that might have everything at different addresses.
3498    
3499    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3500        -R library as well as a -L library.
3501    
3502    14. The debugging options of pcretest (-d on the command line or D on a
3503        pattern) showed incorrect output for anything following an extended class
3504        that contained multibyte characters and which was followed by a quantifier.
3505    
3506    15. Added optional support for general category Unicode character properties
3507        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3508        support. It adds about 90K to the size of the library. The meanings of the
3509        inbuilt class escapes such as \d and \s have NOT been changed.
3510    
3511    16. Updated pcredemo.c to include calls to free() to release the memory for the
3512        compiled pattern.
3513    
3514    17. The generated file chartables.c was being created in the source directory
3515        instead of in the building directory. This caused the build to fail if the
3516        source directory was different from the building directory, and was
3517        read-only.
3518    
3519    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3520        file. No doubt somebody will tell me if they don't make sense... Also added
3521        Dan Mooney's comments about building on OpenVMS.
3522    
3523    19. Added support for partial matching via the PCRE_PARTIAL option for
3524        pcre_exec() and the \P data escape in pcretest.
3525    
3526    20. Extended pcretest with 3 new pattern features:
3527    
3528        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3529              write the compiled pattern to the file whose name is "rest-of-line".
3530              This is a straight binary dump of the data, with the saved pointer to
3531              the character tables forced to be NULL. The study data, if any, is
3532              written too. After writing, pcretest reads a new pattern.
3533    
3534        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3535              compiled pattern from the given file. There must not be any
3536              occurrences of "<" in the file name (pretty unlikely); if there are,
3537              pcretest will instead treat the initial "<" as a pattern delimiter.
3538              After reading in the pattern, pcretest goes on to read data lines as
3539              usual.
3540    
3541        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3542              and 16-bit fields in a compiled pattern, to simulate a pattern that
3543              was compiled on a host of opposite endianness.
3544    
3545    21. The pcre-exec() function can now cope with patterns that were compiled on
3546        hosts of opposite endianness, with this restriction:
3547    
3548          As for any compiled expression that is saved and used later, the tables
3549          pointer field cannot be preserved; the extra_data field in the arguments
3550          to pcre_exec() should be used to pass in a tables address if a value
3551          other than the default internal tables were used at compile time.
3552    
3553    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3554        now diagnosed as an error. Previously, most of the time, a negative number
3555        would have been treated as zero, but if in addition "ovector" was passed as
3556        NULL, a crash could occur.
3557    
3558    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3559        new versions from the libtool 1.5 distribution (the last one is a copy of
3560        a file called libtool.m4). This seems to have fixed the need to patch
3561        "configure" to support Darwin 1.3 (which I used to do). However, I still
3562        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3563        workstation).
3564    
3565    24. Changed the PCRE licence to be the more standard "BSD" licence.
3566    
3567    
3568    Version 4.5 01-Dec-03
3569    ---------------------
3570    
3571     1. There has been some re-arrangement of the code for the match() function so
3572        that it can be compiled in a version that does not call itself recursively.
3573        Instead, it keeps those local variables that need separate instances for
3574        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3575        needs to "recurse". Keeping track of where control must go is done by means
3576        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3577        hide most of the details from the main code, and operates only if
3578        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3579        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of