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

Diff of /code/trunk/ChangeLog

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

revision 164 by ph10, Fri May 4 15:11:14 2007 UTC revision 1248 by ph10, Wed Feb 13 17:36:38 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Version 8.33 xx-xxxx-201x
5    -------------------------
6    
7    1.  Added 'U' to some constants that are compared to unsigned integers, to
8        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9        variables that are added to signed variables, to ensure the result is
10        signed and can be negated.
11    
12    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
13        CMake config files.
14    
15    3.  Revise the creation of config.h.generic so that all boolean macros are
16        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17        overriding via -D on the command line possible.
18    
19    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20        to unsigned int is reported to make a quite noticeable speed difference in
21        a specific Windows environment. Testing on Linux did also appear to show
22        some benefit (and it is clearly not harmful). Also fixed the definition of
23        Xop which should be unsigned.
24    
25    5.  Related to (4), changing the definition of the intermediate variable cc
26        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27        improvements.
28    
29    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
30        unnecessary spaces.
31    
32    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33        and later.
34    
35    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36    
37    9.  Optimizing fast_forward_start_bits in JIT.
38    
39    10. Adding experimental support for callouts in JIT, and fixing some
40        issues revealed during this work. Namely:
41    
42        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43    
44        (b) Minimum length was not checked before the matching is started.
45    
46        WARNING: Callouts are not release ready! <- this line needs to be removed when it is.
47    
48    11. The value of capture_last that is passed to callouts was incorrect in some
49        cases when there was a capture on one path that was subsequently abandoned
50        after a backtrack. Also, the capture_last value is now reset after a
51        recursion, since all captures are also reset in this case.
52    
53    12. The interpreter no longer returns the "too many substrings" error in the
54        case when an overflowing capture is in a branch that is subsequently
55        abandoned after a backtrack.
56    
57    13. In the pathological case when an offset vector of size 2 is used, pcretest
58        now prints out the matched string after a yield of 0 or 1.
59    
60    
61    Version 8.32 30-November-2012
62    -----------------------------
63    
64    1.  Improved JIT compiler optimizations for first character search and single
65        character iterators.
66    
67    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
68        Patch by Daniel Richard G.
69    
70    3.  Single character iterator optimizations in the JIT compiler.
71    
72    4.  Improved JIT compiler optimizations for character ranges.
73    
74    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
75        Reported by Giuseppe D'Angelo.
76    
77    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
78        of a line, was being set incorrectly in cases where .* appeared inside
79        atomic brackets at the start of a pattern, or where there was a subsequent
80        *PRUNE or *SKIP.
81    
82    7.  Improved instruction cache flush for POWER/PowerPC.
83        Patch by Daniel Richard G.
84    
85    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
86        grep:
87    
88        (a) There is now no limit to the number of patterns to be matched.
89    
90        (b) An error is given if a pattern is too long.
91    
92        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
93            are now supported.
94    
95        (d) --exclude-from and --include-from (multiple use) have been added.
96    
97        (e) Exclusions and inclusions now apply to all files and directories, not
98            just to those obtained from scanning a directory recursively.
99    
100        (f) Multiple uses of -f and --file-list are now supported.
101    
102        (g) In a Windows environment, the default for -d has been changed from
103            "read" (the GNU grep default) to "skip", because otherwise the presence
104            of a directory in the file list provokes an error.
105    
106        (h) The documentation has been revised and clarified in places.
107    
108    9.  Improve the matching speed of capturing brackets.
109    
110    10. Changed the meaning of \X so that it now matches a Unicode extended
111        grapheme cluster.
112    
113    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
114        out POSIX threads when JIT support is configured.
115    
116    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
117    
118    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
119        structure could go wrong in environments where size_t is not the same size
120        as int.
121    
122    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
123        set.
124    
125    15. The EBCDIC support had decayed; later updates to the code had included
126        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
127        general tidy up of EBCDIC-related issues, and the documentation was also
128        not quite right. There is now a test that can be run on ASCII systems to
129        check some of the EBCDIC-related things (but is it not a full test).
130    
131    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
132        in a small tidy to the code.
133    
134    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
135    
136    18. If the --only-matching (-o) option in pcregrep is specified multiple
137        times, each one causes appropriate output. For example, -o1 -o2 outputs the
138        substrings matched by the 1st and 2nd capturing parentheses. A separating
139        string can be specified by --om-separator (default empty).
140    
141    19. Improving the first n character searches.
142    
143    20. Turn case lists for horizontal and vertical white space into macros so that
144        they are defined only once.
145    
146    21. This set of changes together give more compatible Unicode case-folding
147        behaviour for characters that have more than one other case when UCP
148        support is available.
149    
150        (a) The Unicode property table now has offsets into a new table of sets of
151            three or more characters that are case-equivalent. The MultiStage2.py
152            script that generates these tables (the pcre_ucd.c file) now scans
153            CaseFolding.txt instead of UnicodeData.txt for character case
154            information.
155    
156        (b) The code for adding characters or ranges of characters to a character
157            class has been abstracted into a generalized function that also handles
158            case-independence. In UTF-mode with UCP support, this uses the new data
159            to handle characters with more than one other case.
160    
161        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
162            whose other case is greater than 256 are now correctly matched
163            caselessly. Previously, the high codepoint matched the low one, but not
164            vice versa.
165    
166        (d) The processing of \h, \H, \v, and \ in character classes now makes use
167            of the new class addition function, using character lists defined as
168            macros alongside the case definitions of 20 above.
169    
170        (e) Caseless back references now work with characters that have more than
171            one other case.
172    
173        (f) General caseless matching of characters with more than one other case
174            is supported.
175    
176    22. Unicode character properties were updated from Unicode 6.2.0
177    
178    23. Improved CMake support under Windows. Patch by Daniel Richard G.
179    
180    24. Add support for 32-bit character strings, and UTF-32
181    
182    25. Major JIT compiler update (code refactoring and bugfixing).
183        Experimental Sparc 32 support is added.
184    
185    26. Applied a modified version of Daniel Richard G's patch to create
186        pcre.h.generic and config.h.generic by "make" instead of in the
187        PrepareRelease script.
188    
189    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
190        pcre_compile.c when checking for a zero character.
191    
192    28. Introducing a native interface for JIT. Through this interface, the compiled
193        machine code can be directly executed. The purpose of this interface is to
194        provide fast pattern matching, so several sanity checks are not performed.
195        However, feature tests are still performed. The new interface provides
196        1.4x speedup compared to the old one.
197    
198    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
199        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
200        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
201    
202    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
203        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
204        the "old" RFC 2279). Instead, it ended up passing a negative length to
205        pcre_exec().
206    
207    31. Add support for GCC's visibility feature to hide internal functions.
208    
209    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
210        "unknown -C option" after outputting 0 or 1.
211    
212    33. There is now support for generating a code coverage report for the test
213        suite in environments where gcc is the compiler and lcov is installed. This
214        is mainly for the benefit of the developers.
215    
216    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
217        unaddressable using valgrind annotations, allowing valgrind to detect
218        invalid memory accesses. This is mainly for the benefit of the developers.
219    
220    25. (*UTF) can now be used to start a pattern in any of the three libraries.
221    
222    26. Give configure error if --enable-cpp but no C++ compiler found.
223    
224    
225    Version 8.31 06-July-2012
226    -------------------------
227    
228    1.  Fixing a wrong JIT test case and some compiler warnings.
229    
230    2.  Removed a bashism from the RunTest script.
231    
232    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
233        to unsigned type, result still unsigned" that was given by an MS compiler
234        on encountering the code "-sizeof(xxx)".
235    
236    4.  Partial matching support is added to the JIT compiler.
237    
238    5.  Fixed several bugs concerned with partial matching of items that consist
239        of more than one character:
240    
241        (a) /^(..)\1/ did not partially match "aba" because checking references was
242            done on an "all or nothing" basis. This also applied to repeated
243            references.
244    
245        (b) \R did not give a hard partial match if \r was found at the end of the
246            subject.
247    
248        (c) \X did not give a hard partial match after matching one or more
249            characters at the end of the subject.
250    
251        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
252            a partial match for the string "\r".
253    
254        (e) When newline was set to CRLF, the metacharacter "." did not recognize
255            a partial match for a CR character at the end of the subject string.
256    
257    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
258        running pcretest, the text "(JIT)" added to the output whenever JIT is
259        actually used to run the match.
260    
261    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
262        or /S+[+] with a digit between 1 and 7.
263    
264    8.  OP_NOT now supports any UTF character not just single-byte ones.
265    
266    9.  (*MARK) control verb is now supported by the JIT compiler.
267    
268    10. The command "./RunTest list" lists the available tests without actually
269        running any of them. (Because I keep forgetting what they all are.)
270    
271    11. Add PCRE_INFO_MAXLOOKBEHIND.
272    
273    12. Applied a (slightly modified) user-supplied patch that improves performance
274        when the heap is used for recursion (compiled with --disable-stack-for-
275        recursion). Instead of malloc and free for each heap frame each time a
276        logical recursion happens, frames are retained on a chain and re-used where
277        possible. This sometimes gives as much as 30% improvement.
278    
279    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
280        call.
281    
282    14. As documented, (*COMMIT) is now confined to within a positive assertion.
283    
284    15. It is now possible to link pcretest with libedit as an alternative to
285        libreadline.
286    
287    16. (*COMMIT) control verb is now supported by the JIT compiler.
288    
289    17. The Unicode data tables have been updated to Unicode 6.1.0.
290    
291    18. Added --file-list option to pcregrep.
292    
293    19. Added binary file support to pcregrep, including the -a, --binary-files,
294        -I, and --text options.
295    
296    20. The madvise function is renamed for posix_madvise for QNX compatibility
297        reasons. Fixed by Giuseppe D'Angelo.
298    
299    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
300    
301    22. Changed the option for creating symbolic links for 16-bit man pages from
302        -s to -sf so that re-installing does not cause issues.
303    
304    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
305    
306    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
307        matches in certain environments (the workspace was not being correctly
308        retained). Also added to pcre_dfa_exec() a simple plausibility check on
309        some of the workspace data at the beginning of a restart.
310    
311    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
312        was not doing so when it should - probably a typo introduced by SVN 528
313        (change 8.10/14).
314    
315    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
316        \w+ when the character tables indicated that \x{c4} was a word character.
317        There were several related cases, all because the tests for doing a table
318        lookup were testing for characters less than 127 instead of 255.
319    
320    27. If a pattern contains capturing parentheses that are not used in a match,
321        their slots in the ovector are set to -1. For those that are higher than
322        any matched groups, this happens at the end of processing. In the case when
323        there were back references that the ovector was too small to contain
324        (causing temporary malloc'd memory to be used during matching), and the
325        highest capturing number was not used, memory off the end of the ovector
326        was incorrectly being set to -1. (It was using the size of the temporary
327        memory instead of the true size.)
328    
329    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
330        ovector size, it uses memory at the end of the block that it has got.
331    
332    29. Check for an overlong MARK name and give an error at compile time. The
333        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
334    
335    30. JIT compiler update.
336    
337    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
338        Rill for the patch.
339    
340    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
341    
342    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
343    
344    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
345        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
346        was enabled.
347    
348    35. Improve JIT code generation for greedy plus quantifier.
349    
350    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
351        1 to "aa" instead of to an empty string. The bug affected repeated groups
352        that could potentially match an empty string.
353    
354    37. Optimizing single character iterators in JIT.
355    
356    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
357        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
358        codepoints that are too big for the mode are faulted, and in a UTF mode,
359        disallowed codepoints are also faulted.
360    
361    39. If PCRE was compiled with UTF support, in three places in the DFA
362        matcher there was code that should only have been obeyed in UTF mode, but
363        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
364        processing when bytes with values greater than 127 were present. In 16-bit
365        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
366        both cases the values are those that cannot be the first data item in a UTF
367        character. The three items that might have provoked this were recursions,
368        possessively repeated groups, and atomic groups.
369    
370    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
371        and pcregrep, because some OS require shared objects to be explicitly
372        passed to ld, causing the link step to fail if they are not.
373    
374    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
375        mode, patterns that started with \h* or \R* might be incorrectly matched.
376    
377    
378    Version 8.30 04-February-2012
379    -----------------------------
380    
381    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
382        name is defined in ctype.h.
383    
384    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
385        only in quite long subpatterns.
386    
387    3.  Removed the function pcre_info(), which has been obsolete and deprecated
388        since it was replaced by pcre_fullinfo() in February 2000.
389    
390    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
391        match a (*MARK), and the match failed at the start of the subject, a
392        reference to memory before the start of the subject could occur. This bug
393        was introduced by fix 17 of release 8.21.
394    
395    5.  A reference to an unset group with zero minimum repetition was giving
396        totally wrong answers (in non-JavaScript-compatibility mode). For example,
397        /(another)?(\1?)test/ matched against "hello world test". This bug was
398        introduced in release 8.13.
399    
400    6.  Add support for 16-bit character strings (a large amount of work involving
401        many changes and refactorings).
402    
403    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
404        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
405        from a file.
406    
407    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
408        rounding is not applied in this particular case).
409    
410    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
411        if they appear, or are escaped, in patterns.
412    
413    10. Get rid of a number of -Wunused-but-set-variable warnings.
414    
415    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
416        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
417        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
418        also returns the mark "x". This bug applied to capturing parentheses,
419        non-capturing parentheses, and atomic parentheses. It also applied to some
420        assertions.
421    
422    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
423        information out of configure.ac instead of relying on pcre.h.generic, which
424        is not stored in the repository.
425    
426    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
427        -lreadline.
428    
429    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
430    
431    15. Applied Graycode's patch to put the top-level frame on the stack rather
432        than the heap when not using the stack for recursion. This gives a
433        performance improvement in many cases when recursion is not deep.
434    
435    16. Experimental code added to "pcretest -C" to output the stack frame size.
436    
437    
438    Version 8.21 12-Dec-2011
439    ------------------------
440    
441    1.  Updating the JIT compiler.
442    
443    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
444        are added as well.
445    
446    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
447        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
448        calling _pcre_jit_exec. Some extra comments are added.
449    
450    4.  (*MARK) settings inside atomic groups that do not contain any capturing
451        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
452        was introduced by change 18 for 8.20.
453    
454    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
455        ECMA-262 standard.
456    
457    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
458        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
459        This bug was probably introduced by change 9 of 8.13.
460    
461    7.  While fixing 6 above, I noticed that a number of other items were being
462        incorrectly rejected as "not fixed length". This arose partly because newer
463        opcodes had not been added to the fixed-length checking code. I have (a)
464        corrected the bug and added tests for these items, and (b) arranged for an
465        error to occur if an unknown opcode is encountered while checking for fixed
466        length instead of just assuming "not fixed length". The items that were
467        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
468        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
469        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
470    
471    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
472        being incorrectly compiled and would have given unpredicatble results.
473    
474    9.  A possessively repeated subpattern with minimum repeat count greater than
475        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
476        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
477        the first (A) could occur when it should not.
478    
479    10. Add a cast and remove a redundant test from the code.
480    
481    11. JIT should use pcre_malloc/pcre_free for allocation.
482    
483    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
484        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
485        is anything other than /usr, -L is still shown).
486    
487    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
488    
489    14. Perl does not support \N without a following name in a [] class; PCRE now
490        also gives an error.
491    
492    15. If a forward reference was repeated with an upper limit of around 2000,
493        it caused the error "internal error: overran compiling workspace". The
494        maximum number of forward references (including repeats) was limited by the
495        internal workspace, and dependent on the LINK_SIZE. The code has been
496        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
497        the default depends on LINK_SIZE. There is a new upper limit (for safety)
498        of around 200,000 forward references. While doing this, I also speeded up
499        the filling in of repeated forward references.
500    
501    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
502        incorrectly expecting the subject to contain another "a" after the start.
503    
504    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
505        in the match, the SKIP should be ignored. This was not happening; instead
506        the SKIP was being treated as NOMATCH. For patterns such as
507        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
508        tested.
509    
510    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
511        now much more compatible with Perl, in particular in cases where the result
512        is a non-match for a non-anchored pattern. For example, if
513        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
514        "m", where previously it did not return a name. A side effect of this
515        change is that for partial matches, the last encountered mark name is
516        returned, as for non matches. A number of tests that were previously not
517        Perl-compatible have been moved into the Perl-compatible test files. The
518        refactoring has had the pleasing side effect of removing one argument from
519        the match() function, thus reducing its stack requirements.
520    
521    19. If the /S+ option was used in pcretest to study a pattern using JIT,
522        subsequent uses of /S (without +) incorrectly behaved like /S+.
523    
524    21. Retrieve executable code size support for the JIT compiler and fixing
525        some warnings.
526    
527    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
528        not work when the shorter character appeared right at the end of the
529        subject string.
530    
531    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
532        systems.
533    
534    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
535        output it when the /M option is used in pcretest.
536    
537    25. The CheckMan script was not being included in the distribution. Also, added
538        an explicit "perl" to run Perl scripts from the PrepareRelease script
539        because this is reportedly needed in Windows.
540    
541    26. If study data was being save in a file and studying had not found a set of
542        "starts with" bytes for the pattern, the data written to the file (though
543        never used) was taken from uninitialized memory and so caused valgrind to
544        complain.
545    
546    27. Updated RunTest.bat as provided by Sheri Pierce.
547    
548    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
549    
550    29. Computation of memory usage for the table of capturing group names was
551        giving an unnecessarily large value.
552    
553    
554    Version 8.20 21-Oct-2011
555    ------------------------
556    
557    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
558        a POSIX class. After further experiments with Perl, which convinced me that
559        Perl has bugs and confusions, a closing square bracket is no longer allowed
560        in a POSIX name. This bug also affected patterns with classes that started
561        with full stops.
562    
563    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
564        captured substring, but while checking the failing first alternative,
565        substring 1 is temporarily captured. If the output vector supplied to
566        pcre_exec() was not big enough for this capture, the yield of the function
567        was still zero ("insufficient space for captured substrings"). This cannot
568        be totally fixed without adding another stack variable, which seems a lot
569        of expense for a edge case. However, I have improved the situation in cases
570        such as /(a)(b)x|abc/ matched against "abc", where the return code
571        indicates that fewer than the maximum number of slots in the ovector have
572        been set.
573    
574    3.  Related to (2) above: when there are more back references in a pattern than
575        slots in the output vector, pcre_exec() uses temporary memory during
576        matching, and copies in the captures as far as possible afterwards. It was
577        using the entire output vector, but this conflicts with the specification
578        that only 2/3 is used for passing back captured substrings. Now it uses
579        only the first 2/3, for compatibility. This is, of course, another edge
580        case.
581    
582    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
583        main code base, and can be used by building with --enable-jit. When this is
584        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
585        runtime --no-jit option is given.
586    
587    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
588        ovector, the return from the function was zero, implying that there were
589        other matches that did not fit. The correct "exactly full" value is now
590        returned.
591    
592    6.  If a subpattern that was called recursively or as a subroutine contained
593        (*PRUNE) or any other control that caused it to give a non-standard return,
594        invalid errors such as "Error -26 (nested recursion at the same subject
595        position)" or even infinite loops could occur.
596    
597    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
598        computing the minimum length on reaching *ACCEPT, and so ended up with the
599        wrong value of 1 rather than 0. Further investigation indicates that
600        computing a minimum subject length in the presence of *ACCEPT is difficult
601        (think back references, subroutine calls), and so I have changed the code
602        so that no minimum is registered for a pattern that contains *ACCEPT.
603    
604    8.  If (*THEN) was present in the first (true) branch of a conditional group,
605        it was not handled as intended. [But see 16 below.]
606    
607    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
608        Sheri Pierce.
609    
610    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
611        the first byte in a match must be "a".
612    
613    11. Change 17 for 8.13 increased the recursion depth for patterns like
614        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
615        pattern contains any instances of (*THEN). If it does not, the old
616        optimizations are restored. It would be nice to do this on a per-group
617        basis, but at the moment that is not feasible.
618    
619    12. In some environments, the output of pcretest -C is CRLF terminated. This
620        broke RunTest's code that checks for the link size. A single white space
621        character after the value is now allowed for.
622    
623    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
624        For "fr", it uses the Windows-specific input and output files.
625    
626    14. If (*THEN) appeared in a group that was called recursively or as a
627        subroutine, it did not work as intended. [But see next item.]
628    
629    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
630        pattern fragments (but not containing any | characters). If A and B are
631        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
632        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
633        D. In other words, Perl considers parentheses that do not contain any |
634        characters to be part of a surrounding alternative, whereas PCRE was
635        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
636        differently. PCRE now behaves in the same way as Perl, except in the case
637        of subroutine/recursion calls such as (?1) which have in any case always
638        been different (but PCRE had them first :-).
639    
640    16. Related to 15 above: Perl does not treat the | in a conditional group as
641        creating alternatives. Such a group is treated in the same way as an
642        ordinary group without any | characters when processing (*THEN). PCRE has
643        been changed to match Perl's behaviour.
644    
645    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
646        RunGrepTest script failed.
647    
648    18. Change 22 for version 13 caused atomic groups to use more stack. This is
649        inevitable for groups that contain captures, but it can lead to a lot of
650        stack use in large patterns. The old behaviour has been restored for atomic
651        groups that do not contain any capturing parentheses.
652    
653    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
654        suppress the check for a minimum subject length at run time. (If it was
655        given to pcre_exec() or pcre_dfa_exec() it did work.)
656    
657    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
658        fail to work when decoding hex characters in data strings in EBCDIC
659        environments.
660    
661    21. It appears that in at least one Mac OS environment, the isxdigit() function
662        is implemented as a macro that evaluates to its argument more than once,
663        contravening the C 90 Standard (I haven't checked a later standard). There
664        was an instance in pcretest which caused it to go wrong when processing
665        \x{...} escapes in subject strings. The has been rewritten to avoid using
666        things like p++ in the argument of isxdigit().
667    
668    
669    Version 8.13 16-Aug-2011
670    ------------------------
671    
672    1.  The Unicode data tables have been updated to Unicode 6.0.0.
673    
674    2.  Two minor typos in pcre_internal.h have been fixed.
675    
676    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
677        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
678        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
679    
680    4.  There were a number of related bugs in the code for matching backrefences
681        caselessly in UTF-8 mode when codes for the characters concerned were
682        different numbers of bytes. For example, U+023A and U+2C65 are an upper
683        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
684        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
685        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
686        2-byte code at the end of the subject (it thought there wasn't enough data
687        left).
688    
689    5.  Comprehensive information about what went wrong is now returned by
690        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
691        as the output vector has at least 2 elements. The offset of the start of
692        the failing character and a reason code are placed in the vector.
693    
694    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
695        now returned is for the first byte of the failing character, instead of the
696        last byte inspected. This is an incompatible change, but I hope it is small
697        enough not to be a problem. It makes the returned offset consistent with
698        pcre_exec() and pcre_dfa_exec().
699    
700    7.  pcretest now gives a text phrase as well as the error number when
701        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
702        failure, the offset and reason code are output.
703    
704    8.  When \R was used with a maximizing quantifier it failed to skip backwards
705        over a \r\n pair if the subsequent match failed. Instead, it just skipped
706        back over a single character (\n). This seems wrong (because it treated the
707        two characters as a single entity when going forwards), conflicts with the
708        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
709        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
710        has been changed.
711    
712    9.  Some internal refactoring has changed the processing so that the handling
713        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
714        time (the PCRE_DOTALL option was changed this way some time ago: version
715        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
716        which was always a bit of a fudge. It also means that there is one less
717        argument for the match() function, which reduces its stack requirements
718        slightly. This change also fixes an incompatibility with Perl: the pattern
719        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
720    
721    10. More internal refactoring has drastically reduced the number of recursive
722        calls to match() for possessively repeated groups such as (abc)++ when
723        using pcre_exec().
724    
725    11. While implementing 10, a number of bugs in the handling of groups were
726        discovered and fixed:
727    
728        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
729        (a|)*(?1) gave a compile-time internal error.
730        ((a|)+)+  did not notice that the outer group could match an empty string.
731        (^a|^)+   was not marked as anchored.
732        (.*a|.*)+ was not marked as matching at start or after a newline.
733    
734    12. Yet more internal refactoring has removed another argument from the match()
735        function. Special calls to this function are now indicated by setting a
736        value in a variable in the "match data" data block.
737    
738    13. Be more explicit in pcre_study() instead of relying on "default" for
739        opcodes that mean there is no starting character; this means that when new
740        ones are added and accidentally left out of pcre_study(), testing should
741        pick them up.
742    
743    14. The -s option of pcretest has been documented for ages as being an old
744        synonym of -m (show memory usage). I have changed it to mean "force study
745        for every regex", that is, assume /S for every regex. This is similar to -i
746        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
747        using it. It makes it easier to run collections of tests with and without
748        study enabled, and thereby test pcre_study() more easily. All the standard
749        tests are now run with and without -s (but some patterns can be marked as
750        "never study" - see 20 below).
751    
752    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
753        restoration of the capturing data to the outer values was not happening
754        correctly.
755    
756    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
757        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
758        pattern had matched an empty string, and so incorrectly returned a no
759        match.
760    
761    17. There was optimizing code for the last branch of non-capturing parentheses,
762        and also for the obeyed branch of a conditional subexpression, which used
763        tail recursion to cut down on stack usage. Unfortunately, now that there is
764        the possibility of (*THEN) occurring in these branches, tail recursion is
765        no longer possible because the return has to be checked for (*THEN). These
766        two optimizations have therefore been removed. [But see 8.20/11 above.]
767    
768    18. If a pattern containing \R was studied, it was assumed that \R always
769        matched two bytes, thus causing the minimum subject length to be
770        incorrectly computed because \R can also match just one byte.
771    
772    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
773        was incorrectly computed.
774    
775    20. If /S is present twice on a test pattern in pcretest input, it now
776        *disables* studying, thereby overriding the use of -s on the command line
777        (see 14 above). This is necessary for one or two tests to keep the output
778        identical in both cases.
779    
780    21. When (*ACCEPT) was used in an assertion that matched an empty string and
781        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
782    
783    22. When an atomic group that contained a capturing parenthesis was
784        successfully matched, but the branch in which it appeared failed, the
785        capturing was not being forgotten if a higher numbered group was later
786        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
787        group 1 to "a", when in fact it should be unset. This applied to multi-
788        branched capturing and non-capturing groups, repeated or not, and also to
789        positive assertions (capturing in negative assertions does not happen
790        in PCRE) and also to nested atomic groups.
791    
792    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
793        subject after a captured substring, to make it easier to tell which of a
794        number of identical substrings has been captured.
795    
796    24. The way atomic groups are processed by pcre_exec() has been changed so that
797        if they are repeated, backtracking one repetition now resets captured
798        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
799        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
800        "aaa". Previously, it would have been "a". As part of this code
801        refactoring, the way recursive calls are handled has also been changed.
802    
803    25. If an assertion condition captured any substrings, they were not passed
804        back unless some other capturing happened later. For example, if
805        (?(?=(a))a) was matched against "a", no capturing was returned.
806    
807    26. When studying a pattern that contained subroutine calls or assertions,
808        the code for finding the minimum length of a possible match was handling
809        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
810        group 1 called group 2 while simultaneously a separate group 2 called group
811        1). A stack overflow occurred in this case. I have fixed this by limiting
812        the recursion depth to 10.
813    
814    27. Updated RunTest.bat in the distribution to the version supplied by Tom
815        Fortmann. This supports explicit test numbers on the command line, and has
816        argument validation and error reporting.
817    
818    28. An instance of \X with an unlimited repeat could fail if at any point the
819        first character it looked at was a mark character.
820    
821    29. Some minor code refactoring concerning Unicode properties and scripts
822        should reduce the stack requirement of match() slightly.
823    
824    30. Added the '=' option to pcretest to check the setting of unused capturing
825        slots at the end of the pattern, which are documented as being -1, but are
826        not included in the return count.
827    
828    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
829        compiled something random. Now it gives a compile-time error (as does
830        Perl).
831    
832    32. A *MARK encountered during the processing of a positive assertion is now
833        recorded and passed back (compatible with Perl).
834    
835    33. If --only-matching or --colour was set on a pcregrep call whose pattern
836        had alternative anchored branches, the search for a second match in a line
837        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
838        matched the line "0102" twice. The same bug affected patterns that started
839        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
840        twice.
841    
842    34. Previously, PCRE did not allow quantification of assertions. However, Perl
843        does, and because of capturing effects, quantifying parenthesized
844        assertions may at times be useful. Quantifiers are now allowed for
845        parenthesized assertions.
846    
847    35. A minor code tidy in pcre_compile() when checking options for \R usage.
848    
849    36. \g was being checked for fancy things in a character class, when it should
850        just be a literal "g".
851    
852    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
853        appearance of a nested POSIX class supersedes an apparent external class.
854        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
855        unescaped square brackets may also appear as part of class names. For
856        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
857        more like Perl. (But see 8.20/1 above.)
858    
859    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
860        was because it thought it was \N{name}, which is not supported).
861    
862    39. Add minix to OS list not supporting the -S option in pcretest.
863    
864    40. PCRE tries to detect cases of infinite recursion at compile time, but it
865        cannot analyze patterns in sufficient detail to catch mutual recursions
866        such as ((?1))((?2)). There is now a runtime test that gives an error if a
867        subgroup is called recursively as a subpattern for a second time at the
868        same position in the subject string. In previous releases this might have
869        been caught by the recursion limit, or it might have run out of stack.
870    
871    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
872        happen only once. PCRE was, however incorrectly giving a compile time error
873        "recursive call could loop indefinitely" because it cannot analyze the
874        pattern in sufficient detail. The compile time test no longer happens when
875        PCRE is compiling a conditional subpattern, but actual runaway loops are
876        now caught at runtime (see 40 above).
877    
878    42. It seems that Perl allows any characters other than a closing parenthesis
879        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
880        has been changed to be the same.
881    
882    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
883        as not to get warnings when autogen.sh is called. Also changed
884        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
885    
886    44. To help people who use pcregrep to scan files containing exceedingly long
887        lines, the following changes have been made:
888    
889        (a) The default value of the buffer size parameter has been increased from
890            8K to 20K. (The actual buffer used is three times this size.)
891    
892        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
893            PCRE is built.
894    
895        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
896            to be set at run time.
897    
898        (d) Numerical values in pcregrep options can be followed by K or M, for
899            example --buffer-size=50K.
900    
901        (e) If a line being scanned overflows pcregrep's buffer, an error is now
902            given and the return code is set to 2.
903    
904    45. Add a pointer to the latest mark to the callout data block.
905    
906    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
907        partial match of an empty string instead of no match. This was specific to
908        the use of ".".
909    
910    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
911        complete match instead of a partial match. This bug was dependent on both
912        the PCRE_UTF8 and PCRE_DOTALL options being set.
913    
914    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
915        starting byte set, because \b was not being ignored.
916    
917    
918    Version 8.12 15-Jan-2011
919    ------------------------
920    
921    1.  Fixed some typos in the markup of the man pages, and wrote a script that
922        checks for such things as part of the documentation building process.
923    
924    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
925        --match-limit and --recursion-limit options (added for 8.11). In
926        particular, this made one of the standard tests fail. (The integer value
927        went into the wrong half of a long int.)
928    
929    3.  If the --colour option was given to pcregrep with -v (invert match), it
930        did strange things, either producing crazy output, or crashing. It should,
931        of course, ignore a request for colour when reporting lines that do not
932        match.
933    
934    4.  Another pcregrep bug caused similar problems if --colour was specified with
935        -M (multiline) and the pattern match finished with a line ending.
936    
937    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
938        matched in multiline mode, the following line was shown as part of the
939        match. This seems wrong, so I have changed it.
940    
941    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
942        the check for further matches in the same line (so they could be coloured)
943        to overrun the end of the current line. If another match was found, it was
944        incorrectly shown (and then shown again when found in the next line).
945    
946    7.  If pcregrep was compiled under Windows, there was a reference to the
947        function pcregrep_exit() before it was defined. I am assuming this was
948        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
949        reported by a user. I've moved the definition above the reference.
950    
951    
952    Version 8.11 10-Dec-2010
953    ------------------------
954    
955    1.  (*THEN) was not working properly if there were untried alternatives prior
956        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
957        backtracked to try for "b" instead of moving to the next alternative branch
958        at the same level (in this case, to look for "c"). The Perl documentation
959        is clear that when (*THEN) is backtracked onto, it goes to the "next
960        alternative in the innermost enclosing group".
961    
962    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
963        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
964        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
965        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
966        (*THEN).
967    
968    3.  If \s appeared in a character class, it removed the VT character from
969        the class, even if it had been included by some previous item, for example
970        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
971        of \s, but is part of the POSIX "space" class.)
972    
973    4.  A partial match never returns an empty string (because you can always
974        match an empty string at the end of the subject); however the checking for
975        an empty string was starting at the "start of match" point. This has been
976        changed to the "earliest inspected character" point, because the returned
977        data for a partial match starts at this character. This means that, for
978        example, /(?<=abc)def/ gives a partial match for the subject "abc"
979        (previously it gave "no match").
980    
981    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
982        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
983        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
984        has an implication that the given string is incomplete (because a partial
985        match is preferred over a full match). For this reason, these items now
986        give a partial match in this situation. [Aside: previously, the one case
987        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
988        match rather than a full match, which was wrong by the old rules, but is
989        now correct.]
990    
991    6.  There was a bug in the handling of #-introduced comments, recognized when
992        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
993        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
994        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
995        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
996        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
997        places in pcre_compile().
998    
999    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1000        comments when looking ahead for named forward references to subpatterns,
1001        the only newline sequence it recognized was NL. It now handles newlines
1002        according to the set newline convention.
1003    
1004    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1005        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1006        cater for a lack of strerror(). These oversights have been fixed.
1007    
1008    9.  Added --match-limit and --recursion-limit to pcregrep.
1009    
1010    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1011    
1012    11. When the -o option was used, pcregrep was setting a return code of 1, even
1013        when matches were found, and --line-buffered was not being honoured.
1014    
1015    12. Added an optional parentheses number to the -o and --only-matching options
1016        of pcregrep.
1017    
1018    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1019        can match an empty string. The code to do it in pcretest and pcredemo
1020        needed fixing:
1021    
1022        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1023            only one byte after an empty string match just before CRLF (this case
1024            just got forgotten; "any" and "anycrlf" were OK).
1025    
1026        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1027            mode when an empty string match preceded an ASCII character followed by
1028            a non-ASCII character. (The code for advancing by one character rather
1029            than one byte was nonsense.)
1030    
1031        (c) The pcredemo.c sample program did not have any code at all to handle
1032            the cases when CRLF is a valid newline sequence.
1033    
1034    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1035        as a starting offset was within the subject string. There is now a new
1036        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1037        negative or greater than the length of the string. In order to test this,
1038        pcretest is extended to allow the setting of negative starting offsets.
1039    
1040    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1041        starting offset points to the beginning of a UTF-8 character was
1042        unnecessarily clumsy. I tidied it up.
1043    
1044    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1045        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1046    
1047    17. Nobody had reported that the --include_dir option, which was added in
1048        release 7.7 should have been called --include-dir (hyphen, not underscore)
1049        for compatibility with GNU grep. I have changed it to --include-dir, but
1050        left --include_dir as an undocumented synonym, and the same for
1051        --exclude-dir, though that is not available in GNU grep, at least as of
1052        release 2.5.4.
1053    
1054    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1055        characters from a string of bytes) have been redefined so as not to use
1056        loops, in order to improve performance in some environments. At the same
1057        time, I abstracted some of the common code into auxiliary macros to save
1058        repetition (this should not affect the compiled code).
1059    
1060    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1061        compile-time error is now given if \c is not followed by an ASCII
1062        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1063        different, and any byte value is allowed.)
1064    
1065    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1066        START_OPTIMIZE option, which is now allowed at compile time - but just
1067        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1068        to pcregrep and other applications that have no direct access to PCRE
1069        options. The new /Y option in pcretest sets this option when calling
1070        pcre_compile().
1071    
1072    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1073        back references. Groups containing recursive back references were forced to
1074        be atomic by that change, but in the case of named groups, the amount of
1075        memory required was incorrectly computed, leading to "Failed: internal
1076        error: code overflow". This has been fixed.
1077    
1078    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1079        pcretest.c, to avoid build problems in some Borland environments.
1080    
1081    
1082    Version 8.10 25-Jun-2010
1083    ------------------------
1084    
1085    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1086        THEN.
1087    
1088    2.  (*ACCEPT) was not working when inside an atomic group.
1089    
1090    3.  Inside a character class, \B is treated as a literal by default, but
1091        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1092        causes the error). The code is unchanged, but I tidied the documentation.
1093    
1094    4.  Inside a character class, PCRE always treated \R and \X as literals,
1095        whereas Perl faults them if its -w option is set. I have changed PCRE so
1096        that it faults them when PCRE_EXTRA is set.
1097    
1098    5.  Added support for \N, which always matches any character other than
1099        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1100    
1101    6.  When compiling pcregrep with newer versions of gcc which may have
1102        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1103        declared with attribute warn_unused_result" were given. Just casting the
1104        result to (void) does not stop the warnings; a more elaborate fudge is
1105        needed. I've used a macro to implement this.
1106    
1107    7.  Minor change to pcretest.c to avoid a compiler warning.
1108    
1109    8.  Added four artifical Unicode properties to help with an option to make
1110        \s etc use properties (see next item). The new properties are: Xan
1111        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1112    
1113    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1114        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1115        this option. Modified pcretest to add /W to test this facility. Added
1116        REG_UCP to make it available via the POSIX interface.
1117    
1118    10. Added --line-buffered to pcregrep.
1119    
1120    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1121        studied, and the match started with a letter with a code point greater than
1122        127 whose first byte was different to the first byte of the other case of
1123        the letter, the other case of this starting letter was not recognized
1124        (#976).
1125    
1126    12. If a pattern that was studied started with a repeated Unicode property
1127        test, for example, \p{Nd}+, there was the theoretical possibility of
1128        setting up an incorrect bitmap of starting bytes, but fortunately it could
1129        not have actually happened in practice until change 8 above was made (it
1130        added property types that matched character-matching opcodes).
1131    
1132    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1133        possible starting bytes for non-anchored patterns.
1134    
1135    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1136        \R, and also a number of cases that involve Unicode properties, both
1137        explicit and implicit when PCRE_UCP is set.
1138    
1139    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1140        input, it could crash or give wrong results if characters with values
1141        greater than 0xc0 were present in the subject string. (Detail: it assumed
1142        UTF-8 input when processing these items.)
1143    
1144    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1145        size_t is 64-bit (#991).
1146    
1147    17. Added a check for running out of memory when PCRE is compiled with
1148        --disable-stack-for-recursion (#990).
1149    
1150    18. If the last data line in a file for pcretest does not have a newline on
1151        the end, a newline was missing in the output.
1152    
1153    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1154        less than 128) in its various bitmaps. However, there is a facility for
1155        generating tables according to the current locale when PCRE is compiled. It
1156        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1157        characters, are recognized by isspace() and therefore were getting set in
1158        these tables, and indeed these tables seem to approximate to ISO 8859. This
1159        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1160        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1161        which of course cannot start UTF-8 characters. I have changed the code so
1162        that only real ASCII characters (less than 128) and the correct starting
1163        bytes for UTF-8 encodings are set for characters greater than 127 when in
1164        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1165        altogether.)
1166    
1167    20. Added the /T option to pcretest so as to be able to run tests with non-
1168        standard character tables, thus making it possible to include the tests
1169        used for 19 above in the standard set of tests.
1170    
1171    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1172        reference to a subpattern the other side of a comment that contains an
1173        opening parenthesis caused either an internal compiling error, or a
1174        reference to the wrong subpattern.
1175    
1176    
1177    Version 8.02 19-Mar-2010
1178    ------------------------
1179    
1180    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1181    
1182    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1183        configured.
1184    
1185    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1186        original author of that file, following a query about its status.
1187    
1188    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1189        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1190    
1191    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1192        quantifier applied to a forward-referencing subroutine call, could compile
1193        incorrect code or give the error "internal error: previously-checked
1194        referenced subpattern not found".
1195    
1196    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1197        variables to point to external functions. For these systems, therefore,
1198        pcre_malloc etc. are now initialized to local functions that call the
1199        relevant global functions.
1200    
1201    7.  There were two entries missing in the vectors called coptable and poptable
1202        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1203        I've fixed the data, and added a kludgy way of testing at compile time that
1204        the lengths are correct (equal to the number of opcodes).
1205    
1206    8.  Following on from 7, I added a similar kludge to check the length of the
1207        eint vector in pcreposix.c.
1208    
1209    9.  Error texts for pcre_compile() are held as one long string to avoid too
1210        much relocation at load time. To find a text, the string is searched,
1211        counting zeros. There was no check for running off the end of the string,
1212        which could happen if a new error number was added without updating the
1213        string.
1214    
1215    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1216    
1217    11. \K was not working if it appeared in an atomic group or in a group that
1218        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1219        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1220        the assertion is positive, but not if it is negative.
1221    
1222    12. Change 11 fortuitously reduced the size of the stack frame used in the
1223        "match()" function of pcre_exec.c by one pointer. Forthcoming
1224        implementation of support for (*MARK) will need an extra pointer on the
1225        stack; I have reserved it now, so that the stack frame size does not
1226        decrease.
1227    
1228    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1229        item in branch that calls a recursion is a subroutine call - as in the
1230        second branch in the above example - was incorrectly given the compile-
1231        time error "recursive call could loop indefinitely" because pcre_compile()
1232        was not correctly checking the subroutine for matching a non-empty string.
1233    
1234    14. The checks for overrunning compiling workspace could trigger after an
1235        overrun had occurred. This is a "should never occur" error, but it can be
1236        triggered by pathological patterns such as hundreds of nested parentheses.
1237        The checks now trigger 100 bytes before the end of the workspace.
1238    
1239    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1240    
1241    
1242    Version 8.01 19-Jan-2010
1243    ------------------------
1244    
1245    1.  If a pattern contained a conditional subpattern with only one branch (in
1246        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1247        computed the wrong minimum data length (which is of course zero for such
1248        subpatterns). This could cause incorrect "no match" results.
1249    
1250    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1251        the pattern is reset in the first branch, pcre_compile() failed with
1252        "internal error: code overflow at offset...". This happened only when
1253        the reset was to the original external option setting. (An optimization
1254        abstracts leading options settings into an external setting, which was the
1255        cause of this.)
1256    
1257    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1258        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1259        assertion pattern did not match (meaning that the assertion was true), it
1260        was incorrectly treated as false if the SKIP had been reached during the
1261        matching. This also applied to assertions used as conditions.
1262    
1263    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1264        assertion subpattern, including such a pattern used as a condition,
1265        unpredictable results occurred, instead of the error return
1266        PCRE_ERROR_DFA_UITEM.
1267    
1268    5.  The C++ GlobalReplace function was not working like Perl for the special
1269        situation when an empty string is matched. It now does the fancy magic
1270        stuff that is necessary.
1271    
1272    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1273        removed. (These were left over from very, very early versions of PCRE.)
1274    
1275    7.  Some cosmetic changes to the code to make life easier when compiling it
1276        as part of something else:
1277    
1278        (a) Change DEBUG to PCRE_DEBUG.
1279    
1280        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1281            called "current" as "current_branch", to prevent a collision with the
1282            Linux macro when compiled as a kernel module.
1283    
1284        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1285            prevent a collision with the Linux macro when compiled as a kernel
1286            module.
1287    
1288    8.  In pcre_compile() there are some checks for integer overflows that used to
1289        cast potentially large values to (double). This has been changed to that
1290        when building, a check for int64_t is made, and if it is found, it is used
1291        instead, thus avoiding the use of floating point arithmetic. (There is no
1292        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1293        double.
1294    
1295    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1296        2005 (difference between two addresses compared to an unsigned value).
1297    
1298    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1299        custom one, because of the following reported problem in Windows:
1300    
1301          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1302              under Win32.
1303          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1304              therefore missing the function definition.
1305          - The compiler thus generates a "C" signature for the test function.
1306          - The linker fails to find the "C" function.
1307          - PCRE fails to configure if asked to do so against libbz2.
1308    
1309    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1310        messages were output:
1311    
1312          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1313          rerunning libtoolize, to keep the correct libtool macros in-tree.
1314          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1315    
1316        I have done both of these things.
1317    
1318    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1319        most of the time, it *can* run out if it is given a pattern that contains a
1320        runaway infinite recursion. I updated the discussion in the pcrestack man
1321        page.
1322    
1323    13. Now that we have gone to the x.xx style of version numbers, the minor
1324        version may start with zero. Using 08 or 09 is a bad idea because users
1325        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1326        interpreted as invalid octal numbers. I've updated the previous comment in
1327        configure.ac, and also added a check that gives an error if 08 or 09 are
1328        used.
1329    
1330    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1331        causing partial matching to fail when the end of the subject matched \W
1332        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1333    
1334    15. There were some discrepancies between the declarations in pcre_internal.h
1335        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1336        their definitions. The declarations used "const uschar *" and the
1337        definitions used USPTR. Even though USPTR is normally defined as "const
1338        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1339        reported that: "This difference in casting confuses some C++ compilers, for
1340        example, SunCC recognizes above declarations as different functions and
1341        generates broken code for hbpcre." I have changed the declarations to use
1342        USPTR.
1343    
1344    16. GNU libtool is named differently on some systems. The autogen.sh script now
1345        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1346        (FreeBSD).
1347    
1348    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1349        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1350        comment: "Figure out how to create a longlong from a string: strtoll and
1351        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1352        instance, but it only takes 2 args instead of 3!"
1353    
1354    18. A subtle bug concerned with back references has been fixed by a change of
1355        specification, with a corresponding code fix. A pattern such as
1356        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1357        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1358        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1359        same bug. Such groups have to be quantified to be useful, or contained
1360        inside another quantified group. (If there's no repetition, the reference
1361        can never match.) The problem arises because, having left the group and
1362        moved on to the rest of the pattern, a later failure that backtracks into
1363        the group uses the captured value from the final iteration of the group
1364        rather than the correct earlier one. I have fixed this in PCRE by forcing
1365        any group that contains a reference to itself to be an atomic group; that
1366        is, there cannot be any backtracking into it once it has completed. This is
1367        similar to recursive and subroutine calls.
1368    
1369    
1370    Version 8.00 19-Oct-09
1371    ----------------------
1372    
1373    1.  The table for translating pcre_compile() error codes into POSIX error codes
1374        was out-of-date, and there was no check on the pcre_compile() error code
1375        being within the table. This could lead to an OK return being given in
1376        error.
1377    
1378    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1379        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1380        in a Windows environment.
1381    
1382    3.  The pcregrep --count option prints the count for each file even when it is
1383        zero, as does GNU grep. However, pcregrep was also printing all files when
1384        --files-with-matches was added. Now, when both options are given, it prints
1385        counts only for those files that have at least one match. (GNU grep just
1386        prints the file name in this circumstance, but including the count seems
1387        more useful - otherwise, why use --count?) Also ensured that the
1388        combination -clh just lists non-zero counts, with no names.
1389    
1390    4.  The long form of the pcregrep -F option was incorrectly implemented as
1391        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1392        but it seems right to fix it, and I didn't think it was worth preserving
1393        the old behaviour.
1394    
1395    5.  The command line items --regex=pattern and --regexp=pattern were not
1396        recognized by pcregrep, which required --regex pattern or --regexp pattern
1397        (with a space rather than an '='). The man page documented the '=' forms,
1398        which are compatible with GNU grep; these now work.
1399    
1400    6.  No libpcreposix.pc file was created for pkg-config; there was just
1401        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1402    
1403    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1404        when UCP support is not needed, by modifying the Python script that
1405        generates it from Unicode data files. This should not matter if the module
1406        is correctly used as a library, but I received one complaint about 50K of
1407        unwanted data. My guess is that the person linked everything into his
1408        program rather than using a library. Anyway, it does no harm.
1409    
1410    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1411        was a minimum greater than 1 for a wide character in a possessive
1412        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1413        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1414        character. Chaos in the form of incorrect output or a compiling loop could
1415        result.
1416    
1417    9.  The restrictions on what a pattern can contain when partial matching is
1418        requested for pcre_exec() have been removed. All patterns can now be
1419        partially matched by this function. In addition, if there are at least two
1420        slots in the offset vector, the offset of the earliest inspected character
1421        for the match and the offset of the end of the subject are set in them when
1422        PCRE_ERROR_PARTIAL is returned.
1423    
1424    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1425        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1426        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1427        and may be more useful for multi-segment matching.
1428    
1429    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1430        used to be given if ever the end of the subject was reached; now it is
1431        given only if matching could not proceed because another character was
1432        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1433        string "Z", which now yields "no match" instead of "partial match". In the
1434        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1435        final character ended with (*FAIL).
1436    
1437    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1438        if the pattern had a "must contain" character that was already found in the
1439        earlier partial match, unless partial matching was again requested. For
1440        example, with the pattern /dog.(body)?/, the "must contain" character is
1441        "g". If the first part-match was for the string "dog", restarting with
1442        "sbody" failed. This bug has been fixed.
1443    
1444    13. The string returned by pcre_dfa_exec() after a partial match has been
1445        changed so that it starts at the first inspected character rather than the
1446        first character of the match. This makes a difference only if the pattern
1447        starts with a lookbehind assertion or \b or \B (\K is not supported by
1448        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1449        matching functions compatible, and I think it's the right thing to do.
1450    
1451    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1452        so that the demonstration program is easily available in environments where
1453        PCRE has not been installed from source.
1454    
1455    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1456        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1457        library.
1458    
1459    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1460        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1461        is not the first non-POSIX option to be added. Clearly some people find
1462        these options useful.
1463    
1464    17. If a caller to the POSIX matching function regexec() passes a non-zero
1465        value for nmatch with a NULL value for pmatch, the value of
1466        nmatch is forced to zero.
1467    
1468    18. RunGrepTest did not have a test for the availability of the -u option of
1469        the diff command, as RunTest does. It now checks in the same way as
1470        RunTest, and also checks for the -b option.
1471    
1472    19. If an odd number of negated classes containing just a single character
1473        interposed, within parentheses, between a forward reference to a named
1474        subpattern and the definition of the subpattern, compilation crashed with
1475        an internal error, complaining that it could not find the referenced
1476        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1477        [The bug was that it was starting one character too far in when skipping
1478        over the character class, thus treating the ] as data rather than
1479        terminating the class. This meant it could skip too much.]
1480    
1481    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1482        /g option in pcretest when the pattern contains \K, which makes it possible
1483        to have an empty string match not at the start, even when the pattern is
1484        anchored. Updated pcretest and pcredemo to use this option.
1485    
1486    21. If the maximum number of capturing subpatterns in a recursion was greater
1487        than the maximum at the outer level, the higher number was returned, but
1488        with unset values at the outer level. The correct (outer level) value is
1489        now given.
1490    
1491    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1492        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1493        make it do so. The string so far is captured, making this feature
1494        compatible with Perl.
1495    
1496    23. The tests have been re-organized, adding tests 11 and 12, to make it
1497        possible to check the Perl 5.10 features against Perl 5.10.
1498    
1499    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1500        pattern matches a fixed length string. PCRE did not allow this; now it
1501        does. Neither allows recursion.
1502    
1503    25. I finally figured out how to implement a request to provide the minimum
1504        length of subject string that was needed in order to match a given pattern.
1505        (It was back references and recursion that I had previously got hung up
1506        on.) This code has now been added to pcre_study(); it finds a lower bound
1507        to the length of subject needed. It is not necessarily the greatest lower
1508        bound, but using it to avoid searching strings that are too short does give
1509        some useful speed-ups. The value is available to calling programs via
1510        pcre_fullinfo().
1511    
1512    26. While implementing 25, I discovered to my embarrassment that pcretest had
1513        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1514        study optimizations had never been tested with that matching function.
1515        Oops. What is worse, even when it was passed study data, there was a bug in
1516        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1517        were also very few tests of studied patterns with pcre_dfa_exec().
1518    
1519    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1520        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1521        on the other side of the coin, they are no longer allowed to have different
1522        names, because these cannot be distinguished in PCRE, and this has caused
1523        confusion. (This is a difference from Perl.)
1524    
1525    28. When duplicate subpattern names are present (necessarily with different
1526        numbers, as required by 27 above), and a test is made by name in a
1527        conditional pattern, either for a subpattern having been matched, or for
1528        recursion in such a pattern, all the associated numbered subpatterns are
1529        tested, and the overall condition is true if the condition is true for any
1530        one of them. This is the way Perl works, and is also more like the way
1531        testing by number works.
1532    
1533    
1534    Version 7.9 11-Apr-09
1535    ---------------------
1536    
1537    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1538        (pcretest), all targets were linked against these libraries. This included
1539        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1540        libraries. This caused unwanted dependencies to be created. This problem
1541        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1542        pcretest is linked with readline.
1543    
1544    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1545        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1546        moved outside it again, because FALSE and TRUE are already defined in AIX,
1547        but BOOL is not.
1548    
1549    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1550        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1551    
1552    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1553        hyphens) following file names and line numbers when outputting matching
1554        lines. This is not true; no spaces are inserted. I have also clarified the
1555        wording for the --colour (or --color) option.
1556    
1557    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1558        was not coloured; this is different to GNU grep, so I have changed it to be
1559        the same.
1560    
1561    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1562        each matching line was coloured. Now it goes on to look for further matches
1563        of any of the test patterns, which is the same behaviour as GNU grep.
1564    
1565    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1566        doesn't make sense to accept an empty string match in pcregrep, so I have
1567        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1568        seems to be how GNU grep behaves.
1569    
1570    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1571        start or after a newline", because the conditional assertion was not being
1572        correctly handled. The rule now is that both the assertion and what follows
1573        in the first alternative must satisfy the test.
1574    
1575    9.  If auto-callout was enabled in a pattern with a conditional group whose
1576        condition was an assertion, PCRE could crash during matching, both with
1577        pcre_exec() and pcre_dfa_exec().
1578    
1579    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1580        used for matching.
1581    
1582    11. Unicode property support in character classes was not working for
1583        characters (bytes) greater than 127 when not in UTF-8 mode.
1584    
1585    12. Added the -M command line option to pcretest.
1586    
1587    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1588    
1589    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1590    
1591    16. Added comments and documentation about mis-use of no_arg in the C++
1592        wrapper.
1593    
1594    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1595        from Martin Jerabek that uses macro names for all relevant character and
1596        string constants.
1597    
1598    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1599        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1600        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1601        these, but not everybody uses configure.
1602    
1603    19. A conditional group that had only one branch was not being correctly
1604        recognized as an item that could match an empty string. This meant that an
1605        enclosing group might also not be so recognized, causing infinite looping
1606        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1607        with the subject "ab", where knowledge that the repeated group can match
1608        nothing is needed in order to break the loop.
1609    
1610    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1611        exec(), but without supplying a callout function, matching went wrong.
1612    
1613    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1614        leak if the size of the offset vector was greater than 30. When the vector
1615        is smaller, the saved offsets during recursion go onto a local stack
1616        vector, but for larger vectors malloc() is used. It was failing to free
1617        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1618        error, in fact).
1619    
1620    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1621        heapframe that is used only when UTF-8 support is enabled. This caused no
1622        problem, but was untidy.
1623    
1624    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1625        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1626        included within another project.
1627    
1628    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1629        slightly modified by me:
1630    
1631          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1632              not building pcregrep.
1633    
1634          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1635              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1636    
1637    25. Forward references, both numeric and by name, in patterns that made use of
1638        duplicate group numbers, could behave incorrectly or give incorrect errors,
1639        because when scanning forward to find the reference group, PCRE was not
1640        taking into account the duplicate group numbers. A pattern such as
1641        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1642    
1643    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1644        the feature of a custom pointer more persuasive (as requested by a user).
1645    
1646    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1647        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1648        pre-defined.
1649    
1650    28. Added support for (*UTF8) at the start of a pattern.
1651    
1652    29. Arrange for flags added by the "release type" setting in CMake to be shown
1653        in the configuration summary.
1654    
1655    
1656    Version 7.8 05-Sep-08
1657    ---------------------
1658    
1659    1.  Replaced UCP searching code with optimized version as implemented for Ad
1660        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1661        stage table and inline lookup instead of a function, giving speed ups of 2
1662        to 5 times on some simple patterns that I tested. Permission was given to
1663        distribute the MultiStage2.py script that generates the tables (it's not in
1664        the tarball, but is in the Subversion repository).
1665    
1666    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1667        scripts.
1668    
1669    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1670        a group with a zero qualifier. The result of the study could be incorrect,
1671        or the function might crash, depending on the pattern.
1672    
1673    4.  Caseless matching was not working for non-ASCII characters in back
1674        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1675        It now works when Unicode Property Support is available.
1676    
1677    5.  In pcretest, an escape such as \x{de} in the data was always generating
1678        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1679        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1680        truncation.
1681    
1682    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1683    
1684    7.  Added two (int) casts to pcregrep when printing the difference of two
1685        pointers, in case they are 64-bit values.
1686    
1687    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1688        test 2 if it fails.
1689    
1690    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1691        and a #define of that name to empty if it is not externally set. This is to
1692        allow users of MSVC to set it if necessary.
1693    
1694    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1695        the convenience functions in the pcre_get.c source file.
1696    
1697    11. An option change at the start of a pattern that had top-level alternatives
1698        could cause overwriting and/or a crash. This command provoked a crash in
1699        some environments:
1700    
1701          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1702    
1703        This potential security problem was recorded as CVE-2008-2371.
1704    
1705    12. For a pattern where the match had to start at the beginning or immediately
1706        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1707        pcre_dfa_exec() could read past the end of the passed subject if there was
1708        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1709        pcretest so that it places the subject at the end of its malloc-ed buffer.
1710    
1711    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1712        exec() might read past the end of the data buffer in UTF-8 mode.
1713    
1714    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1715        the data contained the byte 0x85 as part of a UTF-8 character within its
1716        first line. This applied both to normal and DFA matching.
1717    
1718    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1719        /^[^d]*?$/8 failed to match "abc".
1720    
1721    16. Added a missing copyright notice to pcrecpp_internal.h.
1722    
1723    17. Make it more clear in the documentation that values returned from
1724        pcre_exec() in ovector are byte offsets, not character counts.
1725    
1726    18. Tidied a few places to stop certain compilers from issuing warnings.
1727    
1728    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1729        supplied by Stefan Weber. I made a further small update for 7.8 because
1730        there is a change of source arrangements: the pcre_searchfuncs.c module is
1731        replaced by pcre_ucd.c.
1732    
1733    
1734    Version 7.7 07-May-08
1735    ---------------------
1736    
1737    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1738        a string to a long long, pretend we don't even have a long long." This is
1739        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1740    
1741    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1742        pre-7.6 versions, which defined a global no_arg variable instead of putting
1743        it in the RE class. (See also #8 below.)
1744    
1745    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1746        Lopes.
1747    
1748    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1749    
1750        (1) The include/exclude patterns were being applied to the whole pathnames
1751            of files, instead of just to the final components.
1752    
1753        (2) If there was more than one level of directory, the subdirectories were
1754            skipped unless they satisfied the include/exclude conditions. This is
1755            inconsistent with GNU grep (and could even be seen as contrary to the
1756            pcregrep specification - which I improved to make it absolutely clear).
1757            The action now is always to scan all levels of directory, and just
1758            apply the include/exclude patterns to regular files.
1759    
1760    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1761        --exclude_dir in the tests to avoid scanning .svn directories.
1762    
1763    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1764        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1765        doesn't support NULs in patterns.
1766    
1767    7.  Added some missing "const"s to declarations of static tables in
1768        pcre_compile.c and pcre_dfa_exec.c.
1769    
1770    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1771        caused by fix #2  above. (Subsequently also a second patch to fix the
1772        first patch. And a third patch - this was a messy problem.)
1773    
1774    9.  Applied Craig's patch to remove the use of push_back().
1775    
1776    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1777        matching function regexec().
1778    
1779    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1780        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1781        references. PCRE supports relative numbers with this syntax (I don't think
1782        Oniguruma does).
1783    
1784    12. Previously, a group with a zero repeat such as (...){0} was completely
1785        omitted from the compiled regex. However, this means that if the group
1786        was called as a subroutine from elsewhere in the pattern, things went wrong
1787        (an internal error was given). Such groups are now left in the compiled
1788        pattern, with a new opcode that causes them to be skipped at execution
1789        time.
1790    
1791    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1792        to the way PCRE behaves:
1793    
1794        (a) A lone ] character is dis-allowed (Perl treats it as data).
1795    
1796        (b) A back reference to an unmatched subpattern matches an empty string
1797            (Perl fails the current match path).
1798    
1799        (c) A data ] in a character class must be notated as \] because if the
1800            first data character in a class is ], it defines an empty class. (In
1801            Perl it is not possible to have an empty class.) The empty class []
1802            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1803            The negative empty class [^] matches any one character, independently
1804            of the DOTALL setting.
1805    
1806    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1807        non-existent subpattern following a character class starting with ']' and
1808        containing () gave an internal compiling error instead of "reference to
1809        non-existent subpattern". Fortunately, when the pattern did exist, the
1810        compiled code was correct. (When scanning forwards to check for the
1811        existencd of the subpattern, it was treating the data ']' as terminating
1812        the class, so got the count wrong. When actually compiling, the reference
1813        was subsequently set up correctly.)
1814    
1815    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1816        it was being rejected as not supported by pcre_dfa_exec(), even though
1817        other assertions are supported. I have made pcre_dfa_exec() support
1818        (*FAIL).
1819    
1820    16. The implementation of 13c above involved the invention of a new opcode,
1821        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1822        cannot be changed at match time, I realized I could make a small
1823        improvement to matching performance by compiling OP_ALLANY instead of
1824        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1825        on the OP_ANY path.
1826    
1827    17. Compiling pcretest on Windows with readline support failed without the
1828        following two fixes: (1) Make the unistd.h include conditional on
1829        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1830    
1831    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1832        ncurses library to be included for pcretest when ReadLine support is
1833        requested, but also to allow for it to be overridden. This patch came from
1834        Daniel Bergström.
1835    
1836    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1837        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1838        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1839        spotting this.
1840    
1841    
1842    Version 7.6 28-Jan-08
1843    ---------------------
1844    
1845    1.  A character class containing a very large number of characters with
1846        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1847        overflow.
1848    
1849    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1850        HAVE_LONG_LONG is not defined.
1851    
1852    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1853        bring them up to date and include new features. This patch includes:
1854    
1855        - Fixed PH's badly added libz and libbz2 support.
1856        - Fixed a problem with static linking.
1857        - Added pcredemo. [But later removed - see 7 below.]
1858        - Fixed dftables problem and added an option.
1859        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1860            HAVE_LONG_LONG.
1861        - Added readline support for pcretest.
1862        - Added an listing of the option settings after cmake has run.
1863    
1864    4.  A user submitted a patch to Makefile that makes it easy to create
1865        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1866        Makefile.am that cause it to include this special target, without
1867        affecting anything else. Note that the same mingw target plus all
1868        the other distribution libraries and programs are now supported
1869        when configuring with CMake (see 6 below) instead of with
1870        Configure/Make.
1871    
1872    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1873        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1874        exported in the Windows port". It has not yet been confirmed that the patch
1875        solves the problem, but it does no harm.
1876    
1877    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1878        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1879        with CMake, and also correct the comment about stack recursion.
1880    
1881    7.  Remove the automatic building of pcredemo from the ./configure system and
1882        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1883        of a program that users should build themselves after PCRE is installed, so
1884        building it automatically is not really right. What is more, it gave
1885        trouble in some build environments.
1886    
1887    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1888    
1889    
1890    Version 7.5 10-Jan-08
1891    ---------------------
1892    
1893    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1894        values in parens when parsing an RE using the C++ wrapper."
1895    
1896    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1897        Characters greater than 255 were excluded from the class instead of being
1898        included.
1899    
1900    3.  The same bug as (2) above applied to negated POSIX classes such as
1901        [:^space:].
1902    
1903    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1904        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1905        I have changed it.
1906    
1907    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1908        first named subpattern) and a construct such as (?&a) would reference the
1909        first named subpattern whose name started with "a" (in other words, the
1910        length check was missing). Both these problems are fixed. "Subpattern name
1911        expected" is now given for (?&) (a zero-length name), and this patch also
1912        makes it give the same error for \k'' (previously it complained that that
1913        was a reference to a non-existent subpattern).
1914    
1915    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1916        this is right because (?- can be followed by option settings as well as by
1917        digits. I have, however, made the messages clearer.
1918    
1919    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1920        than the number used in the conditional) now cause a compile-time error.
1921        This is actually not compatible with Perl, which accepts such patterns, but
1922        treats the conditional as always being FALSE (as PCRE used to), but it
1923        seems to me that giving a diagnostic is better.
1924    
1925    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1926        and messages.
1927    
1928    9.  Fix two occurrences of "backslash" in comments that should have been
1929        "backspace".
1930    
1931    10. Remove two redundant lines of code that can never be obeyed (their function
1932        was moved elsewhere).
1933    
1934    11. The program that makes PCRE's Unicode character property table had a bug
1935        which caused it to generate incorrect table entries for sequences of
1936        characters that have the same character type, but are in different scripts.
1937        It amalgamated them into a single range, with the script of the first of
1938        them. In other words, some characters were in the wrong script. There were
1939        thirteen such cases, affecting characters in the following ranges:
1940    
1941          U+002b0 - U+002c1
1942          U+0060c - U+0060d
1943          U+0061e - U+00612
1944          U+0064b - U+0065e
1945          U+0074d - U+0076d
1946          U+01800 - U+01805
1947          U+01d00 - U+01d77
1948          U+01d9b - U+01dbf
1949          U+0200b - U+0200f
1950          U+030fc - U+030fe
1951          U+03260 - U+0327f
1952          U+0fb46 - U+0fbb1
1953          U+10450 - U+1049d
1954    
1955    12. The -o option (show only the matching part of a line) for pcregrep was not
1956        compatible with GNU grep in that, if there was more than one match in a
1957        line, it showed only the first of them. It now behaves in the same way as
1958        GNU grep.
1959    
1960    13. If the -o and -v options were combined for pcregrep, it printed a blank
1961        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1962        does the same. The return code can be used to tell if there were any
1963        non-matching lines.
1964    
1965    14. Added --file-offsets and --line-offsets to pcregrep.
1966    
1967    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1968        infinitely looping recursion. The bug was that positive lookaheads were not
1969        being skipped when checking for a possible empty match (negative lookaheads
1970        and both kinds of lookbehind were skipped).
1971    
1972    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1973        inclusion of <windows.h> to before rather than after the definition of
1974        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1975    
1976    17. Specifying a possessive quantifier with a specific limit for a Unicode
1977        character property caused pcre_compile() to compile bad code, which led at
1978        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1979        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1980        caused the error; without that there was no problem.
1981    
1982    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1983    
1984    19. Added --enable-pcretest-libreadline.
1985    
1986    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1987        RE::GlobalReplace(). As a result, the number of replacements returned was
1988        double what it should be. I removed one of the increments, but Craig sent a
1989        later patch that removed the other one (the right fix) and added unit tests
1990        that check the return values (which was not done before).
1991    
1992    21. Several CMake things:
1993    
1994        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1995            the names libpcre and libpcreposix, not just pcre and pcreposix.
1996    
1997        (2) The above change means that pcretest and pcregrep are now correctly
1998            linked with the newly-built libraries, not previously installed ones.
1999    
2000        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2001    
2002    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2003        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2004        UTF-8 newline character). The key issue is that the pattern starts .*;
2005        this means that the match must be either at the beginning, or after a
2006        newline. The bug was in the code for advancing after a failed match and
2007        checking that the new position followed a newline. It was not taking
2008        account of UTF-8 characters correctly.
2009    
2010    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2011        character classes. PCRE was not treating the sequence [:...:] as a
2012        character class unless the ... were all letters. Perl, however, seems to
2013        allow any characters between [: and :], though of course it rejects as
2014        unknown any "names" that contain non-letters, because all the known class
2015        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2016        for example, whereas PCRE did not - it did not recognize a POSIX character
2017        class. This seemed a bit dangerous, so the code has been changed to be
2018        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2019        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2020        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2021        Perl does, and where it didn't before.
2022    
2023    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2024        Windows environments %n is disabled by default.
2025    
2026    
2027    Version 7.4 21-Sep-07
2028    ---------------------
2029    
2030    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2031        means that a class such as [\s] counted as "explicit reference to CR or
2032        LF". That isn't really right - the whole point of the change was to try to
2033        help when there was an actual mention of one of the two characters. So now
2034        the change happens only if \r or \n (or a literal CR or LF) character is
2035        encountered.
2036    
2037    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2038        of both had grown to the point where there were only 3 bits left.
2039        Fortunately, there was spare space in the data structure, and so I have
2040        moved the internal flags into a new 16-bit field to free up more option
2041        bits.
2042    
2043    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2044        but did not set the internal JCHANGED flag - either of these is enough to
2045        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2046        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2047        start sets both bits.
2048    
2049    4.  Added options (at build time, compile time, exec time) to change \R from
2050        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2051    
2052    5.  doc/pcresyntax.html was missing from the distribution.
2053    
2054    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2055        compatibility, even though it is no longer used.
2056    
2057    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2058        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2059        windows.h file is present (where different names are used). [This was
2060        reversed later after testing - see 16 below.]
2061    
2062    8.  Changed all #include <config.h> to #include "config.h". There were also
2063        some further <pcre.h> cases that I changed to "pcre.h".
2064    
2065    9.  When pcregrep was used with the --colour option, it missed the line ending
2066        sequence off the lines that it output.
2067    
2068    10. It was pointed out to me that arrays of string pointers cause lots of
2069        relocations when a shared library is dynamically loaded. A technique of
2070        using a single long string with a table of offsets can drastically reduce
2071        these. I have refactored PCRE in four places to do this. The result is
2072        dramatic:
2073    
2074          Originally:                          290
2075          After changing UCP table:            187
2076          After changing error message table:   43
2077          After changing table of "verbs"       36
2078          After changing table of Posix names   22
2079    
2080        Thanks to the folks working on Gregex for glib for this insight.
2081    
2082    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2083        unicode-properties was also set.
2084    
2085    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2086    
2087    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2088        checked only for CRLF.
2089    
2090    14. Added casts to pcretest.c to avoid compiler warnings.
2091    
2092    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2093    
2094    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2095        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2096        entirely. This removes changes made in 7 above.
2097    
2098    17. The CMake files have been updated, and there is now more information about
2099        building with CMake in the NON-UNIX-USE document.
2100    
2101    
2102    Version 7.3 28-Aug-07
2103    ---------------------
2104    
2105     1. In the rejigging of the build system that eventually resulted in 7.1, the
2106        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2107        brackets there is not right, since it causes compilers to look for an
2108        installed pcre.h, not the version that is in the source that is being
2109        compiled (which of course may be different). I have changed it back to:
2110    
2111          #include "pcre.h"
2112    
2113        I have a vague recollection that the change was concerned with compiling in
2114        different directories, but in the new build system, that is taken care of
2115        by the VPATH setting the Makefile.
2116    
2117     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2118        when the subject happened to end in the byte 0x85 (e.g. if the last
2119        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2120        characters but of course it shouldn't be taken as a newline when it is part
2121        of another character. The bug was that, for an unlimited repeat of . in
2122        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2123        characters when looking for a newline.
2124    
2125     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2126    
2127     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2128        in debug output.
2129    
2130     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2131        long printing in the pcrecpp unittest when running under MinGW.
2132    
2133     6. ESC_K was left out of the EBCDIC table.
2134    
2135     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2136        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2137        limit also applies to "virtual nesting" when a pattern is recursive, and in
2138        this case 1000 isn't so big. I have been able to remove this limit at the
2139        expense of backing off one optimization in certain circumstances. Normally,
2140        when pcre_exec() would call its internal match() function recursively and
2141        immediately return the result unconditionally, it uses a "tail recursion"
2142        feature to save stack. However, when a subpattern that can match an empty
2143        string has an unlimited repetition quantifier, it no longer makes this
2144        optimization. That gives it a stack frame in which to save the data for
2145        checking that an empty string has been matched. Previously this was taken
2146        from the 1000-entry workspace that had been reserved. So now there is no
2147        explicit limit, but more stack is used.
2148    
2149     8. Applied Daniel's patches to solve problems with the import/export magic
2150        syntax that is required for Windows, and which was going wrong for the
2151        pcreposix and pcrecpp parts of the library. These were overlooked when this
2152        problem was solved for the main library.
2153    
2154     9. There were some crude static tests to avoid integer overflow when computing
2155        the size of patterns that contain repeated groups with explicit upper
2156        limits. As the maximum quantifier is 65535, the maximum group length was
2157        set at 30,000 so that the product of these two numbers did not overflow a
2158        32-bit integer. However, it turns out that people want to use groups that
2159        are longer than 30,000 bytes (though not repeat them that many times).
2160        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2161        made it possible to implement the integer overflow checks in a much more
2162        dynamic way, which I have now done. The artificial limitation on group
2163        length has been removed - we now have only the limit on the total length of
2164        the compiled pattern, which depends on the LINK_SIZE setting.
2165    
2166    10. Fixed a bug in the documentation for get/copy named substring when
2167        duplicate names are permitted. If none of the named substrings are set, the
2168        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2169        empty string.
2170    
2171    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2172        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2173        because the ] is interpreted as the first data character and the
2174        terminating ] is not found. PCRE has been made compatible with Perl in this
2175        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2176        cause memory overwriting.
2177    
2178    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2179        string has been matched (to stop an infinite loop). It was not recognizing
2180        a conditional subpattern that could match an empty string if that
2181        subpattern was within another subpattern. For example, it looped when
2182        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2183        condition was not nested. This bug has been fixed.
2184    
2185    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2186        past the start of the subject in the presence of bytes with the top bit
2187        set, for example "\x8aBCD".
2188    
2189    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2190        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2191    
2192    14. Optimized (?!) to (*FAIL).
2193    
2194    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2195        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2196        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2197        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2198        does: it's just the validity check that is more restrictive.
2199    
2200    16. Inserted checks for integer overflows during escape sequence (backslash)
2201        processing, and also fixed erroneous offset values for syntax errors during
2202        backslash processing.
2203    
2204    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2205        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2206    
2207    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2208        caused an overrun.
2209    
2210    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2211        something other than just ASCII characters) inside a group that had an
2212        unlimited repeat caused a loop at compile time (while checking to see
2213        whether the group could match an empty string).
2214    
2215    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2216        [\P{Any}] did so. (Error in the code for printing property names.)
2217    
2218    21. An orphan \E inside a character class could cause a crash.
2219    
2220    22. A repeated capturing bracket such as (A)? could cause a wild memory
2221        reference during compilation.
2222    
2223    23. There are several functions in pcre_compile() that scan along a compiled
2224        expression for various reasons (e.g. to see if it's fixed length for look
2225        behind). There were bugs in these functions when a repeated \p or \P was
2226        present in the pattern. These operators have additional parameters compared
2227        with \d, etc, and these were not being taken into account when moving along
2228        the compiled data. Specifically:
2229    
2230        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2231            length.
2232    
2233        (b) An item such as \pL+ within a repeated group could cause crashes or
2234            loops.
2235    
2236        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2237            "reference to non-existent subpattern" error.
2238    
2239        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2240    
2241    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2242        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2243    
2244    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2245    
2246    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2247        character were causing crashes (broken optimization).
2248    
2249    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2250        \p or \P) caused a compile-time loop.
2251    
2252    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2253        break. For example, the unstudied pattern [\r\n]A does not match the string
2254        "\r\nA" because change 7.0/46 below moves the current point on by two
2255        characters after failing to match at the start. However, the pattern \nA
2256        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2257        the same is true. There doesn't seem any very clean way out of this, but
2258        what I have chosen to do makes the common cases work: PCRE now takes note
2259        of whether there can be an explicit match for \r or \n anywhere in the
2260        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2261        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2262        pattern has explicit CR or LF references.
2263    
2264    29. Added (*CR) etc for changing newline setting at start of pattern.
2265    
2266    
2267    Version 7.2 19-Jun-07
2268  ---------------------  ---------------------
2269    
2270   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 11  Version 7.2 01-May-07 Line 2274  Version 7.2 01-May-07
2274      to make them independent of the local environment's newline setting.      to make them independent of the local environment's newline setting.
2275    
2276   3. Add code to configure.ac to remove -g from the CFLAGS default settings.   3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2277    
2278   4. Some of the "internals" tests were previously cut out when the link size   4. Some of the "internals" tests were previously cut out when the link size
2279      was not 2, because the output contained actual offsets. The recent new      was not 2, because the output contained actual offsets. The recent new
2280      "Z" feature of pcretest means that these can be cut out, making the tests      "Z" feature of pcretest means that these can be cut out, making the tests
2281      usable with all link sizes.      usable with all link sizes.
2282    
2283   5. Implemented Stan Switzer's goto replacement for longjmp() when not using   5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2284      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
2285      a small improvement under Linux. However, it saves one field in the frame      a small improvement under Linux. However, it saves one field in the frame
2286      in all cases.      in all cases.
2287    
2288     6. Added more features from the forthcoming Perl 5.10:
2289    
2290        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2291            recursion call. It refers to the nth most recently opened parentheses.
2292    
2293        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2294            to be opened parentheses.
2295    
2296        (c) Conditions that refer to capturing parentheses can be specified
2297            relatively, for example, (?(-2)... or (?(+3)...
2298    
2299        (d) \K resets the start of the current match so that everything before
2300            is not part of it.
2301    
2302        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2303    
2304        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2305            reference syntax.
2306    
2307        (g) (?| introduces a group in which the numbering of parentheses in each
2308            alternative starts with the same number.
2309    
2310        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2311    
2312     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2313        PCRE_INFO_JCHANGED.
2314    
2315     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2316        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2317        for detecting groups that can match an empty string.
2318    
2319     9. A pattern with a very large number of alternatives (more than several
2320        hundred) was running out of internal workspace during the pre-compile
2321        phase, where pcre_compile() figures out how much memory will be needed. A
2322        bit of new cunning has reduced the workspace needed for groups with
2323        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2324        workspace instead of running out of the 4096 that are available.
2325    
2326    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2327    
2328    11. Applied patch from Google to remove an optimization that didn't quite work.
2329        The report of the bug said:
2330    
2331          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2332          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2333          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2334    
2335    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2336        it matched the wrong number of bytes.
2337    
2338    
2339  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
2340  ---------------------  ---------------------

Legend:
Removed from v.164  
changed lines
  Added in v.1248

  ViewVC Help
Powered by ViewVC 1.1.5