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

Diff of /code/trunk/ChangeLog

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

revision 158 by ph10, Tue May 1 10:30:14 2007 UTC revision 1245 by zherczeg, Sat Feb 9 11:30:51 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Version 8.33 xx-xxxx-201x
5    -------------------------
6    
7    1.  Added 'U' to some constants that are compared to unsigned integers, to
8        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9        variables that are added to signed variables, to ensure the result is
10        signed and can be negated.
11    
12    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
13        CMake config files.
14    
15    3.  Revise the creation of config.h.generic so that all boolean macros are
16        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17        overriding via -D on the command line possible.
18    
19    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20        to unsigned int is reported to make a quite noticeable speed difference in
21        a specific Windows environment. Testing on Linux did also appear to show
22        some benefit (and it is clearly not harmful). Also fixed the definition of
23        Xop which should be unsigned.
24    
25    5.  Related to (4), changing the definition of the intermediate variable cc
26        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27        improvements.
28    
29    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
30        unnecessary spaces.
31    
32    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        WARNING: Callouts are not release ready! <- this line needs to be removed when it is.
47    
48    
49    Version 8.32 30-November-2012
50    -----------------------------
51    
52    1.  Improved JIT compiler optimizations for first character search and single
53        character iterators.
54    
55    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
56        Patch by Daniel Richard G.
57    
58    3.  Single character iterator optimizations in the JIT compiler.
59    
60    4.  Improved JIT compiler optimizations for character ranges.
61    
62    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
63        Reported by Giuseppe D'Angelo.
64    
65    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
66        of a line, was being set incorrectly in cases where .* appeared inside
67        atomic brackets at the start of a pattern, or where there was a subsequent
68        *PRUNE or *SKIP.
69    
70    7.  Improved instruction cache flush for POWER/PowerPC.
71        Patch by Daniel Richard G.
72    
73    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
74        grep:
75    
76        (a) There is now no limit to the number of patterns to be matched.
77    
78        (b) An error is given if a pattern is too long.
79    
80        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
81            are now supported.
82    
83        (d) --exclude-from and --include-from (multiple use) have been added.
84    
85        (e) Exclusions and inclusions now apply to all files and directories, not
86            just to those obtained from scanning a directory recursively.
87    
88        (f) Multiple uses of -f and --file-list are now supported.
89    
90        (g) In a Windows environment, the default for -d has been changed from
91            "read" (the GNU grep default) to "skip", because otherwise the presence
92            of a directory in the file list provokes an error.
93    
94        (h) The documentation has been revised and clarified in places.
95    
96    9.  Improve the matching speed of capturing brackets.
97    
98    10. Changed the meaning of \X so that it now matches a Unicode extended
99        grapheme cluster.
100    
101    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
102        out POSIX threads when JIT support is configured.
103    
104    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
105    
106    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
107        structure could go wrong in environments where size_t is not the same size
108        as int.
109    
110    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
111        set.
112    
113    15. The EBCDIC support had decayed; later updates to the code had included
114        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
115        general tidy up of EBCDIC-related issues, and the documentation was also
116        not quite right. There is now a test that can be run on ASCII systems to
117        check some of the EBCDIC-related things (but is it not a full test).
118    
119    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
120        in a small tidy to the code.
121    
122    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
123    
124    18. If the --only-matching (-o) option in pcregrep is specified multiple
125        times, each one causes appropriate output. For example, -o1 -o2 outputs the
126        substrings matched by the 1st and 2nd capturing parentheses. A separating
127        string can be specified by --om-separator (default empty).
128    
129    19. Improving the first n character searches.
130    
131    20. Turn case lists for horizontal and vertical white space into macros so that
132        they are defined only once.
133    
134    21. This set of changes together give more compatible Unicode case-folding
135        behaviour for characters that have more than one other case when UCP
136        support is available.
137    
138        (a) The Unicode property table now has offsets into a new table of sets of
139            three or more characters that are case-equivalent. The MultiStage2.py
140            script that generates these tables (the pcre_ucd.c file) now scans
141            CaseFolding.txt instead of UnicodeData.txt for character case
142            information.
143    
144        (b) The code for adding characters or ranges of characters to a character
145            class has been abstracted into a generalized function that also handles
146            case-independence. In UTF-mode with UCP support, this uses the new data
147            to handle characters with more than one other case.
148    
149        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
150            whose other case is greater than 256 are now correctly matched
151            caselessly. Previously, the high codepoint matched the low one, but not
152            vice versa.
153    
154        (d) The processing of \h, \H, \v, and \ in character classes now makes use
155            of the new class addition function, using character lists defined as
156            macros alongside the case definitions of 20 above.
157    
158        (e) Caseless back references now work with characters that have more than
159            one other case.
160    
161        (f) General caseless matching of characters with more than one other case
162            is supported.
163    
164    22. Unicode character properties were updated from Unicode 6.2.0
165    
166    23. Improved CMake support under Windows. Patch by Daniel Richard G.
167    
168    24. Add support for 32-bit character strings, and UTF-32
169    
170    25. Major JIT compiler update (code refactoring and bugfixing).
171        Experimental Sparc 32 support is added.
172    
173    26. Applied a modified version of Daniel Richard G's patch to create
174        pcre.h.generic and config.h.generic by "make" instead of in the
175        PrepareRelease script.
176    
177    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
178        pcre_compile.c when checking for a zero character.
179    
180    28. Introducing a native interface for JIT. Through this interface, the compiled
181        machine code can be directly executed. The purpose of this interface is to
182        provide fast pattern matching, so several sanity checks are not performed.
183        However, feature tests are still performed. The new interface provides
184        1.4x speedup compared to the old one.
185    
186    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
187        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
188        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
189    
190    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
191        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
192        the "old" RFC 2279). Instead, it ended up passing a negative length to
193        pcre_exec().
194    
195    31. Add support for GCC's visibility feature to hide internal functions.
196    
197    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
198        "unknown -C option" after outputting 0 or 1.
199    
200    33. There is now support for generating a code coverage report for the test
201        suite in environments where gcc is the compiler and lcov is installed. This
202        is mainly for the benefit of the developers.
203    
204    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
205        unaddressable using valgrind annotations, allowing valgrind to detect
206        invalid memory accesses. This is mainly for the benefit of the developers.
207    
208    25. (*UTF) can now be used to start a pattern in any of the three libraries.
209    
210    26. Give configure error if --enable-cpp but no C++ compiler found.
211    
212    
213    Version 8.31 06-July-2012
214    -------------------------
215    
216    1.  Fixing a wrong JIT test case and some compiler warnings.
217    
218    2.  Removed a bashism from the RunTest script.
219    
220    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
221        to unsigned type, result still unsigned" that was given by an MS compiler
222        on encountering the code "-sizeof(xxx)".
223    
224    4.  Partial matching support is added to the JIT compiler.
225    
226    5.  Fixed several bugs concerned with partial matching of items that consist
227        of more than one character:
228    
229        (a) /^(..)\1/ did not partially match "aba" because checking references was
230            done on an "all or nothing" basis. This also applied to repeated
231            references.
232    
233        (b) \R did not give a hard partial match if \r was found at the end of the
234            subject.
235    
236        (c) \X did not give a hard partial match after matching one or more
237            characters at the end of the subject.
238    
239        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
240            a partial match for the string "\r".
241    
242        (e) When newline was set to CRLF, the metacharacter "." did not recognize
243            a partial match for a CR character at the end of the subject string.
244    
245    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
246        running pcretest, the text "(JIT)" added to the output whenever JIT is
247        actually used to run the match.
248    
249    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
250        or /S+[+] with a digit between 1 and 7.
251    
252    8.  OP_NOT now supports any UTF character not just single-byte ones.
253    
254    9.  (*MARK) control verb is now supported by the JIT compiler.
255    
256    10. The command "./RunTest list" lists the available tests without actually
257        running any of them. (Because I keep forgetting what they all are.)
258    
259    11. Add PCRE_INFO_MAXLOOKBEHIND.
260    
261    12. Applied a (slightly modified) user-supplied patch that improves performance
262        when the heap is used for recursion (compiled with --disable-stack-for-
263        recursion). Instead of malloc and free for each heap frame each time a
264        logical recursion happens, frames are retained on a chain and re-used where
265        possible. This sometimes gives as much as 30% improvement.
266    
267    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
268        call.
269    
270    14. As documented, (*COMMIT) is now confined to within a positive assertion.
271    
272    15. It is now possible to link pcretest with libedit as an alternative to
273        libreadline.
274    
275    16. (*COMMIT) control verb is now supported by the JIT compiler.
276    
277    17. The Unicode data tables have been updated to Unicode 6.1.0.
278    
279    18. Added --file-list option to pcregrep.
280    
281    19. Added binary file support to pcregrep, including the -a, --binary-files,
282        -I, and --text options.
283    
284    20. The madvise function is renamed for posix_madvise for QNX compatibility
285        reasons. Fixed by Giuseppe D'Angelo.
286    
287    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
288    
289    22. Changed the option for creating symbolic links for 16-bit man pages from
290        -s to -sf so that re-installing does not cause issues.
291    
292    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
293    
294    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
295        matches in certain environments (the workspace was not being correctly
296        retained). Also added to pcre_dfa_exec() a simple plausibility check on
297        some of the workspace data at the beginning of a restart.
298    
299    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
300        was not doing so when it should - probably a typo introduced by SVN 528
301        (change 8.10/14).
302    
303    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
304        \w+ when the character tables indicated that \x{c4} was a word character.
305        There were several related cases, all because the tests for doing a table
306        lookup were testing for characters less than 127 instead of 255.
307    
308    27. If a pattern contains capturing parentheses that are not used in a match,
309        their slots in the ovector are set to -1. For those that are higher than
310        any matched groups, this happens at the end of processing. In the case when
311        there were back references that the ovector was too small to contain
312        (causing temporary malloc'd memory to be used during matching), and the
313        highest capturing number was not used, memory off the end of the ovector
314        was incorrectly being set to -1. (It was using the size of the temporary
315        memory instead of the true size.)
316    
317    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
318        ovector size, it uses memory at the end of the block that it has got.
319    
320    29. Check for an overlong MARK name and give an error at compile time. The
321        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
322    
323    30. JIT compiler update.
324    
325    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
326        Rill for the patch.
327    
328    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
329    
330    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
331    
332    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
333        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
334        was enabled.
335    
336    35. Improve JIT code generation for greedy plus quantifier.
337    
338    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
339        1 to "aa" instead of to an empty string. The bug affected repeated groups
340        that could potentially match an empty string.
341    
342    37. Optimizing single character iterators in JIT.
343    
344    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
345        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
346        codepoints that are too big for the mode are faulted, and in a UTF mode,
347        disallowed codepoints are also faulted.
348    
349    39. If PCRE was compiled with UTF support, in three places in the DFA
350        matcher there was code that should only have been obeyed in UTF mode, but
351        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
352        processing when bytes with values greater than 127 were present. In 16-bit
353        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
354        both cases the values are those that cannot be the first data item in a UTF
355        character. The three items that might have provoked this were recursions,
356        possessively repeated groups, and atomic groups.
357    
358    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
359        and pcregrep, because some OS require shared objects to be explicitly
360        passed to ld, causing the link step to fail if they are not.
361    
362    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
363        mode, patterns that started with \h* or \R* might be incorrectly matched.
364    
365    
366    Version 8.30 04-February-2012
367    -----------------------------
368    
369    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
370        name is defined in ctype.h.
371    
372    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
373        only in quite long subpatterns.
374    
375    3.  Removed the function pcre_info(), which has been obsolete and deprecated
376        since it was replaced by pcre_fullinfo() in February 2000.
377    
378    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
379        match a (*MARK), and the match failed at the start of the subject, a
380        reference to memory before the start of the subject could occur. This bug
381        was introduced by fix 17 of release 8.21.
382    
383    5.  A reference to an unset group with zero minimum repetition was giving
384        totally wrong answers (in non-JavaScript-compatibility mode). For example,
385        /(another)?(\1?)test/ matched against "hello world test". This bug was
386        introduced in release 8.13.
387    
388    6.  Add support for 16-bit character strings (a large amount of work involving
389        many changes and refactorings).
390    
391    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
392        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
393        from a file.
394    
395    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
396        rounding is not applied in this particular case).
397    
398    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
399        if they appear, or are escaped, in patterns.
400    
401    10. Get rid of a number of -Wunused-but-set-variable warnings.
402    
403    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
404        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
405        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
406        also returns the mark "x". This bug applied to capturing parentheses,
407        non-capturing parentheses, and atomic parentheses. It also applied to some
408        assertions.
409    
410    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
411        information out of configure.ac instead of relying on pcre.h.generic, which
412        is not stored in the repository.
413    
414    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
415        -lreadline.
416    
417    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
418    
419    15. Applied Graycode's patch to put the top-level frame on the stack rather
420        than the heap when not using the stack for recursion. This gives a
421        performance improvement in many cases when recursion is not deep.
422    
423    16. Experimental code added to "pcretest -C" to output the stack frame size.
424    
425    
426    Version 8.21 12-Dec-2011
427    ------------------------
428    
429    1.  Updating the JIT compiler.
430    
431    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
432        are added as well.
433    
434    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
435        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
436        calling _pcre_jit_exec. Some extra comments are added.
437    
438    4.  (*MARK) settings inside atomic groups that do not contain any capturing
439        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
440        was introduced by change 18 for 8.20.
441    
442    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
443        ECMA-262 standard.
444    
445    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
446        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
447        This bug was probably introduced by change 9 of 8.13.
448    
449    7.  While fixing 6 above, I noticed that a number of other items were being
450        incorrectly rejected as "not fixed length". This arose partly because newer
451        opcodes had not been added to the fixed-length checking code. I have (a)
452        corrected the bug and added tests for these items, and (b) arranged for an
453        error to occur if an unknown opcode is encountered while checking for fixed
454        length instead of just assuming "not fixed length". The items that were
455        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
456        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
457        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
458    
459    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
460        being incorrectly compiled and would have given unpredicatble results.
461    
462    9.  A possessively repeated subpattern with minimum repeat count greater than
463        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
464        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
465        the first (A) could occur when it should not.
466    
467    10. Add a cast and remove a redundant test from the code.
468    
469    11. JIT should use pcre_malloc/pcre_free for allocation.
470    
471    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
472        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
473        is anything other than /usr, -L is still shown).
474    
475    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
476    
477    14. Perl does not support \N without a following name in a [] class; PCRE now
478        also gives an error.
479    
480    15. If a forward reference was repeated with an upper limit of around 2000,
481        it caused the error "internal error: overran compiling workspace". The
482        maximum number of forward references (including repeats) was limited by the
483        internal workspace, and dependent on the LINK_SIZE. The code has been
484        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
485        the default depends on LINK_SIZE. There is a new upper limit (for safety)
486        of around 200,000 forward references. While doing this, I also speeded up
487        the filling in of repeated forward references.
488    
489    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
490        incorrectly expecting the subject to contain another "a" after the start.
491    
492    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
493        in the match, the SKIP should be ignored. This was not happening; instead
494        the SKIP was being treated as NOMATCH. For patterns such as
495        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
496        tested.
497    
498    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
499        now much more compatible with Perl, in particular in cases where the result
500        is a non-match for a non-anchored pattern. For example, if
501        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
502        "m", where previously it did not return a name. A side effect of this
503        change is that for partial matches, the last encountered mark name is
504        returned, as for non matches. A number of tests that were previously not
505        Perl-compatible have been moved into the Perl-compatible test files. The
506        refactoring has had the pleasing side effect of removing one argument from
507        the match() function, thus reducing its stack requirements.
508    
509    19. If the /S+ option was used in pcretest to study a pattern using JIT,
510        subsequent uses of /S (without +) incorrectly behaved like /S+.
511    
512    21. Retrieve executable code size support for the JIT compiler and fixing
513        some warnings.
514    
515    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
516        not work when the shorter character appeared right at the end of the
517        subject string.
518    
519    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
520        systems.
521    
522    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
523        output it when the /M option is used in pcretest.
524    
525    25. The CheckMan script was not being included in the distribution. Also, added
526        an explicit "perl" to run Perl scripts from the PrepareRelease script
527        because this is reportedly needed in Windows.
528    
529    26. If study data was being save in a file and studying had not found a set of
530        "starts with" bytes for the pattern, the data written to the file (though
531        never used) was taken from uninitialized memory and so caused valgrind to
532        complain.
533    
534    27. Updated RunTest.bat as provided by Sheri Pierce.
535    
536    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
537    
538    29. Computation of memory usage for the table of capturing group names was
539        giving an unnecessarily large value.
540    
541    
542    Version 8.20 21-Oct-2011
543    ------------------------
544    
545    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
546        a POSIX class. After further experiments with Perl, which convinced me that
547        Perl has bugs and confusions, a closing square bracket is no longer allowed
548        in a POSIX name. This bug also affected patterns with classes that started
549        with full stops.
550    
551    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
552        captured substring, but while checking the failing first alternative,
553        substring 1 is temporarily captured. If the output vector supplied to
554        pcre_exec() was not big enough for this capture, the yield of the function
555        was still zero ("insufficient space for captured substrings"). This cannot
556        be totally fixed without adding another stack variable, which seems a lot
557        of expense for a edge case. However, I have improved the situation in cases
558        such as /(a)(b)x|abc/ matched against "abc", where the return code
559        indicates that fewer than the maximum number of slots in the ovector have
560        been set.
561    
562    3.  Related to (2) above: when there are more back references in a pattern than
563        slots in the output vector, pcre_exec() uses temporary memory during
564        matching, and copies in the captures as far as possible afterwards. It was
565        using the entire output vector, but this conflicts with the specification
566        that only 2/3 is used for passing back captured substrings. Now it uses
567        only the first 2/3, for compatibility. This is, of course, another edge
568        case.
569    
570    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
571        main code base, and can be used by building with --enable-jit. When this is
572        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
573        runtime --no-jit option is given.
574    
575    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
576        ovector, the return from the function was zero, implying that there were
577        other matches that did not fit. The correct "exactly full" value is now
578        returned.
579    
580    6.  If a subpattern that was called recursively or as a subroutine contained
581        (*PRUNE) or any other control that caused it to give a non-standard return,
582        invalid errors such as "Error -26 (nested recursion at the same subject
583        position)" or even infinite loops could occur.
584    
585    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
586        computing the minimum length on reaching *ACCEPT, and so ended up with the
587        wrong value of 1 rather than 0. Further investigation indicates that
588        computing a minimum subject length in the presence of *ACCEPT is difficult
589        (think back references, subroutine calls), and so I have changed the code
590        so that no minimum is registered for a pattern that contains *ACCEPT.
591    
592    8.  If (*THEN) was present in the first (true) branch of a conditional group,
593        it was not handled as intended. [But see 16 below.]
594    
595    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
596        Sheri Pierce.
597    
598    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
599        the first byte in a match must be "a".
600    
601    11. Change 17 for 8.13 increased the recursion depth for patterns like
602        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
603        pattern contains any instances of (*THEN). If it does not, the old
604        optimizations are restored. It would be nice to do this on a per-group
605        basis, but at the moment that is not feasible.
606    
607    12. In some environments, the output of pcretest -C is CRLF terminated. This
608        broke RunTest's code that checks for the link size. A single white space
609        character after the value is now allowed for.
610    
611    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
612        For "fr", it uses the Windows-specific input and output files.
613    
614    14. If (*THEN) appeared in a group that was called recursively or as a
615        subroutine, it did not work as intended. [But see next item.]
616    
617    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
618        pattern fragments (but not containing any | characters). If A and B are
619        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
620        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
621        D. In other words, Perl considers parentheses that do not contain any |
622        characters to be part of a surrounding alternative, whereas PCRE was
623        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
624        differently. PCRE now behaves in the same way as Perl, except in the case
625        of subroutine/recursion calls such as (?1) which have in any case always
626        been different (but PCRE had them first :-).
627    
628    16. Related to 15 above: Perl does not treat the | in a conditional group as
629        creating alternatives. Such a group is treated in the same way as an
630        ordinary group without any | characters when processing (*THEN). PCRE has
631        been changed to match Perl's behaviour.
632    
633    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
634        RunGrepTest script failed.
635    
636    18. Change 22 for version 13 caused atomic groups to use more stack. This is
637        inevitable for groups that contain captures, but it can lead to a lot of
638        stack use in large patterns. The old behaviour has been restored for atomic
639        groups that do not contain any capturing parentheses.
640    
641    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
642        suppress the check for a minimum subject length at run time. (If it was
643        given to pcre_exec() or pcre_dfa_exec() it did work.)
644    
645    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
646        fail to work when decoding hex characters in data strings in EBCDIC
647        environments.
648    
649    21. It appears that in at least one Mac OS environment, the isxdigit() function
650        is implemented as a macro that evaluates to its argument more than once,
651        contravening the C 90 Standard (I haven't checked a later standard). There
652        was an instance in pcretest which caused it to go wrong when processing
653        \x{...} escapes in subject strings. The has been rewritten to avoid using
654        things like p++ in the argument of isxdigit().
655    
656    
657    Version 8.13 16-Aug-2011
658    ------------------------
659    
660    1.  The Unicode data tables have been updated to Unicode 6.0.0.
661    
662    2.  Two minor typos in pcre_internal.h have been fixed.
663    
664    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
665        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
666        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
667    
668    4.  There were a number of related bugs in the code for matching backrefences
669        caselessly in UTF-8 mode when codes for the characters concerned were
670        different numbers of bytes. For example, U+023A and U+2C65 are an upper
671        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
672        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
673        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
674        2-byte code at the end of the subject (it thought there wasn't enough data
675        left).
676    
677    5.  Comprehensive information about what went wrong is now returned by
678        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
679        as the output vector has at least 2 elements. The offset of the start of
680        the failing character and a reason code are placed in the vector.
681    
682    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
683        now returned is for the first byte of the failing character, instead of the
684        last byte inspected. This is an incompatible change, but I hope it is small
685        enough not to be a problem. It makes the returned offset consistent with
686        pcre_exec() and pcre_dfa_exec().
687    
688    7.  pcretest now gives a text phrase as well as the error number when
689        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
690        failure, the offset and reason code are output.
691    
692    8.  When \R was used with a maximizing quantifier it failed to skip backwards
693        over a \r\n pair if the subsequent match failed. Instead, it just skipped
694        back over a single character (\n). This seems wrong (because it treated the
695        two characters as a single entity when going forwards), conflicts with the
696        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
697        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
698        has been changed.
699    
700    9.  Some internal refactoring has changed the processing so that the handling
701        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
702        time (the PCRE_DOTALL option was changed this way some time ago: version
703        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
704        which was always a bit of a fudge. It also means that there is one less
705        argument for the match() function, which reduces its stack requirements
706        slightly. This change also fixes an incompatibility with Perl: the pattern
707        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
708    
709    10. More internal refactoring has drastically reduced the number of recursive
710        calls to match() for possessively repeated groups such as (abc)++ when
711        using pcre_exec().
712    
713    11. While implementing 10, a number of bugs in the handling of groups were
714        discovered and fixed:
715    
716        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
717        (a|)*(?1) gave a compile-time internal error.
718        ((a|)+)+  did not notice that the outer group could match an empty string.
719        (^a|^)+   was not marked as anchored.
720        (.*a|.*)+ was not marked as matching at start or after a newline.
721    
722    12. Yet more internal refactoring has removed another argument from the match()
723        function. Special calls to this function are now indicated by setting a
724        value in a variable in the "match data" data block.
725    
726    13. Be more explicit in pcre_study() instead of relying on "default" for
727        opcodes that mean there is no starting character; this means that when new
728        ones are added and accidentally left out of pcre_study(), testing should
729        pick them up.
730    
731    14. The -s option of pcretest has been documented for ages as being an old
732        synonym of -m (show memory usage). I have changed it to mean "force study
733        for every regex", that is, assume /S for every regex. This is similar to -i
734        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
735        using it. It makes it easier to run collections of tests with and without
736        study enabled, and thereby test pcre_study() more easily. All the standard
737        tests are now run with and without -s (but some patterns can be marked as
738        "never study" - see 20 below).
739    
740    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
741        restoration of the capturing data to the outer values was not happening
742        correctly.
743    
744    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
745        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
746        pattern had matched an empty string, and so incorrectly returned a no
747        match.
748    
749    17. There was optimizing code for the last branch of non-capturing parentheses,
750        and also for the obeyed branch of a conditional subexpression, which used
751        tail recursion to cut down on stack usage. Unfortunately, now that there is
752        the possibility of (*THEN) occurring in these branches, tail recursion is
753        no longer possible because the return has to be checked for (*THEN). These
754        two optimizations have therefore been removed. [But see 8.20/11 above.]
755    
756    18. If a pattern containing \R was studied, it was assumed that \R always
757        matched two bytes, thus causing the minimum subject length to be
758        incorrectly computed because \R can also match just one byte.
759    
760    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
761        was incorrectly computed.
762    
763    20. If /S is present twice on a test pattern in pcretest input, it now
764        *disables* studying, thereby overriding the use of -s on the command line
765        (see 14 above). This is necessary for one or two tests to keep the output
766        identical in both cases.
767    
768    21. When (*ACCEPT) was used in an assertion that matched an empty string and
769        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
770    
771    22. When an atomic group that contained a capturing parenthesis was
772        successfully matched, but the branch in which it appeared failed, the
773        capturing was not being forgotten if a higher numbered group was later
774        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
775        group 1 to "a", when in fact it should be unset. This applied to multi-
776        branched capturing and non-capturing groups, repeated or not, and also to
777        positive assertions (capturing in negative assertions does not happen
778        in PCRE) and also to nested atomic groups.
779    
780    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
781        subject after a captured substring, to make it easier to tell which of a
782        number of identical substrings has been captured.
783    
784    24. The way atomic groups are processed by pcre_exec() has been changed so that
785        if they are repeated, backtracking one repetition now resets captured
786        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
787        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
788        "aaa". Previously, it would have been "a". As part of this code
789        refactoring, the way recursive calls are handled has also been changed.
790    
791    25. If an assertion condition captured any substrings, they were not passed
792        back unless some other capturing happened later. For example, if
793        (?(?=(a))a) was matched against "a", no capturing was returned.
794    
795    26. When studying a pattern that contained subroutine calls or assertions,
796        the code for finding the minimum length of a possible match was handling
797        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
798        group 1 called group 2 while simultaneously a separate group 2 called group
799        1). A stack overflow occurred in this case. I have fixed this by limiting
800        the recursion depth to 10.
801    
802    27. Updated RunTest.bat in the distribution to the version supplied by Tom
803        Fortmann. This supports explicit test numbers on the command line, and has
804        argument validation and error reporting.
805    
806    28. An instance of \X with an unlimited repeat could fail if at any point the
807        first character it looked at was a mark character.
808    
809    29. Some minor code refactoring concerning Unicode properties and scripts
810        should reduce the stack requirement of match() slightly.
811    
812    30. Added the '=' option to pcretest to check the setting of unused capturing
813        slots at the end of the pattern, which are documented as being -1, but are
814        not included in the return count.
815    
816    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
817        compiled something random. Now it gives a compile-time error (as does
818        Perl).
819    
820    32. A *MARK encountered during the processing of a positive assertion is now
821        recorded and passed back (compatible with Perl).
822    
823    33. If --only-matching or --colour was set on a pcregrep call whose pattern
824        had alternative anchored branches, the search for a second match in a line
825        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
826        matched the line "0102" twice. The same bug affected patterns that started
827        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
828        twice.
829    
830    34. Previously, PCRE did not allow quantification of assertions. However, Perl
831        does, and because of capturing effects, quantifying parenthesized
832        assertions may at times be useful. Quantifiers are now allowed for
833        parenthesized assertions.
834    
835    35. A minor code tidy in pcre_compile() when checking options for \R usage.
836    
837    36. \g was being checked for fancy things in a character class, when it should
838        just be a literal "g".
839    
840    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
841        appearance of a nested POSIX class supersedes an apparent external class.
842        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
843        unescaped square brackets may also appear as part of class names. For
844        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
845        more like Perl. (But see 8.20/1 above.)
846    
847    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
848        was because it thought it was \N{name}, which is not supported).
849    
850    39. Add minix to OS list not supporting the -S option in pcretest.
851    
852    40. PCRE tries to detect cases of infinite recursion at compile time, but it
853        cannot analyze patterns in sufficient detail to catch mutual recursions
854        such as ((?1))((?2)). There is now a runtime test that gives an error if a
855        subgroup is called recursively as a subpattern for a second time at the
856        same position in the subject string. In previous releases this might have
857        been caught by the recursion limit, or it might have run out of stack.
858    
859    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
860        happen only once. PCRE was, however incorrectly giving a compile time error
861        "recursive call could loop indefinitely" because it cannot analyze the
862        pattern in sufficient detail. The compile time test no longer happens when
863        PCRE is compiling a conditional subpattern, but actual runaway loops are
864        now caught at runtime (see 40 above).
865    
866    42. It seems that Perl allows any characters other than a closing parenthesis
867        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
868        has been changed to be the same.
869    
870    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
871        as not to get warnings when autogen.sh is called. Also changed
872        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
873    
874    44. To help people who use pcregrep to scan files containing exceedingly long
875        lines, the following changes have been made:
876    
877        (a) The default value of the buffer size parameter has been increased from
878            8K to 20K. (The actual buffer used is three times this size.)
879    
880        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
881            PCRE is built.
882    
883        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
884            to be set at run time.
885    
886        (d) Numerical values in pcregrep options can be followed by K or M, for
887            example --buffer-size=50K.
888    
889        (e) If a line being scanned overflows pcregrep's buffer, an error is now
890            given and the return code is set to 2.
891    
892    45. Add a pointer to the latest mark to the callout data block.
893    
894    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
895        partial match of an empty string instead of no match. This was specific to
896        the use of ".".
897    
898    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
899        complete match instead of a partial match. This bug was dependent on both
900        the PCRE_UTF8 and PCRE_DOTALL options being set.
901    
902    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
903        starting byte set, because \b was not being ignored.
904    
905    
906    Version 8.12 15-Jan-2011
907    ------------------------
908    
909    1.  Fixed some typos in the markup of the man pages, and wrote a script that
910        checks for such things as part of the documentation building process.
911    
912    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
913        --match-limit and --recursion-limit options (added for 8.11). In
914        particular, this made one of the standard tests fail. (The integer value
915        went into the wrong half of a long int.)
916    
917    3.  If the --colour option was given to pcregrep with -v (invert match), it
918        did strange things, either producing crazy output, or crashing. It should,
919        of course, ignore a request for colour when reporting lines that do not
920        match.
921    
922    4.  Another pcregrep bug caused similar problems if --colour was specified with
923        -M (multiline) and the pattern match finished with a line ending.
924    
925    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
926        matched in multiline mode, the following line was shown as part of the
927        match. This seems wrong, so I have changed it.
928    
929    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
930        the check for further matches in the same line (so they could be coloured)
931        to overrun the end of the current line. If another match was found, it was
932        incorrectly shown (and then shown again when found in the next line).
933    
934    7.  If pcregrep was compiled under Windows, there was a reference to the
935        function pcregrep_exit() before it was defined. I am assuming this was
936        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
937        reported by a user. I've moved the definition above the reference.
938    
939    
940    Version 8.11 10-Dec-2010
941    ------------------------
942    
943    1.  (*THEN) was not working properly if there were untried alternatives prior
944        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
945        backtracked to try for "b" instead of moving to the next alternative branch
946        at the same level (in this case, to look for "c"). The Perl documentation
947        is clear that when (*THEN) is backtracked onto, it goes to the "next
948        alternative in the innermost enclosing group".
949    
950    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
951        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
952        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
953        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
954        (*THEN).
955    
956    3.  If \s appeared in a character class, it removed the VT character from
957        the class, even if it had been included by some previous item, for example
958        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
959        of \s, but is part of the POSIX "space" class.)
960    
961    4.  A partial match never returns an empty string (because you can always
962        match an empty string at the end of the subject); however the checking for
963        an empty string was starting at the "start of match" point. This has been
964        changed to the "earliest inspected character" point, because the returned
965        data for a partial match starts at this character. This means that, for
966        example, /(?<=abc)def/ gives a partial match for the subject "abc"
967        (previously it gave "no match").
968    
969    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
970        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
971        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
972        has an implication that the given string is incomplete (because a partial
973        match is preferred over a full match). For this reason, these items now
974        give a partial match in this situation. [Aside: previously, the one case
975        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
976        match rather than a full match, which was wrong by the old rules, but is
977        now correct.]
978    
979    6.  There was a bug in the handling of #-introduced comments, recognized when
980        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
981        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
982        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
983        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
984        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
985        places in pcre_compile().
986    
987    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
988        comments when looking ahead for named forward references to subpatterns,
989        the only newline sequence it recognized was NL. It now handles newlines
990        according to the set newline convention.
991    
992    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
993        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
994        cater for a lack of strerror(). These oversights have been fixed.
995    
996    9.  Added --match-limit and --recursion-limit to pcregrep.
997    
998    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
999    
1000    11. When the -o option was used, pcregrep was setting a return code of 1, even
1001        when matches were found, and --line-buffered was not being honoured.
1002    
1003    12. Added an optional parentheses number to the -o and --only-matching options
1004        of pcregrep.
1005    
1006    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1007        can match an empty string. The code to do it in pcretest and pcredemo
1008        needed fixing:
1009    
1010        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1011            only one byte after an empty string match just before CRLF (this case
1012            just got forgotten; "any" and "anycrlf" were OK).
1013    
1014        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1015            mode when an empty string match preceded an ASCII character followed by
1016            a non-ASCII character. (The code for advancing by one character rather
1017            than one byte was nonsense.)
1018    
1019        (c) The pcredemo.c sample program did not have any code at all to handle
1020            the cases when CRLF is a valid newline sequence.
1021    
1022    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1023        as a starting offset was within the subject string. There is now a new
1024        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1025        negative or greater than the length of the string. In order to test this,
1026        pcretest is extended to allow the setting of negative starting offsets.
1027    
1028    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1029        starting offset points to the beginning of a UTF-8 character was
1030        unnecessarily clumsy. I tidied it up.
1031    
1032    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1033        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1034    
1035    17. Nobody had reported that the --include_dir option, which was added in
1036        release 7.7 should have been called --include-dir (hyphen, not underscore)
1037        for compatibility with GNU grep. I have changed it to --include-dir, but
1038        left --include_dir as an undocumented synonym, and the same for
1039        --exclude-dir, though that is not available in GNU grep, at least as of
1040        release 2.5.4.
1041    
1042    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1043        characters from a string of bytes) have been redefined so as not to use
1044        loops, in order to improve performance in some environments. At the same
1045        time, I abstracted some of the common code into auxiliary macros to save
1046        repetition (this should not affect the compiled code).
1047    
1048    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1049        compile-time error is now given if \c is not followed by an ASCII
1050        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1051        different, and any byte value is allowed.)
1052    
1053    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1054        START_OPTIMIZE option, which is now allowed at compile time - but just
1055        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1056        to pcregrep and other applications that have no direct access to PCRE
1057        options. The new /Y option in pcretest sets this option when calling
1058        pcre_compile().
1059    
1060    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1061        back references. Groups containing recursive back references were forced to
1062        be atomic by that change, but in the case of named groups, the amount of
1063        memory required was incorrectly computed, leading to "Failed: internal
1064        error: code overflow". This has been fixed.
1065    
1066    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1067        pcretest.c, to avoid build problems in some Borland environments.
1068    
1069    
1070    Version 8.10 25-Jun-2010
1071    ------------------------
1072    
1073    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1074        THEN.
1075    
1076    2.  (*ACCEPT) was not working when inside an atomic group.
1077    
1078    3.  Inside a character class, \B is treated as a literal by default, but
1079        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1080        causes the error). The code is unchanged, but I tidied the documentation.
1081    
1082    4.  Inside a character class, PCRE always treated \R and \X as literals,
1083        whereas Perl faults them if its -w option is set. I have changed PCRE so
1084        that it faults them when PCRE_EXTRA is set.
1085    
1086    5.  Added support for \N, which always matches any character other than
1087        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1088    
1089    6.  When compiling pcregrep with newer versions of gcc which may have
1090        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1091        declared with attribute warn_unused_result" were given. Just casting the
1092        result to (void) does not stop the warnings; a more elaborate fudge is
1093        needed. I've used a macro to implement this.
1094    
1095    7.  Minor change to pcretest.c to avoid a compiler warning.
1096    
1097    8.  Added four artifical Unicode properties to help with an option to make
1098        \s etc use properties (see next item). The new properties are: Xan
1099        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1100    
1101    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1102        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1103        this option. Modified pcretest to add /W to test this facility. Added
1104        REG_UCP to make it available via the POSIX interface.
1105    
1106    10. Added --line-buffered to pcregrep.
1107    
1108    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1109        studied, and the match started with a letter with a code point greater than
1110        127 whose first byte was different to the first byte of the other case of
1111        the letter, the other case of this starting letter was not recognized
1112        (#976).
1113    
1114    12. If a pattern that was studied started with a repeated Unicode property
1115        test, for example, \p{Nd}+, there was the theoretical possibility of
1116        setting up an incorrect bitmap of starting bytes, but fortunately it could
1117        not have actually happened in practice until change 8 above was made (it
1118        added property types that matched character-matching opcodes).
1119    
1120    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1121        possible starting bytes for non-anchored patterns.
1122    
1123    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1124        \R, and also a number of cases that involve Unicode properties, both
1125        explicit and implicit when PCRE_UCP is set.
1126    
1127    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1128        input, it could crash or give wrong results if characters with values
1129        greater than 0xc0 were present in the subject string. (Detail: it assumed
1130        UTF-8 input when processing these items.)
1131    
1132    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1133        size_t is 64-bit (#991).
1134    
1135    17. Added a check for running out of memory when PCRE is compiled with
1136        --disable-stack-for-recursion (#990).
1137    
1138    18. If the last data line in a file for pcretest does not have a newline on
1139        the end, a newline was missing in the output.
1140    
1141    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1142        less than 128) in its various bitmaps. However, there is a facility for
1143        generating tables according to the current locale when PCRE is compiled. It
1144        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1145        characters, are recognized by isspace() and therefore were getting set in
1146        these tables, and indeed these tables seem to approximate to ISO 8859. This
1147        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1148        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1149        which of course cannot start UTF-8 characters. I have changed the code so
1150        that only real ASCII characters (less than 128) and the correct starting
1151        bytes for UTF-8 encodings are set for characters greater than 127 when in
1152        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1153        altogether.)
1154    
1155    20. Added the /T option to pcretest so as to be able to run tests with non-
1156        standard character tables, thus making it possible to include the tests
1157        used for 19 above in the standard set of tests.
1158    
1159    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1160        reference to a subpattern the other side of a comment that contains an
1161        opening parenthesis caused either an internal compiling error, or a
1162        reference to the wrong subpattern.
1163    
1164    
1165    Version 8.02 19-Mar-2010
1166    ------------------------
1167    
1168    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1169    
1170    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1171        configured.
1172    
1173    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1174        original author of that file, following a query about its status.
1175    
1176    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1177        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1178    
1179    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1180        quantifier applied to a forward-referencing subroutine call, could compile
1181        incorrect code or give the error "internal error: previously-checked
1182        referenced subpattern not found".
1183    
1184    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1185        variables to point to external functions. For these systems, therefore,
1186        pcre_malloc etc. are now initialized to local functions that call the
1187        relevant global functions.
1188    
1189    7.  There were two entries missing in the vectors called coptable and poptable
1190        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1191        I've fixed the data, and added a kludgy way of testing at compile time that
1192        the lengths are correct (equal to the number of opcodes).
1193    
1194    8.  Following on from 7, I added a similar kludge to check the length of the
1195        eint vector in pcreposix.c.
1196    
1197    9.  Error texts for pcre_compile() are held as one long string to avoid too
1198        much relocation at load time. To find a text, the string is searched,
1199        counting zeros. There was no check for running off the end of the string,
1200        which could happen if a new error number was added without updating the
1201        string.
1202    
1203    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1204    
1205    11. \K was not working if it appeared in an atomic group or in a group that
1206        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1207        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1208        the assertion is positive, but not if it is negative.
1209    
1210    12. Change 11 fortuitously reduced the size of the stack frame used in the
1211        "match()" function of pcre_exec.c by one pointer. Forthcoming
1212        implementation of support for (*MARK) will need an extra pointer on the
1213        stack; I have reserved it now, so that the stack frame size does not
1214        decrease.
1215    
1216    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1217        item in branch that calls a recursion is a subroutine call - as in the
1218        second branch in the above example - was incorrectly given the compile-
1219        time error "recursive call could loop indefinitely" because pcre_compile()
1220        was not correctly checking the subroutine for matching a non-empty string.
1221    
1222    14. The checks for overrunning compiling workspace could trigger after an
1223        overrun had occurred. This is a "should never occur" error, but it can be
1224        triggered by pathological patterns such as hundreds of nested parentheses.
1225        The checks now trigger 100 bytes before the end of the workspace.
1226    
1227    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1228    
1229    
1230    Version 8.01 19-Jan-2010
1231    ------------------------
1232    
1233    1.  If a pattern contained a conditional subpattern with only one branch (in
1234        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1235        computed the wrong minimum data length (which is of course zero for such
1236        subpatterns). This could cause incorrect "no match" results.
1237    
1238    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1239        the pattern is reset in the first branch, pcre_compile() failed with
1240        "internal error: code overflow at offset...". This happened only when
1241        the reset was to the original external option setting. (An optimization
1242        abstracts leading options settings into an external setting, which was the
1243        cause of this.)
1244    
1245    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1246        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1247        assertion pattern did not match (meaning that the assertion was true), it
1248        was incorrectly treated as false if the SKIP had been reached during the
1249        matching. This also applied to assertions used as conditions.
1250    
1251    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1252        assertion subpattern, including such a pattern used as a condition,
1253        unpredictable results occurred, instead of the error return
1254        PCRE_ERROR_DFA_UITEM.
1255    
1256    5.  The C++ GlobalReplace function was not working like Perl for the special
1257        situation when an empty string is matched. It now does the fancy magic
1258        stuff that is necessary.
1259    
1260    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1261        removed. (These were left over from very, very early versions of PCRE.)
1262    
1263    7.  Some cosmetic changes to the code to make life easier when compiling it
1264        as part of something else:
1265    
1266        (a) Change DEBUG to PCRE_DEBUG.
1267    
1268        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1269            called "current" as "current_branch", to prevent a collision with the
1270            Linux macro when compiled as a kernel module.
1271    
1272        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1273            prevent a collision with the Linux macro when compiled as a kernel
1274            module.
1275    
1276    8.  In pcre_compile() there are some checks for integer overflows that used to
1277        cast potentially large values to (double). This has been changed to that
1278        when building, a check for int64_t is made, and if it is found, it is used
1279        instead, thus avoiding the use of floating point arithmetic. (There is no
1280        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1281        double.
1282    
1283    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1284        2005 (difference between two addresses compared to an unsigned value).
1285    
1286    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1287        custom one, because of the following reported problem in Windows:
1288    
1289          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1290              under Win32.
1291          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1292              therefore missing the function definition.
1293          - The compiler thus generates a "C" signature for the test function.
1294          - The linker fails to find the "C" function.
1295          - PCRE fails to configure if asked to do so against libbz2.
1296    
1297    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1298        messages were output:
1299    
1300          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1301          rerunning libtoolize, to keep the correct libtool macros in-tree.
1302          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1303    
1304        I have done both of these things.
1305    
1306    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1307        most of the time, it *can* run out if it is given a pattern that contains a
1308        runaway infinite recursion. I updated the discussion in the pcrestack man
1309        page.
1310    
1311    13. Now that we have gone to the x.xx style of version numbers, the minor
1312        version may start with zero. Using 08 or 09 is a bad idea because users
1313        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1314        interpreted as invalid octal numbers. I've updated the previous comment in
1315        configure.ac, and also added a check that gives an error if 08 or 09 are
1316        used.
1317    
1318    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1319        causing partial matching to fail when the end of the subject matched \W
1320        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1321    
1322    15. There were some discrepancies between the declarations in pcre_internal.h
1323        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1324        their definitions. The declarations used "const uschar *" and the
1325        definitions used USPTR. Even though USPTR is normally defined as "const
1326        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1327        reported that: "This difference in casting confuses some C++ compilers, for
1328        example, SunCC recognizes above declarations as different functions and
1329        generates broken code for hbpcre." I have changed the declarations to use
1330        USPTR.
1331    
1332    16. GNU libtool is named differently on some systems. The autogen.sh script now
1333        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1334        (FreeBSD).
1335    
1336    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1337        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1338        comment: "Figure out how to create a longlong from a string: strtoll and
1339        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1340        instance, but it only takes 2 args instead of 3!"
1341    
1342    18. A subtle bug concerned with back references has been fixed by a change of
1343        specification, with a corresponding code fix. A pattern such as
1344        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1345        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1346        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1347        same bug. Such groups have to be quantified to be useful, or contained
1348        inside another quantified group. (If there's no repetition, the reference
1349        can never match.) The problem arises because, having left the group and
1350        moved on to the rest of the pattern, a later failure that backtracks into
1351        the group uses the captured value from the final iteration of the group
1352        rather than the correct earlier one. I have fixed this in PCRE by forcing
1353        any group that contains a reference to itself to be an atomic group; that
1354        is, there cannot be any backtracking into it once it has completed. This is
1355        similar to recursive and subroutine calls.
1356    
1357    
1358    Version 8.00 19-Oct-09
1359    ----------------------
1360    
1361    1.  The table for translating pcre_compile() error codes into POSIX error codes
1362        was out-of-date, and there was no check on the pcre_compile() error code
1363        being within the table. This could lead to an OK return being given in
1364        error.
1365    
1366    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1367        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1368        in a Windows environment.
1369    
1370    3.  The pcregrep --count option prints the count for each file even when it is
1371        zero, as does GNU grep. However, pcregrep was also printing all files when
1372        --files-with-matches was added. Now, when both options are given, it prints
1373        counts only for those files that have at least one match. (GNU grep just
1374        prints the file name in this circumstance, but including the count seems
1375        more useful - otherwise, why use --count?) Also ensured that the
1376        combination -clh just lists non-zero counts, with no names.
1377    
1378    4.  The long form of the pcregrep -F option was incorrectly implemented as
1379        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1380        but it seems right to fix it, and I didn't think it was worth preserving
1381        the old behaviour.
1382    
1383    5.  The command line items --regex=pattern and --regexp=pattern were not
1384        recognized by pcregrep, which required --regex pattern or --regexp pattern
1385        (with a space rather than an '='). The man page documented the '=' forms,
1386        which are compatible with GNU grep; these now work.
1387    
1388    6.  No libpcreposix.pc file was created for pkg-config; there was just
1389        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1390    
1391    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1392        when UCP support is not needed, by modifying the Python script that
1393        generates it from Unicode data files. This should not matter if the module
1394        is correctly used as a library, but I received one complaint about 50K of
1395        unwanted data. My guess is that the person linked everything into his
1396        program rather than using a library. Anyway, it does no harm.
1397    
1398    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1399        was a minimum greater than 1 for a wide character in a possessive
1400        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1401        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1402        character. Chaos in the form of incorrect output or a compiling loop could
1403        result.
1404    
1405    9.  The restrictions on what a pattern can contain when partial matching is
1406        requested for pcre_exec() have been removed. All patterns can now be
1407        partially matched by this function. In addition, if there are at least two
1408        slots in the offset vector, the offset of the earliest inspected character
1409        for the match and the offset of the end of the subject are set in them when
1410        PCRE_ERROR_PARTIAL is returned.
1411    
1412    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1413        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1414        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1415        and may be more useful for multi-segment matching.
1416    
1417    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1418        used to be given if ever the end of the subject was reached; now it is
1419        given only if matching could not proceed because another character was
1420        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1421        string "Z", which now yields "no match" instead of "partial match". In the
1422        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1423        final character ended with (*FAIL).
1424    
1425    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1426        if the pattern had a "must contain" character that was already found in the
1427        earlier partial match, unless partial matching was again requested. For
1428        example, with the pattern /dog.(body)?/, the "must contain" character is
1429        "g". If the first part-match was for the string "dog", restarting with
1430        "sbody" failed. This bug has been fixed.
1431    
1432    13. The string returned by pcre_dfa_exec() after a partial match has been
1433        changed so that it starts at the first inspected character rather than the
1434        first character of the match. This makes a difference only if the pattern
1435        starts with a lookbehind assertion or \b or \B (\K is not supported by
1436        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1437        matching functions compatible, and I think it's the right thing to do.
1438    
1439    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1440        so that the demonstration program is easily available in environments where
1441        PCRE has not been installed from source.
1442    
1443    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1444        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1445        library.
1446    
1447    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1448        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1449        is not the first non-POSIX option to be added. Clearly some people find
1450        these options useful.
1451    
1452    17. If a caller to the POSIX matching function regexec() passes a non-zero
1453        value for nmatch with a NULL value for pmatch, the value of
1454        nmatch is forced to zero.
1455    
1456    18. RunGrepTest did not have a test for the availability of the -u option of
1457        the diff command, as RunTest does. It now checks in the same way as
1458        RunTest, and also checks for the -b option.
1459    
1460    19. If an odd number of negated classes containing just a single character
1461        interposed, within parentheses, between a forward reference to a named
1462        subpattern and the definition of the subpattern, compilation crashed with
1463        an internal error, complaining that it could not find the referenced
1464        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1465        [The bug was that it was starting one character too far in when skipping
1466        over the character class, thus treating the ] as data rather than
1467        terminating the class. This meant it could skip too much.]
1468    
1469    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1470        /g option in pcretest when the pattern contains \K, which makes it possible
1471        to have an empty string match not at the start, even when the pattern is
1472        anchored. Updated pcretest and pcredemo to use this option.
1473    
1474    21. If the maximum number of capturing subpatterns in a recursion was greater
1475        than the maximum at the outer level, the higher number was returned, but
1476        with unset values at the outer level. The correct (outer level) value is
1477        now given.
1478    
1479    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1480        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1481        make it do so. The string so far is captured, making this feature
1482        compatible with Perl.
1483    
1484    23. The tests have been re-organized, adding tests 11 and 12, to make it
1485        possible to check the Perl 5.10 features against Perl 5.10.
1486    
1487    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1488        pattern matches a fixed length string. PCRE did not allow this; now it
1489        does. Neither allows recursion.
1490    
1491    25. I finally figured out how to implement a request to provide the minimum
1492        length of subject string that was needed in order to match a given pattern.
1493        (It was back references and recursion that I had previously got hung up
1494        on.) This code has now been added to pcre_study(); it finds a lower bound
1495        to the length of subject needed. It is not necessarily the greatest lower
1496        bound, but using it to avoid searching strings that are too short does give
1497        some useful speed-ups. The value is available to calling programs via
1498        pcre_fullinfo().
1499    
1500    26. While implementing 25, I discovered to my embarrassment that pcretest had
1501        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1502        study optimizations had never been tested with that matching function.
1503        Oops. What is worse, even when it was passed study data, there was a bug in
1504        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1505        were also very few tests of studied patterns with pcre_dfa_exec().
1506    
1507    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1508        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1509        on the other side of the coin, they are no longer allowed to have different
1510        names, because these cannot be distinguished in PCRE, and this has caused
1511        confusion. (This is a difference from Perl.)
1512    
1513    28. When duplicate subpattern names are present (necessarily with different
1514        numbers, as required by 27 above), and a test is made by name in a
1515        conditional pattern, either for a subpattern having been matched, or for
1516        recursion in such a pattern, all the associated numbered subpatterns are
1517        tested, and the overall condition is true if the condition is true for any
1518        one of them. This is the way Perl works, and is also more like the way
1519        testing by number works.
1520    
1521    
1522    Version 7.9 11-Apr-09
1523    ---------------------
1524    
1525    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1526        (pcretest), all targets were linked against these libraries. This included
1527        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1528        libraries. This caused unwanted dependencies to be created. This problem
1529        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1530        pcretest is linked with readline.
1531    
1532    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1533        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1534        moved outside it again, because FALSE and TRUE are already defined in AIX,
1535        but BOOL is not.
1536    
1537    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1538        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1539    
1540    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1541        hyphens) following file names and line numbers when outputting matching
1542        lines. This is not true; no spaces are inserted. I have also clarified the
1543        wording for the --colour (or --color) option.
1544    
1545    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1546        was not coloured; this is different to GNU grep, so I have changed it to be
1547        the same.
1548    
1549    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1550        each matching line was coloured. Now it goes on to look for further matches
1551        of any of the test patterns, which is the same behaviour as GNU grep.
1552    
1553    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1554        doesn't make sense to accept an empty string match in pcregrep, so I have
1555        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1556        seems to be how GNU grep behaves.
1557    
1558    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1559        start or after a newline", because the conditional assertion was not being
1560        correctly handled. The rule now is that both the assertion and what follows
1561        in the first alternative must satisfy the test.
1562    
1563    9.  If auto-callout was enabled in a pattern with a conditional group whose
1564        condition was an assertion, PCRE could crash during matching, both with
1565        pcre_exec() and pcre_dfa_exec().
1566    
1567    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1568        used for matching.
1569    
1570    11. Unicode property support in character classes was not working for
1571        characters (bytes) greater than 127 when not in UTF-8 mode.
1572    
1573    12. Added the -M command line option to pcretest.
1574    
1575    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1576    
1577    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1578    
1579    16. Added comments and documentation about mis-use of no_arg in the C++
1580        wrapper.
1581    
1582    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1583        from Martin Jerabek that uses macro names for all relevant character and
1584        string constants.
1585    
1586    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1587        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1588        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1589        these, but not everybody uses configure.
1590    
1591    19. A conditional group that had only one branch was not being correctly
1592        recognized as an item that could match an empty string. This meant that an
1593        enclosing group might also not be so recognized, causing infinite looping
1594        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1595        with the subject "ab", where knowledge that the repeated group can match
1596        nothing is needed in order to break the loop.
1597    
1598    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1599        exec(), but without supplying a callout function, matching went wrong.
1600    
1601    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1602        leak if the size of the offset vector was greater than 30. When the vector
1603        is smaller, the saved offsets during recursion go onto a local stack
1604        vector, but for larger vectors malloc() is used. It was failing to free
1605        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1606        error, in fact).
1607    
1608    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1609        heapframe that is used only when UTF-8 support is enabled. This caused no
1610        problem, but was untidy.
1611    
1612    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1613        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1614        included within another project.
1615    
1616    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1617        slightly modified by me:
1618    
1619          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1620              not building pcregrep.
1621    
1622          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1623              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1624    
1625    25. Forward references, both numeric and by name, in patterns that made use of
1626        duplicate group numbers, could behave incorrectly or give incorrect errors,
1627        because when scanning forward to find the reference group, PCRE was not
1628        taking into account the duplicate group numbers. A pattern such as
1629        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1630    
1631    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1632        the feature of a custom pointer more persuasive (as requested by a user).
1633    
1634    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1635        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1636        pre-defined.
1637    
1638    28. Added support for (*UTF8) at the start of a pattern.
1639    
1640    29. Arrange for flags added by the "release type" setting in CMake to be shown
1641        in the configuration summary.
1642    
1643    
1644    Version 7.8 05-Sep-08
1645    ---------------------
1646    
1647    1.  Replaced UCP searching code with optimized version as implemented for Ad
1648        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1649        stage table and inline lookup instead of a function, giving speed ups of 2
1650        to 5 times on some simple patterns that I tested. Permission was given to
1651        distribute the MultiStage2.py script that generates the tables (it's not in
1652        the tarball, but is in the Subversion repository).
1653    
1654    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1655        scripts.
1656    
1657    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1658        a group with a zero qualifier. The result of the study could be incorrect,
1659        or the function might crash, depending on the pattern.
1660    
1661    4.  Caseless matching was not working for non-ASCII characters in back
1662        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1663        It now works when Unicode Property Support is available.
1664    
1665    5.  In pcretest, an escape such as \x{de} in the data was always generating
1666        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1667        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1668        truncation.
1669    
1670    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1671    
1672    7.  Added two (int) casts to pcregrep when printing the difference of two
1673        pointers, in case they are 64-bit values.
1674    
1675    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1676        test 2 if it fails.
1677    
1678    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1679        and a #define of that name to empty if it is not externally set. This is to
1680        allow users of MSVC to set it if necessary.
1681    
1682    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1683        the convenience functions in the pcre_get.c source file.
1684    
1685    11. An option change at the start of a pattern that had top-level alternatives
1686        could cause overwriting and/or a crash. This command provoked a crash in
1687        some environments:
1688    
1689          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1690    
1691        This potential security problem was recorded as CVE-2008-2371.
1692    
1693    12. For a pattern where the match had to start at the beginning or immediately
1694        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1695        pcre_dfa_exec() could read past the end of the passed subject if there was
1696        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1697        pcretest so that it places the subject at the end of its malloc-ed buffer.
1698    
1699    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1700        exec() might read past the end of the data buffer in UTF-8 mode.
1701    
1702    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1703        the data contained the byte 0x85 as part of a UTF-8 character within its
1704        first line. This applied both to normal and DFA matching.
1705    
1706    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1707        /^[^d]*?$/8 failed to match "abc".
1708    
1709    16. Added a missing copyright notice to pcrecpp_internal.h.
1710    
1711    17. Make it more clear in the documentation that values returned from
1712        pcre_exec() in ovector are byte offsets, not character counts.
1713    
1714    18. Tidied a few places to stop certain compilers from issuing warnings.
1715    
1716    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1717        supplied by Stefan Weber. I made a further small update for 7.8 because
1718        there is a change of source arrangements: the pcre_searchfuncs.c module is
1719        replaced by pcre_ucd.c.
1720    
1721    
1722    Version 7.7 07-May-08
1723    ---------------------
1724    
1725    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1726        a string to a long long, pretend we don't even have a long long." This is
1727        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1728    
1729    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1730        pre-7.6 versions, which defined a global no_arg variable instead of putting
1731        it in the RE class. (See also #8 below.)
1732    
1733    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1734        Lopes.
1735    
1736    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1737    
1738        (1) The include/exclude patterns were being applied to the whole pathnames
1739            of files, instead of just to the final components.
1740    
1741        (2) If there was more than one level of directory, the subdirectories were
1742            skipped unless they satisfied the include/exclude conditions. This is
1743            inconsistent with GNU grep (and could even be seen as contrary to the
1744            pcregrep specification - which I improved to make it absolutely clear).
1745            The action now is always to scan all levels of directory, and just
1746            apply the include/exclude patterns to regular files.
1747    
1748    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1749        --exclude_dir in the tests to avoid scanning .svn directories.
1750    
1751    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1752        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1753        doesn't support NULs in patterns.
1754    
1755    7.  Added some missing "const"s to declarations of static tables in
1756        pcre_compile.c and pcre_dfa_exec.c.
1757    
1758    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1759        caused by fix #2  above. (Subsequently also a second patch to fix the
1760        first patch. And a third patch - this was a messy problem.)
1761    
1762    9.  Applied Craig's patch to remove the use of push_back().
1763    
1764    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1765        matching function regexec().
1766    
1767    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1768        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1769        references. PCRE supports relative numbers with this syntax (I don't think
1770        Oniguruma does).
1771    
1772    12. Previously, a group with a zero repeat such as (...){0} was completely
1773        omitted from the compiled regex. However, this means that if the group
1774        was called as a subroutine from elsewhere in the pattern, things went wrong
1775        (an internal error was given). Such groups are now left in the compiled
1776        pattern, with a new opcode that causes them to be skipped at execution
1777        time.
1778    
1779    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1780        to the way PCRE behaves:
1781    
1782        (a) A lone ] character is dis-allowed (Perl treats it as data).
1783    
1784        (b) A back reference to an unmatched subpattern matches an empty string
1785            (Perl fails the current match path).
1786    
1787        (c) A data ] in a character class must be notated as \] because if the
1788            first data character in a class is ], it defines an empty class. (In
1789            Perl it is not possible to have an empty class.) The empty class []
1790            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1791            The negative empty class [^] matches any one character, independently
1792            of the DOTALL setting.
1793    
1794    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1795        non-existent subpattern following a character class starting with ']' and
1796        containing () gave an internal compiling error instead of "reference to
1797        non-existent subpattern". Fortunately, when the pattern did exist, the
1798        compiled code was correct. (When scanning forwards to check for the
1799        existencd of the subpattern, it was treating the data ']' as terminating
1800        the class, so got the count wrong. When actually compiling, the reference
1801        was subsequently set up correctly.)
1802    
1803    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1804        it was being rejected as not supported by pcre_dfa_exec(), even though
1805        other assertions are supported. I have made pcre_dfa_exec() support
1806        (*FAIL).
1807    
1808    16. The implementation of 13c above involved the invention of a new opcode,
1809        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1810        cannot be changed at match time, I realized I could make a small
1811        improvement to matching performance by compiling OP_ALLANY instead of
1812        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1813        on the OP_ANY path.
1814    
1815    17. Compiling pcretest on Windows with readline support failed without the
1816        following two fixes: (1) Make the unistd.h include conditional on
1817        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1818    
1819    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1820        ncurses library to be included for pcretest when ReadLine support is
1821        requested, but also to allow for it to be overridden. This patch came from
1822        Daniel Bergström.
1823    
1824    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1825        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1826        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1827        spotting this.
1828    
1829    
1830    Version 7.6 28-Jan-08
1831    ---------------------
1832    
1833    1.  A character class containing a very large number of characters with
1834        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1835        overflow.
1836    
1837    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1838        HAVE_LONG_LONG is not defined.
1839    
1840    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1841        bring them up to date and include new features. This patch includes:
1842    
1843        - Fixed PH's badly added libz and libbz2 support.
1844        - Fixed a problem with static linking.
1845        - Added pcredemo. [But later removed - see 7 below.]
1846        - Fixed dftables problem and added an option.
1847        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1848            HAVE_LONG_LONG.
1849        - Added readline support for pcretest.
1850        - Added an listing of the option settings after cmake has run.
1851    
1852    4.  A user submitted a patch to Makefile that makes it easy to create
1853        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1854        Makefile.am that cause it to include this special target, without
1855        affecting anything else. Note that the same mingw target plus all
1856        the other distribution libraries and programs are now supported
1857        when configuring with CMake (see 6 below) instead of with
1858        Configure/Make.
1859    
1860    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1861        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1862        exported in the Windows port". It has not yet been confirmed that the patch
1863        solves the problem, but it does no harm.
1864    
1865    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1866        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1867        with CMake, and also correct the comment about stack recursion.
1868    
1869    7.  Remove the automatic building of pcredemo from the ./configure system and
1870        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1871        of a program that users should build themselves after PCRE is installed, so
1872        building it automatically is not really right. What is more, it gave
1873        trouble in some build environments.
1874    
1875    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1876    
1877    
1878    Version 7.5 10-Jan-08
1879    ---------------------
1880    
1881    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1882        values in parens when parsing an RE using the C++ wrapper."
1883    
1884    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1885        Characters greater than 255 were excluded from the class instead of being
1886        included.
1887    
1888    3.  The same bug as (2) above applied to negated POSIX classes such as
1889        [:^space:].
1890    
1891    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1892        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1893        I have changed it.
1894    
1895    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1896        first named subpattern) and a construct such as (?&a) would reference the
1897        first named subpattern whose name started with "a" (in other words, the
1898        length check was missing). Both these problems are fixed. "Subpattern name
1899        expected" is now given for (?&) (a zero-length name), and this patch also
1900        makes it give the same error for \k'' (previously it complained that that
1901        was a reference to a non-existent subpattern).
1902    
1903    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1904        this is right because (?- can be followed by option settings as well as by
1905        digits. I have, however, made the messages clearer.
1906    
1907    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1908        than the number used in the conditional) now cause a compile-time error.
1909        This is actually not compatible with Perl, which accepts such patterns, but
1910        treats the conditional as always being FALSE (as PCRE used to), but it
1911        seems to me that giving a diagnostic is better.
1912    
1913    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1914        and messages.
1915    
1916    9.  Fix two occurrences of "backslash" in comments that should have been
1917        "backspace".
1918    
1919    10. Remove two redundant lines of code that can never be obeyed (their function
1920        was moved elsewhere).
1921    
1922    11. The program that makes PCRE's Unicode character property table had a bug
1923        which caused it to generate incorrect table entries for sequences of
1924        characters that have the same character type, but are in different scripts.
1925        It amalgamated them into a single range, with the script of the first of
1926        them. In other words, some characters were in the wrong script. There were
1927        thirteen such cases, affecting characters in the following ranges:
1928    
1929          U+002b0 - U+002c1
1930          U+0060c - U+0060d
1931          U+0061e - U+00612
1932          U+0064b - U+0065e
1933          U+0074d - U+0076d
1934          U+01800 - U+01805
1935          U+01d00 - U+01d77
1936          U+01d9b - U+01dbf
1937          U+0200b - U+0200f
1938          U+030fc - U+030fe
1939          U+03260 - U+0327f
1940          U+0fb46 - U+0fbb1
1941          U+10450 - U+1049d
1942    
1943    12. The -o option (show only the matching part of a line) for pcregrep was not
1944        compatible with GNU grep in that, if there was more than one match in a
1945        line, it showed only the first of them. It now behaves in the same way as
1946        GNU grep.
1947    
1948    13. If the -o and -v options were combined for pcregrep, it printed a blank
1949        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1950        does the same. The return code can be used to tell if there were any
1951        non-matching lines.
1952    
1953    14. Added --file-offsets and --line-offsets to pcregrep.
1954    
1955    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1956        infinitely looping recursion. The bug was that positive lookaheads were not
1957        being skipped when checking for a possible empty match (negative lookaheads
1958        and both kinds of lookbehind were skipped).
1959    
1960    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1961        inclusion of <windows.h> to before rather than after the definition of
1962        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1963    
1964    17. Specifying a possessive quantifier with a specific limit for a Unicode
1965        character property caused pcre_compile() to compile bad code, which led at
1966        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1967        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1968        caused the error; without that there was no problem.
1969    
1970    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1971    
1972    19. Added --enable-pcretest-libreadline.
1973    
1974    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1975        RE::GlobalReplace(). As a result, the number of replacements returned was
1976        double what it should be. I removed one of the increments, but Craig sent a
1977        later patch that removed the other one (the right fix) and added unit tests
1978        that check the return values (which was not done before).
1979    
1980    21. Several CMake things:
1981    
1982        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1983            the names libpcre and libpcreposix, not just pcre and pcreposix.
1984    
1985        (2) The above change means that pcretest and pcregrep are now correctly
1986            linked with the newly-built libraries, not previously installed ones.
1987    
1988        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1989    
1990    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1991        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1992        UTF-8 newline character). The key issue is that the pattern starts .*;
1993        this means that the match must be either at the beginning, or after a
1994        newline. The bug was in the code for advancing after a failed match and
1995        checking that the new position followed a newline. It was not taking
1996        account of UTF-8 characters correctly.
1997    
1998    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1999        character classes. PCRE was not treating the sequence [:...:] as a
2000        character class unless the ... were all letters. Perl, however, seems to
2001        allow any characters between [: and :], though of course it rejects as
2002        unknown any "names" that contain non-letters, because all the known class
2003        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2004        for example, whereas PCRE did not - it did not recognize a POSIX character
2005        class. This seemed a bit dangerous, so the code has been changed to be
2006        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2007        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2008        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2009        Perl does, and where it didn't before.
2010    
2011    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2012        Windows environments %n is disabled by default.
2013    
2014    
2015    Version 7.4 21-Sep-07
2016    ---------------------
2017    
2018    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2019        means that a class such as [\s] counted as "explicit reference to CR or
2020        LF". That isn't really right - the whole point of the change was to try to
2021        help when there was an actual mention of one of the two characters. So now
2022        the change happens only if \r or \n (or a literal CR or LF) character is
2023        encountered.
2024    
2025    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2026        of both had grown to the point where there were only 3 bits left.
2027        Fortunately, there was spare space in the data structure, and so I have
2028        moved the internal flags into a new 16-bit field to free up more option
2029        bits.
2030    
2031    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2032        but did not set the internal JCHANGED flag - either of these is enough to
2033        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2034        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2035        start sets both bits.
2036    
2037    4.  Added options (at build time, compile time, exec time) to change \R from
2038        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2039    
2040    5.  doc/pcresyntax.html was missing from the distribution.
2041    
2042    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2043        compatibility, even though it is no longer used.
2044    
2045    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2046        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2047        windows.h file is present (where different names are used). [This was
2048        reversed later after testing - see 16 below.]
2049    
2050    8.  Changed all #include <config.h> to #include "config.h". There were also
2051        some further <pcre.h> cases that I changed to "pcre.h".
2052    
2053    9.  When pcregrep was used with the --colour option, it missed the line ending
2054        sequence off the lines that it output.
2055    
2056    10. It was pointed out to me that arrays of string pointers cause lots of
2057        relocations when a shared library is dynamically loaded. A technique of
2058        using a single long string with a table of offsets can drastically reduce
2059        these. I have refactored PCRE in four places to do this. The result is
2060        dramatic:
2061    
2062          Originally:                          290
2063          After changing UCP table:            187
2064          After changing error message table:   43
2065          After changing table of "verbs"       36
2066          After changing table of Posix names   22
2067    
2068        Thanks to the folks working on Gregex for glib for this insight.
2069    
2070    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2071        unicode-properties was also set.
2072    
2073    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2074    
2075    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2076        checked only for CRLF.
2077    
2078    14. Added casts to pcretest.c to avoid compiler warnings.
2079    
2080    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2081    
2082    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2083        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2084        entirely. This removes changes made in 7 above.
2085    
2086    17. The CMake files have been updated, and there is now more information about
2087        building with CMake in the NON-UNIX-USE document.
2088    
2089    
2090    Version 7.3 28-Aug-07
2091    ---------------------
2092    
2093     1. In the rejigging of the build system that eventually resulted in 7.1, the
2094        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2095        brackets there is not right, since it causes compilers to look for an
2096        installed pcre.h, not the version that is in the source that is being
2097        compiled (which of course may be different). I have changed it back to:
2098    
2099          #include "pcre.h"
2100    
2101        I have a vague recollection that the change was concerned with compiling in
2102        different directories, but in the new build system, that is taken care of
2103        by the VPATH setting the Makefile.
2104    
2105     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2106        when the subject happened to end in the byte 0x85 (e.g. if the last
2107        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2108        characters but of course it shouldn't be taken as a newline when it is part
2109        of another character. The bug was that, for an unlimited repeat of . in
2110        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2111        characters when looking for a newline.
2112    
2113     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2114    
2115     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2116        in debug output.
2117    
2118     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2119        long printing in the pcrecpp unittest when running under MinGW.
2120    
2121     6. ESC_K was left out of the EBCDIC table.
2122    
2123     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2124        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2125        limit also applies to "virtual nesting" when a pattern is recursive, and in
2126        this case 1000 isn't so big. I have been able to remove this limit at the
2127        expense of backing off one optimization in certain circumstances. Normally,
2128        when pcre_exec() would call its internal match() function recursively and
2129        immediately return the result unconditionally, it uses a "tail recursion"
2130        feature to save stack. However, when a subpattern that can match an empty
2131        string has an unlimited repetition quantifier, it no longer makes this
2132        optimization. That gives it a stack frame in which to save the data for
2133        checking that an empty string has been matched. Previously this was taken
2134        from the 1000-entry workspace that had been reserved. So now there is no
2135        explicit limit, but more stack is used.
2136    
2137     8. Applied Daniel's patches to solve problems with the import/export magic
2138        syntax that is required for Windows, and which was going wrong for the
2139        pcreposix and pcrecpp parts of the library. These were overlooked when this
2140        problem was solved for the main library.
2141    
2142     9. There were some crude static tests to avoid integer overflow when computing
2143        the size of patterns that contain repeated groups with explicit upper
2144        limits. As the maximum quantifier is 65535, the maximum group length was
2145        set at 30,000 so that the product of these two numbers did not overflow a
2146        32-bit integer. However, it turns out that people want to use groups that
2147        are longer than 30,000 bytes (though not repeat them that many times).
2148        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2149        made it possible to implement the integer overflow checks in a much more
2150        dynamic way, which I have now done. The artificial limitation on group
2151        length has been removed - we now have only the limit on the total length of
2152        the compiled pattern, which depends on the LINK_SIZE setting.
2153    
2154    10. Fixed a bug in the documentation for get/copy named substring when
2155        duplicate names are permitted. If none of the named substrings are set, the
2156        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2157        empty string.
2158    
2159    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2160        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2161        because the ] is interpreted as the first data character and the
2162        terminating ] is not found. PCRE has been made compatible with Perl in this
2163        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2164        cause memory overwriting.
2165    
2166    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2167        string has been matched (to stop an infinite loop). It was not recognizing
2168        a conditional subpattern that could match an empty string if that
2169        subpattern was within another subpattern. For example, it looped when
2170        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2171        condition was not nested. This bug has been fixed.
2172    
2173    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2174        past the start of the subject in the presence of bytes with the top bit
2175        set, for example "\x8aBCD".
2176    
2177    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2178        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2179    
2180    14. Optimized (?!) to (*FAIL).
2181    
2182    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2183        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2184        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2185        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2186        does: it's just the validity check that is more restrictive.
2187    
2188    16. Inserted checks for integer overflows during escape sequence (backslash)
2189        processing, and also fixed erroneous offset values for syntax errors during
2190        backslash processing.
2191    
2192    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2193        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2194    
2195    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2196        caused an overrun.
2197    
2198    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2199        something other than just ASCII characters) inside a group that had an
2200        unlimited repeat caused a loop at compile time (while checking to see
2201        whether the group could match an empty string).
2202    
2203    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2204        [\P{Any}] did so. (Error in the code for printing property names.)
2205    
2206    21. An orphan \E inside a character class could cause a crash.
2207    
2208    22. A repeated capturing bracket such as (A)? could cause a wild memory
2209        reference during compilation.
2210    
2211    23. There are several functions in pcre_compile() that scan along a compiled
2212        expression for various reasons (e.g. to see if it's fixed length for look
2213        behind). There were bugs in these functions when a repeated \p or \P was
2214        present in the pattern. These operators have additional parameters compared
2215        with \d, etc, and these were not being taken into account when moving along
2216        the compiled data. Specifically:
2217    
2218        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2219            length.
2220    
2221        (b) An item such as \pL+ within a repeated group could cause crashes or
2222            loops.
2223    
2224        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2225            "reference to non-existent subpattern" error.
2226    
2227        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2228    
2229    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2230        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2231    
2232    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2233    
2234    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2235        character were causing crashes (broken optimization).
2236    
2237    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2238        \p or \P) caused a compile-time loop.
2239    
2240    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2241        break. For example, the unstudied pattern [\r\n]A does not match the string
2242        "\r\nA" because change 7.0/46 below moves the current point on by two
2243        characters after failing to match at the start. However, the pattern \nA
2244        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2245        the same is true. There doesn't seem any very clean way out of this, but
2246        what I have chosen to do makes the common cases work: PCRE now takes note
2247        of whether there can be an explicit match for \r or \n anywhere in the
2248        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2249        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2250        pattern has explicit CR or LF references.
2251    
2252    29. Added (*CR) etc for changing newline setting at start of pattern.
2253    
2254    
2255    Version 7.2 19-Jun-07
2256  ---------------------  ---------------------
2257    
2258   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2259      which is apparently normally available under Windows.      which is apparently normally available under Windows.
2260    
2261   2. Re-jig the pcregrep tests with different newline settings in an attempt   2. Re-jig the pcregrep tests with different newline settings in an attempt
2262      to make them independent of the local environment's newline setting.      to make them independent of the local environment's newline setting.
2263    
2264     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2265    
2266     4. Some of the "internals" tests were previously cut out when the link size
2267        was not 2, because the output contained actual offsets. The recent new
2268        "Z" feature of pcretest means that these can be cut out, making the tests
2269        usable with all link sizes.
2270    
2271     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2272        stack recursion. This gives a massive performance boost under BSD, but just
2273        a small improvement under Linux. However, it saves one field in the frame
2274        in all cases.
2275    
2276     6. Added more features from the forthcoming Perl 5.10:
2277    
2278        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2279            recursion call. It refers to the nth most recently opened parentheses.
2280    
2281        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2282            to be opened parentheses.
2283    
2284        (c) Conditions that refer to capturing parentheses can be specified
2285            relatively, for example, (?(-2)... or (?(+3)...
2286    
2287        (d) \K resets the start of the current match so that everything before
2288            is not part of it.
2289    
2290        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2291    
2292        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2293            reference syntax.
2294    
2295        (g) (?| introduces a group in which the numbering of parentheses in each
2296            alternative starts with the same number.
2297    
2298        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2299    
2300     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2301        PCRE_INFO_JCHANGED.
2302    
2303     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2304        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2305        for detecting groups that can match an empty string.
2306    
2307     9. A pattern with a very large number of alternatives (more than several
2308        hundred) was running out of internal workspace during the pre-compile
2309        phase, where pcre_compile() figures out how much memory will be needed. A
2310        bit of new cunning has reduced the workspace needed for groups with
2311        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2312        workspace instead of running out of the 4096 that are available.
2313    
2314    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2315    
2316    11. Applied patch from Google to remove an optimization that didn't quite work.
2317        The report of the bug said:
2318    
2319          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2320          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2321          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2322    
2323    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2324        it matched the wrong number of bytes.
2325    
2326    
2327  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
2328  ---------------------  ---------------------

Legend:
Removed from v.158  
changed lines
  Added in v.1245

  ViewVC Help
Powered by ViewVC 1.1.5