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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.195  
changed lines
  Added in v.1257

  ViewVC Help
Powered by ViewVC 1.1.5