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

Diff of /code/trunk/ChangeLog

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

revision 193 by ph10, Mon Jul 30 11:34:18 2007 UTC revision 1244 by zherczeg, Tue Feb 5 12:03:43 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 05-Jul-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    
40    Version 8.32 30-November-2012
41    -----------------------------
42    
43    1.  Improved JIT compiler optimizations for first character search and single
44        character iterators.
45    
46    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
47        Patch by Daniel Richard G.
48    
49    3.  Single character iterator optimizations in the JIT compiler.
50    
51    4.  Improved JIT compiler optimizations for character ranges.
52    
53    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
54        Reported by Giuseppe D'Angelo.
55    
56    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
57        of a line, was being set incorrectly in cases where .* appeared inside
58        atomic brackets at the start of a pattern, or where there was a subsequent
59        *PRUNE or *SKIP.
60    
61    7.  Improved instruction cache flush for POWER/PowerPC.
62        Patch by Daniel Richard G.
63    
64    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
65        grep:
66    
67        (a) There is now no limit to the number of patterns to be matched.
68    
69        (b) An error is given if a pattern is too long.
70    
71        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
72            are now supported.
73    
74        (d) --exclude-from and --include-from (multiple use) have been added.
75    
76        (e) Exclusions and inclusions now apply to all files and directories, not
77            just to those obtained from scanning a directory recursively.
78    
79        (f) Multiple uses of -f and --file-list are now supported.
80    
81        (g) In a Windows environment, the default for -d has been changed from
82            "read" (the GNU grep default) to "skip", because otherwise the presence
83            of a directory in the file list provokes an error.
84    
85        (h) The documentation has been revised and clarified in places.
86    
87    9.  Improve the matching speed of capturing brackets.
88    
89    10. Changed the meaning of \X so that it now matches a Unicode extended
90        grapheme cluster.
91    
92    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
93        out POSIX threads when JIT support is configured.
94    
95    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
96    
97    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
98        structure could go wrong in environments where size_t is not the same size
99        as int.
100    
101    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
102        set.
103    
104    15. The EBCDIC support had decayed; later updates to the code had included
105        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
106        general tidy up of EBCDIC-related issues, and the documentation was also
107        not quite right. There is now a test that can be run on ASCII systems to
108        check some of the EBCDIC-related things (but is it not a full test).
109    
110    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
111        in a small tidy to the code.
112    
113    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
114    
115    18. If the --only-matching (-o) option in pcregrep is specified multiple
116        times, each one causes appropriate output. For example, -o1 -o2 outputs the
117        substrings matched by the 1st and 2nd capturing parentheses. A separating
118        string can be specified by --om-separator (default empty).
119    
120    19. Improving the first n character searches.
121    
122    20. Turn case lists for horizontal and vertical white space into macros so that
123        they are defined only once.
124    
125    21. This set of changes together give more compatible Unicode case-folding
126        behaviour for characters that have more than one other case when UCP
127        support is available.
128    
129        (a) The Unicode property table now has offsets into a new table of sets of
130            three or more characters that are case-equivalent. The MultiStage2.py
131            script that generates these tables (the pcre_ucd.c file) now scans
132            CaseFolding.txt instead of UnicodeData.txt for character case
133            information.
134    
135        (b) The code for adding characters or ranges of characters to a character
136            class has been abstracted into a generalized function that also handles
137            case-independence. In UTF-mode with UCP support, this uses the new data
138            to handle characters with more than one other case.
139    
140        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
141            whose other case is greater than 256 are now correctly matched
142            caselessly. Previously, the high codepoint matched the low one, but not
143            vice versa.
144    
145        (d) The processing of \h, \H, \v, and \ in character classes now makes use
146            of the new class addition function, using character lists defined as
147            macros alongside the case definitions of 20 above.
148    
149        (e) Caseless back references now work with characters that have more than
150            one other case.
151    
152        (f) General caseless matching of characters with more than one other case
153            is supported.
154    
155    22. Unicode character properties were updated from Unicode 6.2.0
156    
157    23. Improved CMake support under Windows. Patch by Daniel Richard G.
158    
159    24. Add support for 32-bit character strings, and UTF-32
160    
161    25. Major JIT compiler update (code refactoring and bugfixing).
162        Experimental Sparc 32 support is added.
163    
164    26. Applied a modified version of Daniel Richard G's patch to create
165        pcre.h.generic and config.h.generic by "make" instead of in the
166        PrepareRelease script.
167    
168    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
169        pcre_compile.c when checking for a zero character.
170    
171    28. Introducing a native interface for JIT. Through this interface, the compiled
172        machine code can be directly executed. The purpose of this interface is to
173        provide fast pattern matching, so several sanity checks are not performed.
174        However, feature tests are still performed. The new interface provides
175        1.4x speedup compared to the old one.
176    
177    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
178        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
179        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
180    
181    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
182        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
183        the "old" RFC 2279). Instead, it ended up passing a negative length to
184        pcre_exec().
185    
186    31. Add support for GCC's visibility feature to hide internal functions.
187    
188    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
189        "unknown -C option" after outputting 0 or 1.
190    
191    33. There is now support for generating a code coverage report for the test
192        suite in environments where gcc is the compiler and lcov is installed. This
193        is mainly for the benefit of the developers.
194    
195    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
196        unaddressable using valgrind annotations, allowing valgrind to detect
197        invalid memory accesses. This is mainly for the benefit of the developers.
198    
199    25. (*UTF) can now be used to start a pattern in any of the three libraries.
200    
201    26. Give configure error if --enable-cpp but no C++ compiler found.
202    
203    
204    Version 8.31 06-July-2012
205    -------------------------
206    
207    1.  Fixing a wrong JIT test case and some compiler warnings.
208    
209    2.  Removed a bashism from the RunTest script.
210    
211    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
212        to unsigned type, result still unsigned" that was given by an MS compiler
213        on encountering the code "-sizeof(xxx)".
214    
215    4.  Partial matching support is added to the JIT compiler.
216    
217    5.  Fixed several bugs concerned with partial matching of items that consist
218        of more than one character:
219    
220        (a) /^(..)\1/ did not partially match "aba" because checking references was
221            done on an "all or nothing" basis. This also applied to repeated
222            references.
223    
224        (b) \R did not give a hard partial match if \r was found at the end of the
225            subject.
226    
227        (c) \X did not give a hard partial match after matching one or more
228            characters at the end of the subject.
229    
230        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
231            a partial match for the string "\r".
232    
233        (e) When newline was set to CRLF, the metacharacter "." did not recognize
234            a partial match for a CR character at the end of the subject string.
235    
236    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
237        running pcretest, the text "(JIT)" added to the output whenever JIT is
238        actually used to run the match.
239    
240    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
241        or /S+[+] with a digit between 1 and 7.
242    
243    8.  OP_NOT now supports any UTF character not just single-byte ones.
244    
245    9.  (*MARK) control verb is now supported by the JIT compiler.
246    
247    10. The command "./RunTest list" lists the available tests without actually
248        running any of them. (Because I keep forgetting what they all are.)
249    
250    11. Add PCRE_INFO_MAXLOOKBEHIND.
251    
252    12. Applied a (slightly modified) user-supplied patch that improves performance
253        when the heap is used for recursion (compiled with --disable-stack-for-
254        recursion). Instead of malloc and free for each heap frame each time a
255        logical recursion happens, frames are retained on a chain and re-used where
256        possible. This sometimes gives as much as 30% improvement.
257    
258    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
259        call.
260    
261    14. As documented, (*COMMIT) is now confined to within a positive assertion.
262    
263    15. It is now possible to link pcretest with libedit as an alternative to
264        libreadline.
265    
266    16. (*COMMIT) control verb is now supported by the JIT compiler.
267    
268    17. The Unicode data tables have been updated to Unicode 6.1.0.
269    
270    18. Added --file-list option to pcregrep.
271    
272    19. Added binary file support to pcregrep, including the -a, --binary-files,
273        -I, and --text options.
274    
275    20. The madvise function is renamed for posix_madvise for QNX compatibility
276        reasons. Fixed by Giuseppe D'Angelo.
277    
278    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
279    
280    22. Changed the option for creating symbolic links for 16-bit man pages from
281        -s to -sf so that re-installing does not cause issues.
282    
283    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
284    
285    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
286        matches in certain environments (the workspace was not being correctly
287        retained). Also added to pcre_dfa_exec() a simple plausibility check on
288        some of the workspace data at the beginning of a restart.
289    
290    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
291        was not doing so when it should - probably a typo introduced by SVN 528
292        (change 8.10/14).
293    
294    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
295        \w+ when the character tables indicated that \x{c4} was a word character.
296        There were several related cases, all because the tests for doing a table
297        lookup were testing for characters less than 127 instead of 255.
298    
299    27. If a pattern contains capturing parentheses that are not used in a match,
300        their slots in the ovector are set to -1. For those that are higher than
301        any matched groups, this happens at the end of processing. In the case when
302        there were back references that the ovector was too small to contain
303        (causing temporary malloc'd memory to be used during matching), and the
304        highest capturing number was not used, memory off the end of the ovector
305        was incorrectly being set to -1. (It was using the size of the temporary
306        memory instead of the true size.)
307    
308    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
309        ovector size, it uses memory at the end of the block that it has got.
310    
311    29. Check for an overlong MARK name and give an error at compile time. The
312        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
313    
314    30. JIT compiler update.
315    
316    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
317        Rill for the patch.
318    
319    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
320    
321    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
322    
323    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
324        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
325        was enabled.
326    
327    35. Improve JIT code generation for greedy plus quantifier.
328    
329    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
330        1 to "aa" instead of to an empty string. The bug affected repeated groups
331        that could potentially match an empty string.
332    
333    37. Optimizing single character iterators in JIT.
334    
335    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
336        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
337        codepoints that are too big for the mode are faulted, and in a UTF mode,
338        disallowed codepoints are also faulted.
339    
340    39. If PCRE was compiled with UTF support, in three places in the DFA
341        matcher there was code that should only have been obeyed in UTF mode, but
342        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
343        processing when bytes with values greater than 127 were present. In 16-bit
344        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
345        both cases the values are those that cannot be the first data item in a UTF
346        character. The three items that might have provoked this were recursions,
347        possessively repeated groups, and atomic groups.
348    
349    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
350        and pcregrep, because some OS require shared objects to be explicitly
351        passed to ld, causing the link step to fail if they are not.
352    
353    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
354        mode, patterns that started with \h* or \R* might be incorrectly matched.
355    
356    
357    Version 8.30 04-February-2012
358    -----------------------------
359    
360    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
361        name is defined in ctype.h.
362    
363    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
364        only in quite long subpatterns.
365    
366    3.  Removed the function pcre_info(), which has been obsolete and deprecated
367        since it was replaced by pcre_fullinfo() in February 2000.
368    
369    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
370        match a (*MARK), and the match failed at the start of the subject, a
371        reference to memory before the start of the subject could occur. This bug
372        was introduced by fix 17 of release 8.21.
373    
374    5.  A reference to an unset group with zero minimum repetition was giving
375        totally wrong answers (in non-JavaScript-compatibility mode). For example,
376        /(another)?(\1?)test/ matched against "hello world test". This bug was
377        introduced in release 8.13.
378    
379    6.  Add support for 16-bit character strings (a large amount of work involving
380        many changes and refactorings).
381    
382    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
383        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
384        from a file.
385    
386    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
387        rounding is not applied in this particular case).
388    
389    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
390        if they appear, or are escaped, in patterns.
391    
392    10. Get rid of a number of -Wunused-but-set-variable warnings.
393    
394    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
395        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
396        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
397        also returns the mark "x". This bug applied to capturing parentheses,
398        non-capturing parentheses, and atomic parentheses. It also applied to some
399        assertions.
400    
401    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
402        information out of configure.ac instead of relying on pcre.h.generic, which
403        is not stored in the repository.
404    
405    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
406        -lreadline.
407    
408    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
409    
410    15. Applied Graycode's patch to put the top-level frame on the stack rather
411        than the heap when not using the stack for recursion. This gives a
412        performance improvement in many cases when recursion is not deep.
413    
414    16. Experimental code added to "pcretest -C" to output the stack frame size.
415    
416    
417    Version 8.21 12-Dec-2011
418    ------------------------
419    
420    1.  Updating the JIT compiler.
421    
422    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
423        are added as well.
424    
425    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
426        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
427        calling _pcre_jit_exec. Some extra comments are added.
428    
429    4.  (*MARK) settings inside atomic groups that do not contain any capturing
430        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
431        was introduced by change 18 for 8.20.
432    
433    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
434        ECMA-262 standard.
435    
436    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
437        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
438        This bug was probably introduced by change 9 of 8.13.
439    
440    7.  While fixing 6 above, I noticed that a number of other items were being
441        incorrectly rejected as "not fixed length". This arose partly because newer
442        opcodes had not been added to the fixed-length checking code. I have (a)
443        corrected the bug and added tests for these items, and (b) arranged for an
444        error to occur if an unknown opcode is encountered while checking for fixed
445        length instead of just assuming "not fixed length". The items that were
446        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
447        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
448        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
449    
450    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
451        being incorrectly compiled and would have given unpredicatble results.
452    
453    9.  A possessively repeated subpattern with minimum repeat count greater than
454        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
455        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
456        the first (A) could occur when it should not.
457    
458    10. Add a cast and remove a redundant test from the code.
459    
460    11. JIT should use pcre_malloc/pcre_free for allocation.
461    
462    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
463        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
464        is anything other than /usr, -L is still shown).
465    
466    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
467    
468    14. Perl does not support \N without a following name in a [] class; PCRE now
469        also gives an error.
470    
471    15. If a forward reference was repeated with an upper limit of around 2000,
472        it caused the error "internal error: overran compiling workspace". The
473        maximum number of forward references (including repeats) was limited by the
474        internal workspace, and dependent on the LINK_SIZE. The code has been
475        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
476        the default depends on LINK_SIZE. There is a new upper limit (for safety)
477        of around 200,000 forward references. While doing this, I also speeded up
478        the filling in of repeated forward references.
479    
480    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
481        incorrectly expecting the subject to contain another "a" after the start.
482    
483    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
484        in the match, the SKIP should be ignored. This was not happening; instead
485        the SKIP was being treated as NOMATCH. For patterns such as
486        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
487        tested.
488    
489    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
490        now much more compatible with Perl, in particular in cases where the result
491        is a non-match for a non-anchored pattern. For example, if
492        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
493        "m", where previously it did not return a name. A side effect of this
494        change is that for partial matches, the last encountered mark name is
495        returned, as for non matches. A number of tests that were previously not
496        Perl-compatible have been moved into the Perl-compatible test files. The
497        refactoring has had the pleasing side effect of removing one argument from
498        the match() function, thus reducing its stack requirements.
499    
500    19. If the /S+ option was used in pcretest to study a pattern using JIT,
501        subsequent uses of /S (without +) incorrectly behaved like /S+.
502    
503    21. Retrieve executable code size support for the JIT compiler and fixing
504        some warnings.
505    
506    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
507        not work when the shorter character appeared right at the end of the
508        subject string.
509    
510    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
511        systems.
512    
513    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
514        output it when the /M option is used in pcretest.
515    
516    25. The CheckMan script was not being included in the distribution. Also, added
517        an explicit "perl" to run Perl scripts from the PrepareRelease script
518        because this is reportedly needed in Windows.
519    
520    26. If study data was being save in a file and studying had not found a set of
521        "starts with" bytes for the pattern, the data written to the file (though
522        never used) was taken from uninitialized memory and so caused valgrind to
523        complain.
524    
525    27. Updated RunTest.bat as provided by Sheri Pierce.
526    
527    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
528    
529    29. Computation of memory usage for the table of capturing group names was
530        giving an unnecessarily large value.
531    
532    
533    Version 8.20 21-Oct-2011
534    ------------------------
535    
536    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
537        a POSIX class. After further experiments with Perl, which convinced me that
538        Perl has bugs and confusions, a closing square bracket is no longer allowed
539        in a POSIX name. This bug also affected patterns with classes that started
540        with full stops.
541    
542    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
543        captured substring, but while checking the failing first alternative,
544        substring 1 is temporarily captured. If the output vector supplied to
545        pcre_exec() was not big enough for this capture, the yield of the function
546        was still zero ("insufficient space for captured substrings"). This cannot
547        be totally fixed without adding another stack variable, which seems a lot
548        of expense for a edge case. However, I have improved the situation in cases
549        such as /(a)(b)x|abc/ matched against "abc", where the return code
550        indicates that fewer than the maximum number of slots in the ovector have
551        been set.
552    
553    3.  Related to (2) above: when there are more back references in a pattern than
554        slots in the output vector, pcre_exec() uses temporary memory during
555        matching, and copies in the captures as far as possible afterwards. It was
556        using the entire output vector, but this conflicts with the specification
557        that only 2/3 is used for passing back captured substrings. Now it uses
558        only the first 2/3, for compatibility. This is, of course, another edge
559        case.
560    
561    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
562        main code base, and can be used by building with --enable-jit. When this is
563        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
564        runtime --no-jit option is given.
565    
566    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
567        ovector, the return from the function was zero, implying that there were
568        other matches that did not fit. The correct "exactly full" value is now
569        returned.
570    
571    6.  If a subpattern that was called recursively or as a subroutine contained
572        (*PRUNE) or any other control that caused it to give a non-standard return,
573        invalid errors such as "Error -26 (nested recursion at the same subject
574        position)" or even infinite loops could occur.
575    
576    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
577        computing the minimum length on reaching *ACCEPT, and so ended up with the
578        wrong value of 1 rather than 0. Further investigation indicates that
579        computing a minimum subject length in the presence of *ACCEPT is difficult
580        (think back references, subroutine calls), and so I have changed the code
581        so that no minimum is registered for a pattern that contains *ACCEPT.
582    
583    8.  If (*THEN) was present in the first (true) branch of a conditional group,
584        it was not handled as intended. [But see 16 below.]
585    
586    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
587        Sheri Pierce.
588    
589    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
590        the first byte in a match must be "a".
591    
592    11. Change 17 for 8.13 increased the recursion depth for patterns like
593        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
594        pattern contains any instances of (*THEN). If it does not, the old
595        optimizations are restored. It would be nice to do this on a per-group
596        basis, but at the moment that is not feasible.
597    
598    12. In some environments, the output of pcretest -C is CRLF terminated. This
599        broke RunTest's code that checks for the link size. A single white space
600        character after the value is now allowed for.
601    
602    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
603        For "fr", it uses the Windows-specific input and output files.
604    
605    14. If (*THEN) appeared in a group that was called recursively or as a
606        subroutine, it did not work as intended. [But see next item.]
607    
608    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
609        pattern fragments (but not containing any | characters). If A and B are
610        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
611        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
612        D. In other words, Perl considers parentheses that do not contain any |
613        characters to be part of a surrounding alternative, whereas PCRE was
614        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
615        differently. PCRE now behaves in the same way as Perl, except in the case
616        of subroutine/recursion calls such as (?1) which have in any case always
617        been different (but PCRE had them first :-).
618    
619    16. Related to 15 above: Perl does not treat the | in a conditional group as
620        creating alternatives. Such a group is treated in the same way as an
621        ordinary group without any | characters when processing (*THEN). PCRE has
622        been changed to match Perl's behaviour.
623    
624    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
625        RunGrepTest script failed.
626    
627    18. Change 22 for version 13 caused atomic groups to use more stack. This is
628        inevitable for groups that contain captures, but it can lead to a lot of
629        stack use in large patterns. The old behaviour has been restored for atomic
630        groups that do not contain any capturing parentheses.
631    
632    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
633        suppress the check for a minimum subject length at run time. (If it was
634        given to pcre_exec() or pcre_dfa_exec() it did work.)
635    
636    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
637        fail to work when decoding hex characters in data strings in EBCDIC
638        environments.
639    
640    21. It appears that in at least one Mac OS environment, the isxdigit() function
641        is implemented as a macro that evaluates to its argument more than once,
642        contravening the C 90 Standard (I haven't checked a later standard). There
643        was an instance in pcretest which caused it to go wrong when processing
644        \x{...} escapes in subject strings. The has been rewritten to avoid using
645        things like p++ in the argument of isxdigit().
646    
647    
648    Version 8.13 16-Aug-2011
649    ------------------------
650    
651    1.  The Unicode data tables have been updated to Unicode 6.0.0.
652    
653    2.  Two minor typos in pcre_internal.h have been fixed.
654    
655    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
656        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
657        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
658    
659    4.  There were a number of related bugs in the code for matching backrefences
660        caselessly in UTF-8 mode when codes for the characters concerned were
661        different numbers of bytes. For example, U+023A and U+2C65 are an upper
662        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
663        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
664        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
665        2-byte code at the end of the subject (it thought there wasn't enough data
666        left).
667    
668    5.  Comprehensive information about what went wrong is now returned by
669        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
670        as the output vector has at least 2 elements. The offset of the start of
671        the failing character and a reason code are placed in the vector.
672    
673    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
674        now returned is for the first byte of the failing character, instead of the
675        last byte inspected. This is an incompatible change, but I hope it is small
676        enough not to be a problem. It makes the returned offset consistent with
677        pcre_exec() and pcre_dfa_exec().
678    
679    7.  pcretest now gives a text phrase as well as the error number when
680        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
681        failure, the offset and reason code are output.
682    
683    8.  When \R was used with a maximizing quantifier it failed to skip backwards
684        over a \r\n pair if the subsequent match failed. Instead, it just skipped
685        back over a single character (\n). This seems wrong (because it treated the
686        two characters as a single entity when going forwards), conflicts with the
687        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
688        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
689        has been changed.
690    
691    9.  Some internal refactoring has changed the processing so that the handling
692        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
693        time (the PCRE_DOTALL option was changed this way some time ago: version
694        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
695        which was always a bit of a fudge. It also means that there is one less
696        argument for the match() function, which reduces its stack requirements
697        slightly. This change also fixes an incompatibility with Perl: the pattern
698        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
699    
700    10. More internal refactoring has drastically reduced the number of recursive
701        calls to match() for possessively repeated groups such as (abc)++ when
702        using pcre_exec().
703    
704    11. While implementing 10, a number of bugs in the handling of groups were
705        discovered and fixed:
706    
707        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
708        (a|)*(?1) gave a compile-time internal error.
709        ((a|)+)+  did not notice that the outer group could match an empty string.
710        (^a|^)+   was not marked as anchored.
711        (.*a|.*)+ was not marked as matching at start or after a newline.
712    
713    12. Yet more internal refactoring has removed another argument from the match()
714        function. Special calls to this function are now indicated by setting a
715        value in a variable in the "match data" data block.
716    
717    13. Be more explicit in pcre_study() instead of relying on "default" for
718        opcodes that mean there is no starting character; this means that when new
719        ones are added and accidentally left out of pcre_study(), testing should
720        pick them up.
721    
722    14. The -s option of pcretest has been documented for ages as being an old
723        synonym of -m (show memory usage). I have changed it to mean "force study
724        for every regex", that is, assume /S for every regex. This is similar to -i
725        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
726        using it. It makes it easier to run collections of tests with and without
727        study enabled, and thereby test pcre_study() more easily. All the standard
728        tests are now run with and without -s (but some patterns can be marked as
729        "never study" - see 20 below).
730    
731    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
732        restoration of the capturing data to the outer values was not happening
733        correctly.
734    
735    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
736        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
737        pattern had matched an empty string, and so incorrectly returned a no
738        match.
739    
740    17. There was optimizing code for the last branch of non-capturing parentheses,
741        and also for the obeyed branch of a conditional subexpression, which used
742        tail recursion to cut down on stack usage. Unfortunately, now that there is
743        the possibility of (*THEN) occurring in these branches, tail recursion is
744        no longer possible because the return has to be checked for (*THEN). These
745        two optimizations have therefore been removed. [But see 8.20/11 above.]
746    
747    18. If a pattern containing \R was studied, it was assumed that \R always
748        matched two bytes, thus causing the minimum subject length to be
749        incorrectly computed because \R can also match just one byte.
750    
751    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
752        was incorrectly computed.
753    
754    20. If /S is present twice on a test pattern in pcretest input, it now
755        *disables* studying, thereby overriding the use of -s on the command line
756        (see 14 above). This is necessary for one or two tests to keep the output
757        identical in both cases.
758    
759    21. When (*ACCEPT) was used in an assertion that matched an empty string and
760        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
761    
762    22. When an atomic group that contained a capturing parenthesis was
763        successfully matched, but the branch in which it appeared failed, the
764        capturing was not being forgotten if a higher numbered group was later
765        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
766        group 1 to "a", when in fact it should be unset. This applied to multi-
767        branched capturing and non-capturing groups, repeated or not, and also to
768        positive assertions (capturing in negative assertions does not happen
769        in PCRE) and also to nested atomic groups.
770    
771    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
772        subject after a captured substring, to make it easier to tell which of a
773        number of identical substrings has been captured.
774    
775    24. The way atomic groups are processed by pcre_exec() has been changed so that
776        if they are repeated, backtracking one repetition now resets captured
777        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
778        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
779        "aaa". Previously, it would have been "a". As part of this code
780        refactoring, the way recursive calls are handled has also been changed.
781    
782    25. If an assertion condition captured any substrings, they were not passed
783        back unless some other capturing happened later. For example, if
784        (?(?=(a))a) was matched against "a", no capturing was returned.
785    
786    26. When studying a pattern that contained subroutine calls or assertions,
787        the code for finding the minimum length of a possible match was handling
788        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
789        group 1 called group 2 while simultaneously a separate group 2 called group
790        1). A stack overflow occurred in this case. I have fixed this by limiting
791        the recursion depth to 10.
792    
793    27. Updated RunTest.bat in the distribution to the version supplied by Tom
794        Fortmann. This supports explicit test numbers on the command line, and has
795        argument validation and error reporting.
796    
797    28. An instance of \X with an unlimited repeat could fail if at any point the
798        first character it looked at was a mark character.
799    
800    29. Some minor code refactoring concerning Unicode properties and scripts
801        should reduce the stack requirement of match() slightly.
802    
803    30. Added the '=' option to pcretest to check the setting of unused capturing
804        slots at the end of the pattern, which are documented as being -1, but are
805        not included in the return count.
806    
807    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
808        compiled something random. Now it gives a compile-time error (as does
809        Perl).
810    
811    32. A *MARK encountered during the processing of a positive assertion is now
812        recorded and passed back (compatible with Perl).
813    
814    33. If --only-matching or --colour was set on a pcregrep call whose pattern
815        had alternative anchored branches, the search for a second match in a line
816        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
817        matched the line "0102" twice. The same bug affected patterns that started
818        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
819        twice.
820    
821    34. Previously, PCRE did not allow quantification of assertions. However, Perl
822        does, and because of capturing effects, quantifying parenthesized
823        assertions may at times be useful. Quantifiers are now allowed for
824        parenthesized assertions.
825    
826    35. A minor code tidy in pcre_compile() when checking options for \R usage.
827    
828    36. \g was being checked for fancy things in a character class, when it should
829        just be a literal "g".
830    
831    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
832        appearance of a nested POSIX class supersedes an apparent external class.
833        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
834        unescaped square brackets may also appear as part of class names. For
835        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
836        more like Perl. (But see 8.20/1 above.)
837    
838    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
839        was because it thought it was \N{name}, which is not supported).
840    
841    39. Add minix to OS list not supporting the -S option in pcretest.
842    
843    40. PCRE tries to detect cases of infinite recursion at compile time, but it
844        cannot analyze patterns in sufficient detail to catch mutual recursions
845        such as ((?1))((?2)). There is now a runtime test that gives an error if a
846        subgroup is called recursively as a subpattern for a second time at the
847        same position in the subject string. In previous releases this might have
848        been caught by the recursion limit, or it might have run out of stack.
849    
850    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
851        happen only once. PCRE was, however incorrectly giving a compile time error
852        "recursive call could loop indefinitely" because it cannot analyze the
853        pattern in sufficient detail. The compile time test no longer happens when
854        PCRE is compiling a conditional subpattern, but actual runaway loops are
855        now caught at runtime (see 40 above).
856    
857    42. It seems that Perl allows any characters other than a closing parenthesis
858        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
859        has been changed to be the same.
860    
861    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
862        as not to get warnings when autogen.sh is called. Also changed
863        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
864    
865    44. To help people who use pcregrep to scan files containing exceedingly long
866        lines, the following changes have been made:
867    
868        (a) The default value of the buffer size parameter has been increased from
869            8K to 20K. (The actual buffer used is three times this size.)
870    
871        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
872            PCRE is built.
873    
874        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
875            to be set at run time.
876    
877        (d) Numerical values in pcregrep options can be followed by K or M, for
878            example --buffer-size=50K.
879    
880        (e) If a line being scanned overflows pcregrep's buffer, an error is now
881            given and the return code is set to 2.
882    
883    45. Add a pointer to the latest mark to the callout data block.
884    
885    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
886        partial match of an empty string instead of no match. This was specific to
887        the use of ".".
888    
889    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
890        complete match instead of a partial match. This bug was dependent on both
891        the PCRE_UTF8 and PCRE_DOTALL options being set.
892    
893    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
894        starting byte set, because \b was not being ignored.
895    
896    
897    Version 8.12 15-Jan-2011
898    ------------------------
899    
900    1.  Fixed some typos in the markup of the man pages, and wrote a script that
901        checks for such things as part of the documentation building process.
902    
903    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
904        --match-limit and --recursion-limit options (added for 8.11). In
905        particular, this made one of the standard tests fail. (The integer value
906        went into the wrong half of a long int.)
907    
908    3.  If the --colour option was given to pcregrep with -v (invert match), it
909        did strange things, either producing crazy output, or crashing. It should,
910        of course, ignore a request for colour when reporting lines that do not
911        match.
912    
913    4.  Another pcregrep bug caused similar problems if --colour was specified with
914        -M (multiline) and the pattern match finished with a line ending.
915    
916    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
917        matched in multiline mode, the following line was shown as part of the
918        match. This seems wrong, so I have changed it.
919    
920    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
921        the check for further matches in the same line (so they could be coloured)
922        to overrun the end of the current line. If another match was found, it was
923        incorrectly shown (and then shown again when found in the next line).
924    
925    7.  If pcregrep was compiled under Windows, there was a reference to the
926        function pcregrep_exit() before it was defined. I am assuming this was
927        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
928        reported by a user. I've moved the definition above the reference.
929    
930    
931    Version 8.11 10-Dec-2010
932    ------------------------
933    
934    1.  (*THEN) was not working properly if there were untried alternatives prior
935        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
936        backtracked to try for "b" instead of moving to the next alternative branch
937        at the same level (in this case, to look for "c"). The Perl documentation
938        is clear that when (*THEN) is backtracked onto, it goes to the "next
939        alternative in the innermost enclosing group".
940    
941    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
942        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
943        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
944        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
945        (*THEN).
946    
947    3.  If \s appeared in a character class, it removed the VT character from
948        the class, even if it had been included by some previous item, for example
949        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
950        of \s, but is part of the POSIX "space" class.)
951    
952    4.  A partial match never returns an empty string (because you can always
953        match an empty string at the end of the subject); however the checking for
954        an empty string was starting at the "start of match" point. This has been
955        changed to the "earliest inspected character" point, because the returned
956        data for a partial match starts at this character. This means that, for
957        example, /(?<=abc)def/ gives a partial match for the subject "abc"
958        (previously it gave "no match").
959    
960    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
961        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
962        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
963        has an implication that the given string is incomplete (because a partial
964        match is preferred over a full match). For this reason, these items now
965        give a partial match in this situation. [Aside: previously, the one case
966        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
967        match rather than a full match, which was wrong by the old rules, but is
968        now correct.]
969    
970    6.  There was a bug in the handling of #-introduced comments, recognized when
971        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
972        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
973        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
974        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
975        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
976        places in pcre_compile().
977    
978    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
979        comments when looking ahead for named forward references to subpatterns,
980        the only newline sequence it recognized was NL. It now handles newlines
981        according to the set newline convention.
982    
983    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
984        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
985        cater for a lack of strerror(). These oversights have been fixed.
986    
987    9.  Added --match-limit and --recursion-limit to pcregrep.
988    
989    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
990    
991    11. When the -o option was used, pcregrep was setting a return code of 1, even
992        when matches were found, and --line-buffered was not being honoured.
993    
994    12. Added an optional parentheses number to the -o and --only-matching options
995        of pcregrep.
996    
997    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
998        can match an empty string. The code to do it in pcretest and pcredemo
999        needed fixing:
1000    
1001        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1002            only one byte after an empty string match just before CRLF (this case
1003            just got forgotten; "any" and "anycrlf" were OK).
1004    
1005        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1006            mode when an empty string match preceded an ASCII character followed by
1007            a non-ASCII character. (The code for advancing by one character rather
1008            than one byte was nonsense.)
1009    
1010        (c) The pcredemo.c sample program did not have any code at all to handle
1011            the cases when CRLF is a valid newline sequence.
1012    
1013    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1014        as a starting offset was within the subject string. There is now a new
1015        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1016        negative or greater than the length of the string. In order to test this,
1017        pcretest is extended to allow the setting of negative starting offsets.
1018    
1019    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1020        starting offset points to the beginning of a UTF-8 character was
1021        unnecessarily clumsy. I tidied it up.
1022    
1023    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1024        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1025    
1026    17. Nobody had reported that the --include_dir option, which was added in
1027        release 7.7 should have been called --include-dir (hyphen, not underscore)
1028        for compatibility with GNU grep. I have changed it to --include-dir, but
1029        left --include_dir as an undocumented synonym, and the same for
1030        --exclude-dir, though that is not available in GNU grep, at least as of
1031        release 2.5.4.
1032    
1033    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1034        characters from a string of bytes) have been redefined so as not to use
1035        loops, in order to improve performance in some environments. At the same
1036        time, I abstracted some of the common code into auxiliary macros to save
1037        repetition (this should not affect the compiled code).
1038    
1039    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1040        compile-time error is now given if \c is not followed by an ASCII
1041        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1042        different, and any byte value is allowed.)
1043    
1044    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1045        START_OPTIMIZE option, which is now allowed at compile time - but just
1046        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1047        to pcregrep and other applications that have no direct access to PCRE
1048        options. The new /Y option in pcretest sets this option when calling
1049        pcre_compile().
1050    
1051    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1052        back references. Groups containing recursive back references were forced to
1053        be atomic by that change, but in the case of named groups, the amount of
1054        memory required was incorrectly computed, leading to "Failed: internal
1055        error: code overflow". This has been fixed.
1056    
1057    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1058        pcretest.c, to avoid build problems in some Borland environments.
1059    
1060    
1061    Version 8.10 25-Jun-2010
1062    ------------------------
1063    
1064    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1065        THEN.
1066    
1067    2.  (*ACCEPT) was not working when inside an atomic group.
1068    
1069    3.  Inside a character class, \B is treated as a literal by default, but
1070        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1071        causes the error). The code is unchanged, but I tidied the documentation.
1072    
1073    4.  Inside a character class, PCRE always treated \R and \X as literals,
1074        whereas Perl faults them if its -w option is set. I have changed PCRE so
1075        that it faults them when PCRE_EXTRA is set.
1076    
1077    5.  Added support for \N, which always matches any character other than
1078        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1079    
1080    6.  When compiling pcregrep with newer versions of gcc which may have
1081        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1082        declared with attribute warn_unused_result" were given. Just casting the
1083        result to (void) does not stop the warnings; a more elaborate fudge is
1084        needed. I've used a macro to implement this.
1085    
1086    7.  Minor change to pcretest.c to avoid a compiler warning.
1087    
1088    8.  Added four artifical Unicode properties to help with an option to make
1089        \s etc use properties (see next item). The new properties are: Xan
1090        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1091    
1092    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1093        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1094        this option. Modified pcretest to add /W to test this facility. Added
1095        REG_UCP to make it available via the POSIX interface.
1096    
1097    10. Added --line-buffered to pcregrep.
1098    
1099    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1100        studied, and the match started with a letter with a code point greater than
1101        127 whose first byte was different to the first byte of the other case of
1102        the letter, the other case of this starting letter was not recognized
1103        (#976).
1104    
1105    12. If a pattern that was studied started with a repeated Unicode property
1106        test, for example, \p{Nd}+, there was the theoretical possibility of
1107        setting up an incorrect bitmap of starting bytes, but fortunately it could
1108        not have actually happened in practice until change 8 above was made (it
1109        added property types that matched character-matching opcodes).
1110    
1111    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1112        possible starting bytes for non-anchored patterns.
1113    
1114    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1115        \R, and also a number of cases that involve Unicode properties, both
1116        explicit and implicit when PCRE_UCP is set.
1117    
1118    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1119        input, it could crash or give wrong results if characters with values
1120        greater than 0xc0 were present in the subject string. (Detail: it assumed
1121        UTF-8 input when processing these items.)
1122    
1123    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1124        size_t is 64-bit (#991).
1125    
1126    17. Added a check for running out of memory when PCRE is compiled with
1127        --disable-stack-for-recursion (#990).
1128    
1129    18. If the last data line in a file for pcretest does not have a newline on
1130        the end, a newline was missing in the output.
1131    
1132    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1133        less than 128) in its various bitmaps. However, there is a facility for
1134        generating tables according to the current locale when PCRE is compiled. It
1135        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1136        characters, are recognized by isspace() and therefore were getting set in
1137        these tables, and indeed these tables seem to approximate to ISO 8859. This
1138        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1139        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1140        which of course cannot start UTF-8 characters. I have changed the code so
1141        that only real ASCII characters (less than 128) and the correct starting
1142        bytes for UTF-8 encodings are set for characters greater than 127 when in
1143        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1144        altogether.)
1145    
1146    20. Added the /T option to pcretest so as to be able to run tests with non-
1147        standard character tables, thus making it possible to include the tests
1148        used for 19 above in the standard set of tests.
1149    
1150    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1151        reference to a subpattern the other side of a comment that contains an
1152        opening parenthesis caused either an internal compiling error, or a
1153        reference to the wrong subpattern.
1154    
1155    
1156    Version 8.02 19-Mar-2010
1157    ------------------------
1158    
1159    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1160    
1161    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1162        configured.
1163    
1164    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1165        original author of that file, following a query about its status.
1166    
1167    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1168        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1169    
1170    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1171        quantifier applied to a forward-referencing subroutine call, could compile
1172        incorrect code or give the error "internal error: previously-checked
1173        referenced subpattern not found".
1174    
1175    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1176        variables to point to external functions. For these systems, therefore,
1177        pcre_malloc etc. are now initialized to local functions that call the
1178        relevant global functions.
1179    
1180    7.  There were two entries missing in the vectors called coptable and poptable
1181        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1182        I've fixed the data, and added a kludgy way of testing at compile time that
1183        the lengths are correct (equal to the number of opcodes).
1184    
1185    8.  Following on from 7, I added a similar kludge to check the length of the
1186        eint vector in pcreposix.c.
1187    
1188    9.  Error texts for pcre_compile() are held as one long string to avoid too
1189        much relocation at load time. To find a text, the string is searched,
1190        counting zeros. There was no check for running off the end of the string,
1191        which could happen if a new error number was added without updating the
1192        string.
1193    
1194    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1195    
1196    11. \K was not working if it appeared in an atomic group or in a group that
1197        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1198        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1199        the assertion is positive, but not if it is negative.
1200    
1201    12. Change 11 fortuitously reduced the size of the stack frame used in the
1202        "match()" function of pcre_exec.c by one pointer. Forthcoming
1203        implementation of support for (*MARK) will need an extra pointer on the
1204        stack; I have reserved it now, so that the stack frame size does not
1205        decrease.
1206    
1207    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1208        item in branch that calls a recursion is a subroutine call - as in the
1209        second branch in the above example - was incorrectly given the compile-
1210        time error "recursive call could loop indefinitely" because pcre_compile()
1211        was not correctly checking the subroutine for matching a non-empty string.
1212    
1213    14. The checks for overrunning compiling workspace could trigger after an
1214        overrun had occurred. This is a "should never occur" error, but it can be
1215        triggered by pathological patterns such as hundreds of nested parentheses.
1216        The checks now trigger 100 bytes before the end of the workspace.
1217    
1218    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1219    
1220    
1221    Version 8.01 19-Jan-2010
1222    ------------------------
1223    
1224    1.  If a pattern contained a conditional subpattern with only one branch (in
1225        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1226        computed the wrong minimum data length (which is of course zero for such
1227        subpatterns). This could cause incorrect "no match" results.
1228    
1229    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1230        the pattern is reset in the first branch, pcre_compile() failed with
1231        "internal error: code overflow at offset...". This happened only when
1232        the reset was to the original external option setting. (An optimization
1233        abstracts leading options settings into an external setting, which was the
1234        cause of this.)
1235    
1236    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1237        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1238        assertion pattern did not match (meaning that the assertion was true), it
1239        was incorrectly treated as false if the SKIP had been reached during the
1240        matching. This also applied to assertions used as conditions.
1241    
1242    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1243        assertion subpattern, including such a pattern used as a condition,
1244        unpredictable results occurred, instead of the error return
1245        PCRE_ERROR_DFA_UITEM.
1246    
1247    5.  The C++ GlobalReplace function was not working like Perl for the special
1248        situation when an empty string is matched. It now does the fancy magic
1249        stuff that is necessary.
1250    
1251    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1252        removed. (These were left over from very, very early versions of PCRE.)
1253    
1254    7.  Some cosmetic changes to the code to make life easier when compiling it
1255        as part of something else:
1256    
1257        (a) Change DEBUG to PCRE_DEBUG.
1258    
1259        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1260            called "current" as "current_branch", to prevent a collision with the
1261            Linux macro when compiled as a kernel module.
1262    
1263        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1264            prevent a collision with the Linux macro when compiled as a kernel
1265            module.
1266    
1267    8.  In pcre_compile() there are some checks for integer overflows that used to
1268        cast potentially large values to (double). This has been changed to that
1269        when building, a check for int64_t is made, and if it is found, it is used
1270        instead, thus avoiding the use of floating point arithmetic. (There is no
1271        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1272        double.
1273    
1274    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1275        2005 (difference between two addresses compared to an unsigned value).
1276    
1277    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1278        custom one, because of the following reported problem in Windows:
1279    
1280          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1281              under Win32.
1282          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1283              therefore missing the function definition.
1284          - The compiler thus generates a "C" signature for the test function.
1285          - The linker fails to find the "C" function.
1286          - PCRE fails to configure if asked to do so against libbz2.
1287    
1288    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1289        messages were output:
1290    
1291          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1292          rerunning libtoolize, to keep the correct libtool macros in-tree.
1293          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1294    
1295        I have done both of these things.
1296    
1297    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1298        most of the time, it *can* run out if it is given a pattern that contains a
1299        runaway infinite recursion. I updated the discussion in the pcrestack man
1300        page.
1301    
1302    13. Now that we have gone to the x.xx style of version numbers, the minor
1303        version may start with zero. Using 08 or 09 is a bad idea because users
1304        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1305        interpreted as invalid octal numbers. I've updated the previous comment in
1306        configure.ac, and also added a check that gives an error if 08 or 09 are
1307        used.
1308    
1309    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1310        causing partial matching to fail when the end of the subject matched \W
1311        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1312    
1313    15. There were some discrepancies between the declarations in pcre_internal.h
1314        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1315        their definitions. The declarations used "const uschar *" and the
1316        definitions used USPTR. Even though USPTR is normally defined as "const
1317        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1318        reported that: "This difference in casting confuses some C++ compilers, for
1319        example, SunCC recognizes above declarations as different functions and
1320        generates broken code for hbpcre." I have changed the declarations to use
1321        USPTR.
1322    
1323    16. GNU libtool is named differently on some systems. The autogen.sh script now
1324        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1325        (FreeBSD).
1326    
1327    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1328        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1329        comment: "Figure out how to create a longlong from a string: strtoll and
1330        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1331        instance, but it only takes 2 args instead of 3!"
1332    
1333    18. A subtle bug concerned with back references has been fixed by a change of
1334        specification, with a corresponding code fix. A pattern such as
1335        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1336        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1337        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1338        same bug. Such groups have to be quantified to be useful, or contained
1339        inside another quantified group. (If there's no repetition, the reference
1340        can never match.) The problem arises because, having left the group and
1341        moved on to the rest of the pattern, a later failure that backtracks into
1342        the group uses the captured value from the final iteration of the group
1343        rather than the correct earlier one. I have fixed this in PCRE by forcing
1344        any group that contains a reference to itself to be an atomic group; that
1345        is, there cannot be any backtracking into it once it has completed. This is
1346        similar to recursive and subroutine calls.
1347    
1348    
1349    Version 8.00 19-Oct-09
1350    ----------------------
1351    
1352    1.  The table for translating pcre_compile() error codes into POSIX error codes
1353        was out-of-date, and there was no check on the pcre_compile() error code
1354        being within the table. This could lead to an OK return being given in
1355        error.
1356    
1357    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1358        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1359        in a Windows environment.
1360    
1361    3.  The pcregrep --count option prints the count for each file even when it is
1362        zero, as does GNU grep. However, pcregrep was also printing all files when
1363        --files-with-matches was added. Now, when both options are given, it prints
1364        counts only for those files that have at least one match. (GNU grep just
1365        prints the file name in this circumstance, but including the count seems
1366        more useful - otherwise, why use --count?) Also ensured that the
1367        combination -clh just lists non-zero counts, with no names.
1368    
1369    4.  The long form of the pcregrep -F option was incorrectly implemented as
1370        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1371        but it seems right to fix it, and I didn't think it was worth preserving
1372        the old behaviour.
1373    
1374    5.  The command line items --regex=pattern and --regexp=pattern were not
1375        recognized by pcregrep, which required --regex pattern or --regexp pattern
1376        (with a space rather than an '='). The man page documented the '=' forms,
1377        which are compatible with GNU grep; these now work.
1378    
1379    6.  No libpcreposix.pc file was created for pkg-config; there was just
1380        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1381    
1382    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1383        when UCP support is not needed, by modifying the Python script that
1384        generates it from Unicode data files. This should not matter if the module
1385        is correctly used as a library, but I received one complaint about 50K of
1386        unwanted data. My guess is that the person linked everything into his
1387        program rather than using a library. Anyway, it does no harm.
1388    
1389    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1390        was a minimum greater than 1 for a wide character in a possessive
1391        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1392        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1393        character. Chaos in the form of incorrect output or a compiling loop could
1394        result.
1395    
1396    9.  The restrictions on what a pattern can contain when partial matching is
1397        requested for pcre_exec() have been removed. All patterns can now be
1398        partially matched by this function. In addition, if there are at least two
1399        slots in the offset vector, the offset of the earliest inspected character
1400        for the match and the offset of the end of the subject are set in them when
1401        PCRE_ERROR_PARTIAL is returned.
1402    
1403    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1404        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1405        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1406        and may be more useful for multi-segment matching.
1407    
1408    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1409        used to be given if ever the end of the subject was reached; now it is
1410        given only if matching could not proceed because another character was
1411        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1412        string "Z", which now yields "no match" instead of "partial match". In the
1413        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1414        final character ended with (*FAIL).
1415    
1416    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1417        if the pattern had a "must contain" character that was already found in the
1418        earlier partial match, unless partial matching was again requested. For
1419        example, with the pattern /dog.(body)?/, the "must contain" character is
1420        "g". If the first part-match was for the string "dog", restarting with
1421        "sbody" failed. This bug has been fixed.
1422    
1423    13. The string returned by pcre_dfa_exec() after a partial match has been
1424        changed so that it starts at the first inspected character rather than the
1425        first character of the match. This makes a difference only if the pattern
1426        starts with a lookbehind assertion or \b or \B (\K is not supported by
1427        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1428        matching functions compatible, and I think it's the right thing to do.
1429    
1430    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1431        so that the demonstration program is easily available in environments where
1432        PCRE has not been installed from source.
1433    
1434    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1435        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1436        library.
1437    
1438    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1439        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1440        is not the first non-POSIX option to be added. Clearly some people find
1441        these options useful.
1442    
1443    17. If a caller to the POSIX matching function regexec() passes a non-zero
1444        value for nmatch with a NULL value for pmatch, the value of
1445        nmatch is forced to zero.
1446    
1447    18. RunGrepTest did not have a test for the availability of the -u option of
1448        the diff command, as RunTest does. It now checks in the same way as
1449        RunTest, and also checks for the -b option.
1450    
1451    19. If an odd number of negated classes containing just a single character
1452        interposed, within parentheses, between a forward reference to a named
1453        subpattern and the definition of the subpattern, compilation crashed with
1454        an internal error, complaining that it could not find the referenced
1455        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1456        [The bug was that it was starting one character too far in when skipping
1457        over the character class, thus treating the ] as data rather than
1458        terminating the class. This meant it could skip too much.]
1459    
1460    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1461        /g option in pcretest when the pattern contains \K, which makes it possible
1462        to have an empty string match not at the start, even when the pattern is
1463        anchored. Updated pcretest and pcredemo to use this option.
1464    
1465    21. If the maximum number of capturing subpatterns in a recursion was greater
1466        than the maximum at the outer level, the higher number was returned, but
1467        with unset values at the outer level. The correct (outer level) value is
1468        now given.
1469    
1470    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1471        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1472        make it do so. The string so far is captured, making this feature
1473        compatible with Perl.
1474    
1475    23. The tests have been re-organized, adding tests 11 and 12, to make it
1476        possible to check the Perl 5.10 features against Perl 5.10.
1477    
1478    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1479        pattern matches a fixed length string. PCRE did not allow this; now it
1480        does. Neither allows recursion.
1481    
1482    25. I finally figured out how to implement a request to provide the minimum
1483        length of subject string that was needed in order to match a given pattern.
1484        (It was back references and recursion that I had previously got hung up
1485        on.) This code has now been added to pcre_study(); it finds a lower bound
1486        to the length of subject needed. It is not necessarily the greatest lower
1487        bound, but using it to avoid searching strings that are too short does give
1488        some useful speed-ups. The value is available to calling programs via
1489        pcre_fullinfo().
1490    
1491    26. While implementing 25, I discovered to my embarrassment that pcretest had
1492        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1493        study optimizations had never been tested with that matching function.
1494        Oops. What is worse, even when it was passed study data, there was a bug in
1495        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1496        were also very few tests of studied patterns with pcre_dfa_exec().
1497    
1498    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1499        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1500        on the other side of the coin, they are no longer allowed to have different
1501        names, because these cannot be distinguished in PCRE, and this has caused
1502        confusion. (This is a difference from Perl.)
1503    
1504    28. When duplicate subpattern names are present (necessarily with different
1505        numbers, as required by 27 above), and a test is made by name in a
1506        conditional pattern, either for a subpattern having been matched, or for
1507        recursion in such a pattern, all the associated numbered subpatterns are
1508        tested, and the overall condition is true if the condition is true for any
1509        one of them. This is the way Perl works, and is also more like the way
1510        testing by number works.
1511    
1512    
1513    Version 7.9 11-Apr-09
1514    ---------------------
1515    
1516    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1517        (pcretest), all targets were linked against these libraries. This included
1518        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1519        libraries. This caused unwanted dependencies to be created. This problem
1520        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1521        pcretest is linked with readline.
1522    
1523    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1524        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1525        moved outside it again, because FALSE and TRUE are already defined in AIX,
1526        but BOOL is not.
1527    
1528    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1529        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1530    
1531    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1532        hyphens) following file names and line numbers when outputting matching
1533        lines. This is not true; no spaces are inserted. I have also clarified the
1534        wording for the --colour (or --color) option.
1535    
1536    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1537        was not coloured; this is different to GNU grep, so I have changed it to be
1538        the same.
1539    
1540    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1541        each matching line was coloured. Now it goes on to look for further matches
1542        of any of the test patterns, which is the same behaviour as GNU grep.
1543    
1544    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1545        doesn't make sense to accept an empty string match in pcregrep, so I have
1546        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1547        seems to be how GNU grep behaves.
1548    
1549    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1550        start or after a newline", because the conditional assertion was not being
1551        correctly handled. The rule now is that both the assertion and what follows
1552        in the first alternative must satisfy the test.
1553    
1554    9.  If auto-callout was enabled in a pattern with a conditional group whose
1555        condition was an assertion, PCRE could crash during matching, both with
1556        pcre_exec() and pcre_dfa_exec().
1557    
1558    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1559        used for matching.
1560    
1561    11. Unicode property support in character classes was not working for
1562        characters (bytes) greater than 127 when not in UTF-8 mode.
1563    
1564    12. Added the -M command line option to pcretest.
1565    
1566    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1567    
1568    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1569    
1570    16. Added comments and documentation about mis-use of no_arg in the C++
1571        wrapper.
1572    
1573    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1574        from Martin Jerabek that uses macro names for all relevant character and
1575        string constants.
1576    
1577    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1578        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1579        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1580        these, but not everybody uses configure.
1581    
1582    19. A conditional group that had only one branch was not being correctly
1583        recognized as an item that could match an empty string. This meant that an
1584        enclosing group might also not be so recognized, causing infinite looping
1585        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1586        with the subject "ab", where knowledge that the repeated group can match
1587        nothing is needed in order to break the loop.
1588    
1589    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1590        exec(), but without supplying a callout function, matching went wrong.
1591    
1592    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1593        leak if the size of the offset vector was greater than 30. When the vector
1594        is smaller, the saved offsets during recursion go onto a local stack
1595        vector, but for larger vectors malloc() is used. It was failing to free
1596        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1597        error, in fact).
1598    
1599    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1600        heapframe that is used only when UTF-8 support is enabled. This caused no
1601        problem, but was untidy.
1602    
1603    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1604        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1605        included within another project.
1606    
1607    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1608        slightly modified by me:
1609    
1610          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1611              not building pcregrep.
1612    
1613          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1614              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1615    
1616    25. Forward references, both numeric and by name, in patterns that made use of
1617        duplicate group numbers, could behave incorrectly or give incorrect errors,
1618        because when scanning forward to find the reference group, PCRE was not
1619        taking into account the duplicate group numbers. A pattern such as
1620        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1621    
1622    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1623        the feature of a custom pointer more persuasive (as requested by a user).
1624    
1625    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1626        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1627        pre-defined.
1628    
1629    28. Added support for (*UTF8) at the start of a pattern.
1630    
1631    29. Arrange for flags added by the "release type" setting in CMake to be shown
1632        in the configuration summary.
1633    
1634    
1635    Version 7.8 05-Sep-08
1636    ---------------------
1637    
1638    1.  Replaced UCP searching code with optimized version as implemented for Ad
1639        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1640        stage table and inline lookup instead of a function, giving speed ups of 2
1641        to 5 times on some simple patterns that I tested. Permission was given to
1642        distribute the MultiStage2.py script that generates the tables (it's not in
1643        the tarball, but is in the Subversion repository).
1644    
1645    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1646        scripts.
1647    
1648    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1649        a group with a zero qualifier. The result of the study could be incorrect,
1650        or the function might crash, depending on the pattern.
1651    
1652    4.  Caseless matching was not working for non-ASCII characters in back
1653        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1654        It now works when Unicode Property Support is available.
1655    
1656    5.  In pcretest, an escape such as \x{de} in the data was always generating
1657        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1658        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1659        truncation.
1660    
1661    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1662    
1663    7.  Added two (int) casts to pcregrep when printing the difference of two
1664        pointers, in case they are 64-bit values.
1665    
1666    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1667        test 2 if it fails.
1668    
1669    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1670        and a #define of that name to empty if it is not externally set. This is to
1671        allow users of MSVC to set it if necessary.
1672    
1673    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1674        the convenience functions in the pcre_get.c source file.
1675    
1676    11. An option change at the start of a pattern that had top-level alternatives
1677        could cause overwriting and/or a crash. This command provoked a crash in
1678        some environments:
1679    
1680          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1681    
1682        This potential security problem was recorded as CVE-2008-2371.
1683    
1684    12. For a pattern where the match had to start at the beginning or immediately
1685        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1686        pcre_dfa_exec() could read past the end of the passed subject if there was
1687        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1688        pcretest so that it places the subject at the end of its malloc-ed buffer.
1689    
1690    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1691        exec() might read past the end of the data buffer in UTF-8 mode.
1692    
1693    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1694        the data contained the byte 0x85 as part of a UTF-8 character within its
1695        first line. This applied both to normal and DFA matching.
1696    
1697    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1698        /^[^d]*?$/8 failed to match "abc".
1699    
1700    16. Added a missing copyright notice to pcrecpp_internal.h.
1701    
1702    17. Make it more clear in the documentation that values returned from
1703        pcre_exec() in ovector are byte offsets, not character counts.
1704    
1705    18. Tidied a few places to stop certain compilers from issuing warnings.
1706    
1707    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1708        supplied by Stefan Weber. I made a further small update for 7.8 because
1709        there is a change of source arrangements: the pcre_searchfuncs.c module is
1710        replaced by pcre_ucd.c.
1711    
1712    
1713    Version 7.7 07-May-08
1714    ---------------------
1715    
1716    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1717        a string to a long long, pretend we don't even have a long long." This is
1718        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1719    
1720    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1721        pre-7.6 versions, which defined a global no_arg variable instead of putting
1722        it in the RE class. (See also #8 below.)
1723    
1724    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1725        Lopes.
1726    
1727    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1728    
1729        (1) The include/exclude patterns were being applied to the whole pathnames
1730            of files, instead of just to the final components.
1731    
1732        (2) If there was more than one level of directory, the subdirectories were
1733            skipped unless they satisfied the include/exclude conditions. This is
1734            inconsistent with GNU grep (and could even be seen as contrary to the
1735            pcregrep specification - which I improved to make it absolutely clear).
1736            The action now is always to scan all levels of directory, and just
1737            apply the include/exclude patterns to regular files.
1738    
1739    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1740        --exclude_dir in the tests to avoid scanning .svn directories.
1741    
1742    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1743        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1744        doesn't support NULs in patterns.
1745    
1746    7.  Added some missing "const"s to declarations of static tables in
1747        pcre_compile.c and pcre_dfa_exec.c.
1748    
1749    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1750        caused by fix #2  above. (Subsequently also a second patch to fix the
1751        first patch. And a third patch - this was a messy problem.)
1752    
1753    9.  Applied Craig's patch to remove the use of push_back().
1754    
1755    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1756        matching function regexec().
1757    
1758    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1759        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1760        references. PCRE supports relative numbers with this syntax (I don't think
1761        Oniguruma does).
1762    
1763    12. Previously, a group with a zero repeat such as (...){0} was completely
1764        omitted from the compiled regex. However, this means that if the group
1765        was called as a subroutine from elsewhere in the pattern, things went wrong
1766        (an internal error was given). Such groups are now left in the compiled
1767        pattern, with a new opcode that causes them to be skipped at execution
1768        time.
1769    
1770    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1771        to the way PCRE behaves:
1772    
1773        (a) A lone ] character is dis-allowed (Perl treats it as data).
1774    
1775        (b) A back reference to an unmatched subpattern matches an empty string
1776            (Perl fails the current match path).
1777    
1778        (c) A data ] in a character class must be notated as \] because if the
1779            first data character in a class is ], it defines an empty class. (In
1780            Perl it is not possible to have an empty class.) The empty class []
1781            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1782            The negative empty class [^] matches any one character, independently
1783            of the DOTALL setting.
1784    
1785    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1786        non-existent subpattern following a character class starting with ']' and
1787        containing () gave an internal compiling error instead of "reference to
1788        non-existent subpattern". Fortunately, when the pattern did exist, the
1789        compiled code was correct. (When scanning forwards to check for the
1790        existencd of the subpattern, it was treating the data ']' as terminating
1791        the class, so got the count wrong. When actually compiling, the reference
1792        was subsequently set up correctly.)
1793    
1794    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1795        it was being rejected as not supported by pcre_dfa_exec(), even though
1796        other assertions are supported. I have made pcre_dfa_exec() support
1797        (*FAIL).
1798    
1799    16. The implementation of 13c above involved the invention of a new opcode,
1800        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1801        cannot be changed at match time, I realized I could make a small
1802        improvement to matching performance by compiling OP_ALLANY instead of
1803        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1804        on the OP_ANY path.
1805    
1806    17. Compiling pcretest on Windows with readline support failed without the
1807        following two fixes: (1) Make the unistd.h include conditional on
1808        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1809    
1810    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1811        ncurses library to be included for pcretest when ReadLine support is
1812        requested, but also to allow for it to be overridden. This patch came from
1813        Daniel Bergström.
1814    
1815    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1816        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1817        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1818        spotting this.
1819    
1820    
1821    Version 7.6 28-Jan-08
1822    ---------------------
1823    
1824    1.  A character class containing a very large number of characters with
1825        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1826        overflow.
1827    
1828    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1829        HAVE_LONG_LONG is not defined.
1830    
1831    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1832        bring them up to date and include new features. This patch includes:
1833    
1834        - Fixed PH's badly added libz and libbz2 support.
1835        - Fixed a problem with static linking.
1836        - Added pcredemo. [But later removed - see 7 below.]
1837        - Fixed dftables problem and added an option.
1838        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1839            HAVE_LONG_LONG.
1840        - Added readline support for pcretest.
1841        - Added an listing of the option settings after cmake has run.
1842    
1843    4.  A user submitted a patch to Makefile that makes it easy to create
1844        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1845        Makefile.am that cause it to include this special target, without
1846        affecting anything else. Note that the same mingw target plus all
1847        the other distribution libraries and programs are now supported
1848        when configuring with CMake (see 6 below) instead of with
1849        Configure/Make.
1850    
1851    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1852        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1853        exported in the Windows port". It has not yet been confirmed that the patch
1854        solves the problem, but it does no harm.
1855    
1856    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1857        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1858        with CMake, and also correct the comment about stack recursion.
1859    
1860    7.  Remove the automatic building of pcredemo from the ./configure system and
1861        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1862        of a program that users should build themselves after PCRE is installed, so
1863        building it automatically is not really right. What is more, it gave
1864        trouble in some build environments.
1865    
1866    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1867    
1868    
1869    Version 7.5 10-Jan-08
1870    ---------------------
1871    
1872    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1873        values in parens when parsing an RE using the C++ wrapper."
1874    
1875    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1876        Characters greater than 255 were excluded from the class instead of being
1877        included.
1878    
1879    3.  The same bug as (2) above applied to negated POSIX classes such as
1880        [:^space:].
1881    
1882    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1883        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1884        I have changed it.
1885    
1886    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1887        first named subpattern) and a construct such as (?&a) would reference the
1888        first named subpattern whose name started with "a" (in other words, the
1889        length check was missing). Both these problems are fixed. "Subpattern name
1890        expected" is now given for (?&) (a zero-length name), and this patch also
1891        makes it give the same error for \k'' (previously it complained that that
1892        was a reference to a non-existent subpattern).
1893    
1894    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1895        this is right because (?- can be followed by option settings as well as by
1896        digits. I have, however, made the messages clearer.
1897    
1898    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1899        than the number used in the conditional) now cause a compile-time error.
1900        This is actually not compatible with Perl, which accepts such patterns, but
1901        treats the conditional as always being FALSE (as PCRE used to), but it
1902        seems to me that giving a diagnostic is better.
1903    
1904    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1905        and messages.
1906    
1907    9.  Fix two occurrences of "backslash" in comments that should have been
1908        "backspace".
1909    
1910    10. Remove two redundant lines of code that can never be obeyed (their function
1911        was moved elsewhere).
1912    
1913    11. The program that makes PCRE's Unicode character property table had a bug
1914        which caused it to generate incorrect table entries for sequences of
1915        characters that have the same character type, but are in different scripts.
1916        It amalgamated them into a single range, with the script of the first of
1917        them. In other words, some characters were in the wrong script. There were
1918        thirteen such cases, affecting characters in the following ranges:
1919    
1920          U+002b0 - U+002c1
1921          U+0060c - U+0060d
1922          U+0061e - U+00612
1923          U+0064b - U+0065e
1924          U+0074d - U+0076d
1925          U+01800 - U+01805
1926          U+01d00 - U+01d77
1927          U+01d9b - U+01dbf
1928          U+0200b - U+0200f
1929          U+030fc - U+030fe
1930          U+03260 - U+0327f
1931          U+0fb46 - U+0fbb1
1932          U+10450 - U+1049d
1933    
1934    12. The -o option (show only the matching part of a line) for pcregrep was not
1935        compatible with GNU grep in that, if there was more than one match in a
1936        line, it showed only the first of them. It now behaves in the same way as
1937        GNU grep.
1938    
1939    13. If the -o and -v options were combined for pcregrep, it printed a blank
1940        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1941        does the same. The return code can be used to tell if there were any
1942        non-matching lines.
1943    
1944    14. Added --file-offsets and --line-offsets to pcregrep.
1945    
1946    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1947        infinitely looping recursion. The bug was that positive lookaheads were not
1948        being skipped when checking for a possible empty match (negative lookaheads
1949        and both kinds of lookbehind were skipped).
1950    
1951    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1952        inclusion of <windows.h> to before rather than after the definition of
1953        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1954    
1955    17. Specifying a possessive quantifier with a specific limit for a Unicode
1956        character property caused pcre_compile() to compile bad code, which led at
1957        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1958        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1959        caused the error; without that there was no problem.
1960    
1961    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1962    
1963    19. Added --enable-pcretest-libreadline.
1964    
1965    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1966        RE::GlobalReplace(). As a result, the number of replacements returned was
1967        double what it should be. I removed one of the increments, but Craig sent a
1968        later patch that removed the other one (the right fix) and added unit tests
1969        that check the return values (which was not done before).
1970    
1971    21. Several CMake things:
1972    
1973        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1974            the names libpcre and libpcreposix, not just pcre and pcreposix.
1975    
1976        (2) The above change means that pcretest and pcregrep are now correctly
1977            linked with the newly-built libraries, not previously installed ones.
1978    
1979        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1980    
1981    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1982        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1983        UTF-8 newline character). The key issue is that the pattern starts .*;
1984        this means that the match must be either at the beginning, or after a
1985        newline. The bug was in the code for advancing after a failed match and
1986        checking that the new position followed a newline. It was not taking
1987        account of UTF-8 characters correctly.
1988    
1989    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1990        character classes. PCRE was not treating the sequence [:...:] as a
1991        character class unless the ... were all letters. Perl, however, seems to
1992        allow any characters between [: and :], though of course it rejects as
1993        unknown any "names" that contain non-letters, because all the known class
1994        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1995        for example, whereas PCRE did not - it did not recognize a POSIX character
1996        class. This seemed a bit dangerous, so the code has been changed to be
1997        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1998        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1999        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2000        Perl does, and where it didn't before.
2001    
2002    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2003        Windows environments %n is disabled by default.
2004    
2005    
2006    Version 7.4 21-Sep-07
2007    ---------------------
2008    
2009    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2010        means that a class such as [\s] counted as "explicit reference to CR or
2011        LF". That isn't really right - the whole point of the change was to try to
2012        help when there was an actual mention of one of the two characters. So now
2013        the change happens only if \r or \n (or a literal CR or LF) character is
2014        encountered.
2015    
2016    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2017        of both had grown to the point where there were only 3 bits left.
2018        Fortunately, there was spare space in the data structure, and so I have
2019        moved the internal flags into a new 16-bit field to free up more option
2020        bits.
2021    
2022    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2023        but did not set the internal JCHANGED flag - either of these is enough to
2024        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2025        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2026        start sets both bits.
2027    
2028    4.  Added options (at build time, compile time, exec time) to change \R from
2029        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2030    
2031    5.  doc/pcresyntax.html was missing from the distribution.
2032    
2033    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2034        compatibility, even though it is no longer used.
2035    
2036    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2037        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2038        windows.h file is present (where different names are used). [This was
2039        reversed later after testing - see 16 below.]
2040    
2041    8.  Changed all #include <config.h> to #include "config.h". There were also
2042        some further <pcre.h> cases that I changed to "pcre.h".
2043    
2044    9.  When pcregrep was used with the --colour option, it missed the line ending
2045        sequence off the lines that it output.
2046    
2047    10. It was pointed out to me that arrays of string pointers cause lots of
2048        relocations when a shared library is dynamically loaded. A technique of
2049        using a single long string with a table of offsets can drastically reduce
2050        these. I have refactored PCRE in four places to do this. The result is
2051        dramatic:
2052    
2053          Originally:                          290
2054          After changing UCP table:            187
2055          After changing error message table:   43
2056          After changing table of "verbs"       36
2057          After changing table of Posix names   22
2058    
2059        Thanks to the folks working on Gregex for glib for this insight.
2060    
2061    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2062        unicode-properties was also set.
2063    
2064    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2065    
2066    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2067        checked only for CRLF.
2068    
2069    14. Added casts to pcretest.c to avoid compiler warnings.
2070    
2071    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2072    
2073    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2074        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2075        entirely. This removes changes made in 7 above.
2076    
2077    17. The CMake files have been updated, and there is now more information about
2078        building with CMake in the NON-UNIX-USE document.
2079    
2080    
2081    Version 7.3 28-Aug-07
2082  ---------------------  ---------------------
2083    
2084   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 9  Version 7.3 05-Jul-07 Line 2086  Version 7.3 05-Jul-07
2086      brackets there is not right, since it causes compilers to look for an      brackets there is not right, since it causes compilers to look for an
2087      installed pcre.h, not the version that is in the source that is being      installed pcre.h, not the version that is in the source that is being
2088      compiled (which of course may be different). I have changed it back to:      compiled (which of course may be different). I have changed it back to:
2089    
2090        #include "pcre.h"        #include "pcre.h"
2091    
2092      I have a vague recollection that the change was concerned with compiling in      I have a vague recollection that the change was concerned with compiling in
2093      different directories, but in the new build system, that is taken care of      different directories, but in the new build system, that is taken care of
2094      by the VPATH setting the Makefile.      by the VPATH setting the Makefile.
2095    
2096   2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed   2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2097      when the subject happened to end in the byte 0x85 (e.g. if the last      when the subject happened to end in the byte 0x85 (e.g. if the last
2098      character was \x{1ec5}). *Character* 0x85 is one of the "any" newline      character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
# Line 23  Version 7.3 05-Jul-07 Line 2100  Version 7.3 05-Jul-07
2100      of another character. The bug was that, for an unlimited repeat of . in      of another character. The bug was that, for an unlimited repeat of . in
2101      not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by      not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2102      characters when looking for a newline.      characters when looking for a newline.
   
  3. A small performance improvement in the DOTALL UTF-8 mode .* case.  
   
  4. Debugging: adjusted the names of opcodes for different kinds of parentheses  
     in debug output.  
   
  5. Arrange to use "%I64d" instead of "%lld" for long printing in the pcrecpp  
     unittest when running under MinGW.  
2103    
2104     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2105    
2106     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2107        in debug output.
2108    
2109     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2110        long printing in the pcrecpp unittest when running under MinGW.
2111    
2112     6. ESC_K was left out of the EBCDIC table.
2113    
2114     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2115        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2116        limit also applies to "virtual nesting" when a pattern is recursive, and in
2117        this case 1000 isn't so big. I have been able to remove this limit at the
2118        expense of backing off one optimization in certain circumstances. Normally,
2119        when pcre_exec() would call its internal match() function recursively and
2120        immediately return the result unconditionally, it uses a "tail recursion"
2121        feature to save stack. However, when a subpattern that can match an empty
2122        string has an unlimited repetition quantifier, it no longer makes this
2123        optimization. That gives it a stack frame in which to save the data for
2124        checking that an empty string has been matched. Previously this was taken
2125        from the 1000-entry workspace that had been reserved. So now there is no
2126        explicit limit, but more stack is used.
2127    
2128     8. Applied Daniel's patches to solve problems with the import/export magic
2129        syntax that is required for Windows, and which was going wrong for the
2130        pcreposix and pcrecpp parts of the library. These were overlooked when this
2131        problem was solved for the main library.
2132    
2133     9. There were some crude static tests to avoid integer overflow when computing
2134        the size of patterns that contain repeated groups with explicit upper
2135        limits. As the maximum quantifier is 65535, the maximum group length was
2136        set at 30,000 so that the product of these two numbers did not overflow a
2137        32-bit integer. However, it turns out that people want to use groups that
2138        are longer than 30,000 bytes (though not repeat them that many times).
2139        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2140        made it possible to implement the integer overflow checks in a much more
2141        dynamic way, which I have now done. The artificial limitation on group
2142        length has been removed - we now have only the limit on the total length of
2143        the compiled pattern, which depends on the LINK_SIZE setting.
2144    
2145    10. Fixed a bug in the documentation for get/copy named substring when
2146        duplicate names are permitted. If none of the named substrings are set, the
2147        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2148        empty string.
2149    
2150    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2151        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2152        because the ] is interpreted as the first data character and the
2153        terminating ] is not found. PCRE has been made compatible with Perl in this
2154        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2155        cause memory overwriting.
2156    
2157    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2158        string has been matched (to stop an infinite loop). It was not recognizing
2159        a conditional subpattern that could match an empty string if that
2160        subpattern was within another subpattern. For example, it looped when
2161        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2162        condition was not nested. This bug has been fixed.
2163    
2164    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2165        past the start of the subject in the presence of bytes with the top bit
2166        set, for example "\x8aBCD".
2167    
2168    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2169        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2170    
2171    14. Optimized (?!) to (*FAIL).
2172    
2173    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2174        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2175        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2176        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2177        does: it's just the validity check that is more restrictive.
2178    
2179    16. Inserted checks for integer overflows during escape sequence (backslash)
2180        processing, and also fixed erroneous offset values for syntax errors during
2181        backslash processing.
2182    
2183    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2184        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2185    
2186    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2187        caused an overrun.
2188    
2189    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2190        something other than just ASCII characters) inside a group that had an
2191        unlimited repeat caused a loop at compile time (while checking to see
2192        whether the group could match an empty string).
2193    
2194    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2195        [\P{Any}] did so. (Error in the code for printing property names.)
2196    
2197    21. An orphan \E inside a character class could cause a crash.
2198    
2199    22. A repeated capturing bracket such as (A)? could cause a wild memory
2200        reference during compilation.
2201    
2202    23. There are several functions in pcre_compile() that scan along a compiled
2203        expression for various reasons (e.g. to see if it's fixed length for look
2204        behind). There were bugs in these functions when a repeated \p or \P was
2205        present in the pattern. These operators have additional parameters compared
2206        with \d, etc, and these were not being taken into account when moving along
2207        the compiled data. Specifically:
2208    
2209        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2210            length.
2211    
2212        (b) An item such as \pL+ within a repeated group could cause crashes or
2213            loops.
2214    
2215        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2216            "reference to non-existent subpattern" error.
2217    
2218        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2219    
2220    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2221        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2222    
2223    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2224    
2225    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2226        character were causing crashes (broken optimization).
2227    
2228    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2229        \p or \P) caused a compile-time loop.
2230    
2231    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2232        break. For example, the unstudied pattern [\r\n]A does not match the string
2233        "\r\nA" because change 7.0/46 below moves the current point on by two
2234        characters after failing to match at the start. However, the pattern \nA
2235        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2236        the same is true. There doesn't seem any very clean way out of this, but
2237        what I have chosen to do makes the common cases work: PCRE now takes note
2238        of whether there can be an explicit match for \r or \n anywhere in the
2239        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2240        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2241        pattern has explicit CR or LF references.
2242    
2243    29. Added (*CR) etc for changing newline setting at start of pattern.
2244    
2245    
2246  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07

Legend:
Removed from v.193  
changed lines
  Added in v.1244

  ViewVC Help
Powered by ViewVC 1.1.5