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

Diff of /code/trunk/ChangeLog

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

revision 108 by ph10, Wed Mar 7 11:11:23 2007 UTC revision 1235 by ph10, Mon Dec 10 13:36:29 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 05-Mar-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    
20    Version 8.32 30-November-2012
21    -----------------------------
22    
23    1.  Improved JIT compiler optimizations for first character search and single
24        character iterators.
25    
26    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
27        Patch by Daniel Richard G.
28    
29    3.  Single character iterator optimizations in the JIT compiler.
30    
31    4.  Improved JIT compiler optimizations for character ranges.
32    
33    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
34        Reported by Giuseppe D'Angelo.
35    
36    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
37        of a line, was being set incorrectly in cases where .* appeared inside
38        atomic brackets at the start of a pattern, or where there was a subsequent
39        *PRUNE or *SKIP.
40    
41    7.  Improved instruction cache flush for POWER/PowerPC.
42        Patch by Daniel Richard G.
43    
44    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
45        grep:
46    
47        (a) There is now no limit to the number of patterns to be matched.
48    
49        (b) An error is given if a pattern is too long.
50    
51        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
52            are now supported.
53    
54        (d) --exclude-from and --include-from (multiple use) have been added.
55    
56        (e) Exclusions and inclusions now apply to all files and directories, not
57            just to those obtained from scanning a directory recursively.
58    
59        (f) Multiple uses of -f and --file-list are now supported.
60    
61        (g) In a Windows environment, the default for -d has been changed from
62            "read" (the GNU grep default) to "skip", because otherwise the presence
63            of a directory in the file list provokes an error.
64    
65        (h) The documentation has been revised and clarified in places.
66    
67    9.  Improve the matching speed of capturing brackets.
68    
69    10. Changed the meaning of \X so that it now matches a Unicode extended
70        grapheme cluster.
71    
72    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
73        out POSIX threads when JIT support is configured.
74    
75    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
76    
77    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
78        structure could go wrong in environments where size_t is not the same size
79        as int.
80    
81    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
82        set.
83    
84    15. The EBCDIC support had decayed; later updates to the code had included
85        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
86        general tidy up of EBCDIC-related issues, and the documentation was also
87        not quite right. There is now a test that can be run on ASCII systems to
88        check some of the EBCDIC-related things (but is it not a full test).
89    
90    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
91        in a small tidy to the code.
92    
93    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
94    
95    18. If the --only-matching (-o) option in pcregrep is specified multiple
96        times, each one causes appropriate output. For example, -o1 -o2 outputs the
97        substrings matched by the 1st and 2nd capturing parentheses. A separating
98        string can be specified by --om-separator (default empty).
99    
100    19. Improving the first n character searches.
101    
102    20. Turn case lists for horizontal and vertical white space into macros so that
103        they are defined only once.
104    
105    21. This set of changes together give more compatible Unicode case-folding
106        behaviour for characters that have more than one other case when UCP
107        support is available.
108    
109        (a) The Unicode property table now has offsets into a new table of sets of
110            three or more characters that are case-equivalent. The MultiStage2.py
111            script that generates these tables (the pcre_ucd.c file) now scans
112            CaseFolding.txt instead of UnicodeData.txt for character case
113            information.
114    
115        (b) The code for adding characters or ranges of characters to a character
116            class has been abstracted into a generalized function that also handles
117            case-independence. In UTF-mode with UCP support, this uses the new data
118            to handle characters with more than one other case.
119    
120        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
121            whose other case is greater than 256 are now correctly matched
122            caselessly. Previously, the high codepoint matched the low one, but not
123            vice versa.
124    
125        (d) The processing of \h, \H, \v, and \ in character classes now makes use
126            of the new class addition function, using character lists defined as
127            macros alongside the case definitions of 20 above.
128    
129        (e) Caseless back references now work with characters that have more than
130            one other case.
131    
132        (f) General caseless matching of characters with more than one other case
133            is supported.
134    
135    22. Unicode character properties were updated from Unicode 6.2.0
136    
137    23. Improved CMake support under Windows. Patch by Daniel Richard G.
138    
139    24. Add support for 32-bit character strings, and UTF-32
140    
141    25. Major JIT compiler update (code refactoring and bugfixing).
142        Experimental Sparc 32 support is added.
143    
144    26. Applied a modified version of Daniel Richard G's patch to create
145        pcre.h.generic and config.h.generic by "make" instead of in the
146        PrepareRelease script.
147    
148    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
149        pcre_compile.c when checking for a zero character.
150    
151    28. Introducing a native interface for JIT. Through this interface, the compiled
152        machine code can be directly executed. The purpose of this interface is to
153        provide fast pattern matching, so several sanity checks are not performed.
154        However, feature tests are still performed. The new interface provides
155        1.4x speedup compared to the old one.
156    
157    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
158        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
159        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
160    
161    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
162        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
163        the "old" RFC 2279). Instead, it ended up passing a negative length to
164        pcre_exec().
165    
166    31. Add support for GCC's visibility feature to hide internal functions.
167    
168    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
169        "unknown -C option" after outputting 0 or 1.
170    
171    33. There is now support for generating a code coverage report for the test
172        suite in environments where gcc is the compiler and lcov is installed. This
173        is mainly for the benefit of the developers.
174    
175    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
176        unaddressable using valgrind annotations, allowing valgrind to detect
177        invalid memory accesses. This is mainly for the benefit of the developers.
178    
179    25. (*UTF) can now be used to start a pattern in any of the three libraries.
180    
181    26. Give configure error if --enable-cpp but no C++ compiler found.
182    
183    
184    Version 8.31 06-July-2012
185    -------------------------
186    
187    1.  Fixing a wrong JIT test case and some compiler warnings.
188    
189    2.  Removed a bashism from the RunTest script.
190    
191    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
192        to unsigned type, result still unsigned" that was given by an MS compiler
193        on encountering the code "-sizeof(xxx)".
194    
195    4.  Partial matching support is added to the JIT compiler.
196    
197    5.  Fixed several bugs concerned with partial matching of items that consist
198        of more than one character:
199    
200        (a) /^(..)\1/ did not partially match "aba" because checking references was
201            done on an "all or nothing" basis. This also applied to repeated
202            references.
203    
204        (b) \R did not give a hard partial match if \r was found at the end of the
205            subject.
206    
207        (c) \X did not give a hard partial match after matching one or more
208            characters at the end of the subject.
209    
210        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
211            a partial match for the string "\r".
212    
213        (e) When newline was set to CRLF, the metacharacter "." did not recognize
214            a partial match for a CR character at the end of the subject string.
215    
216    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
217        running pcretest, the text "(JIT)" added to the output whenever JIT is
218        actually used to run the match.
219    
220    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
221        or /S+[+] with a digit between 1 and 7.
222    
223    8.  OP_NOT now supports any UTF character not just single-byte ones.
224    
225    9.  (*MARK) control verb is now supported by the JIT compiler.
226    
227    10. The command "./RunTest list" lists the available tests without actually
228        running any of them. (Because I keep forgetting what they all are.)
229    
230    11. Add PCRE_INFO_MAXLOOKBEHIND.
231    
232    12. Applied a (slightly modified) user-supplied patch that improves performance
233        when the heap is used for recursion (compiled with --disable-stack-for-
234        recursion). Instead of malloc and free for each heap frame each time a
235        logical recursion happens, frames are retained on a chain and re-used where
236        possible. This sometimes gives as much as 30% improvement.
237    
238    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
239        call.
240    
241    14. As documented, (*COMMIT) is now confined to within a positive assertion.
242    
243    15. It is now possible to link pcretest with libedit as an alternative to
244        libreadline.
245    
246    16. (*COMMIT) control verb is now supported by the JIT compiler.
247    
248    17. The Unicode data tables have been updated to Unicode 6.1.0.
249    
250    18. Added --file-list option to pcregrep.
251    
252    19. Added binary file support to pcregrep, including the -a, --binary-files,
253        -I, and --text options.
254    
255    20. The madvise function is renamed for posix_madvise for QNX compatibility
256        reasons. Fixed by Giuseppe D'Angelo.
257    
258    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
259    
260    22. Changed the option for creating symbolic links for 16-bit man pages from
261        -s to -sf so that re-installing does not cause issues.
262    
263    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
264    
265    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
266        matches in certain environments (the workspace was not being correctly
267        retained). Also added to pcre_dfa_exec() a simple plausibility check on
268        some of the workspace data at the beginning of a restart.
269    
270    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
271        was not doing so when it should - probably a typo introduced by SVN 528
272        (change 8.10/14).
273    
274    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
275        \w+ when the character tables indicated that \x{c4} was a word character.
276        There were several related cases, all because the tests for doing a table
277        lookup were testing for characters less than 127 instead of 255.
278    
279    27. If a pattern contains capturing parentheses that are not used in a match,
280        their slots in the ovector are set to -1. For those that are higher than
281        any matched groups, this happens at the end of processing. In the case when
282        there were back references that the ovector was too small to contain
283        (causing temporary malloc'd memory to be used during matching), and the
284        highest capturing number was not used, memory off the end of the ovector
285        was incorrectly being set to -1. (It was using the size of the temporary
286        memory instead of the true size.)
287    
288    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
289        ovector size, it uses memory at the end of the block that it has got.
290    
291    29. Check for an overlong MARK name and give an error at compile time. The
292        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
293    
294    30. JIT compiler update.
295    
296    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
297        Rill for the patch.
298    
299    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
300    
301    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
302    
303    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
304        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
305        was enabled.
306    
307    35. Improve JIT code generation for greedy plus quantifier.
308    
309    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
310        1 to "aa" instead of to an empty string. The bug affected repeated groups
311        that could potentially match an empty string.
312    
313    37. Optimizing single character iterators in JIT.
314    
315    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
316        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
317        codepoints that are too big for the mode are faulted, and in a UTF mode,
318        disallowed codepoints are also faulted.
319    
320    39. If PCRE was compiled with UTF support, in three places in the DFA
321        matcher there was code that should only have been obeyed in UTF mode, but
322        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
323        processing when bytes with values greater than 127 were present. In 16-bit
324        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
325        both cases the values are those that cannot be the first data item in a UTF
326        character. The three items that might have provoked this were recursions,
327        possessively repeated groups, and atomic groups.
328    
329    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
330        and pcregrep, because some OS require shared objects to be explicitly
331        passed to ld, causing the link step to fail if they are not.
332    
333    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
334        mode, patterns that started with \h* or \R* might be incorrectly matched.
335    
336    
337    Version 8.30 04-February-2012
338    -----------------------------
339    
340    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
341        name is defined in ctype.h.
342    
343    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
344        only in quite long subpatterns.
345    
346    3.  Removed the function pcre_info(), which has been obsolete and deprecated
347        since it was replaced by pcre_fullinfo() in February 2000.
348    
349    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
350        match a (*MARK), and the match failed at the start of the subject, a
351        reference to memory before the start of the subject could occur. This bug
352        was introduced by fix 17 of release 8.21.
353    
354    5.  A reference to an unset group with zero minimum repetition was giving
355        totally wrong answers (in non-JavaScript-compatibility mode). For example,
356        /(another)?(\1?)test/ matched against "hello world test". This bug was
357        introduced in release 8.13.
358    
359    6.  Add support for 16-bit character strings (a large amount of work involving
360        many changes and refactorings).
361    
362    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
363        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
364        from a file.
365    
366    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
367        rounding is not applied in this particular case).
368    
369    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
370        if they appear, or are escaped, in patterns.
371    
372    10. Get rid of a number of -Wunused-but-set-variable warnings.
373    
374    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
375        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
376        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
377        also returns the mark "x". This bug applied to capturing parentheses,
378        non-capturing parentheses, and atomic parentheses. It also applied to some
379        assertions.
380    
381    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
382        information out of configure.ac instead of relying on pcre.h.generic, which
383        is not stored in the repository.
384    
385    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
386        -lreadline.
387    
388    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
389    
390    15. Applied Graycode's patch to put the top-level frame on the stack rather
391        than the heap when not using the stack for recursion. This gives a
392        performance improvement in many cases when recursion is not deep.
393    
394    16. Experimental code added to "pcretest -C" to output the stack frame size.
395    
396    
397    Version 8.21 12-Dec-2011
398    ------------------------
399    
400    1.  Updating the JIT compiler.
401    
402    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
403        are added as well.
404    
405    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
406        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
407        calling _pcre_jit_exec. Some extra comments are added.
408    
409    4.  (*MARK) settings inside atomic groups that do not contain any capturing
410        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
411        was introduced by change 18 for 8.20.
412    
413    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
414        ECMA-262 standard.
415    
416    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
417        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
418        This bug was probably introduced by change 9 of 8.13.
419    
420    7.  While fixing 6 above, I noticed that a number of other items were being
421        incorrectly rejected as "not fixed length". This arose partly because newer
422        opcodes had not been added to the fixed-length checking code. I have (a)
423        corrected the bug and added tests for these items, and (b) arranged for an
424        error to occur if an unknown opcode is encountered while checking for fixed
425        length instead of just assuming "not fixed length". The items that were
426        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
427        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
428        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
429    
430    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
431        being incorrectly compiled and would have given unpredicatble results.
432    
433    9.  A possessively repeated subpattern with minimum repeat count greater than
434        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
435        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
436        the first (A) could occur when it should not.
437    
438    10. Add a cast and remove a redundant test from the code.
439    
440    11. JIT should use pcre_malloc/pcre_free for allocation.
441    
442    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
443        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
444        is anything other than /usr, -L is still shown).
445    
446    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
447    
448    14. Perl does not support \N without a following name in a [] class; PCRE now
449        also gives an error.
450    
451    15. If a forward reference was repeated with an upper limit of around 2000,
452        it caused the error "internal error: overran compiling workspace". The
453        maximum number of forward references (including repeats) was limited by the
454        internal workspace, and dependent on the LINK_SIZE. The code has been
455        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
456        the default depends on LINK_SIZE. There is a new upper limit (for safety)
457        of around 200,000 forward references. While doing this, I also speeded up
458        the filling in of repeated forward references.
459    
460    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
461        incorrectly expecting the subject to contain another "a" after the start.
462    
463    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
464        in the match, the SKIP should be ignored. This was not happening; instead
465        the SKIP was being treated as NOMATCH. For patterns such as
466        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
467        tested.
468    
469    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
470        now much more compatible with Perl, in particular in cases where the result
471        is a non-match for a non-anchored pattern. For example, if
472        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
473        "m", where previously it did not return a name. A side effect of this
474        change is that for partial matches, the last encountered mark name is
475        returned, as for non matches. A number of tests that were previously not
476        Perl-compatible have been moved into the Perl-compatible test files. The
477        refactoring has had the pleasing side effect of removing one argument from
478        the match() function, thus reducing its stack requirements.
479    
480    19. If the /S+ option was used in pcretest to study a pattern using JIT,
481        subsequent uses of /S (without +) incorrectly behaved like /S+.
482    
483    21. Retrieve executable code size support for the JIT compiler and fixing
484        some warnings.
485    
486    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
487        not work when the shorter character appeared right at the end of the
488        subject string.
489    
490    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
491        systems.
492    
493    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
494        output it when the /M option is used in pcretest.
495    
496    25. The CheckMan script was not being included in the distribution. Also, added
497        an explicit "perl" to run Perl scripts from the PrepareRelease script
498        because this is reportedly needed in Windows.
499    
500    26. If study data was being save in a file and studying had not found a set of
501        "starts with" bytes for the pattern, the data written to the file (though
502        never used) was taken from uninitialized memory and so caused valgrind to
503        complain.
504    
505    27. Updated RunTest.bat as provided by Sheri Pierce.
506    
507    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
508    
509    29. Computation of memory usage for the table of capturing group names was
510        giving an unnecessarily large value.
511    
512    
513    Version 8.20 21-Oct-2011
514    ------------------------
515    
516    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
517        a POSIX class. After further experiments with Perl, which convinced me that
518        Perl has bugs and confusions, a closing square bracket is no longer allowed
519        in a POSIX name. This bug also affected patterns with classes that started
520        with full stops.
521    
522    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
523        captured substring, but while checking the failing first alternative,
524        substring 1 is temporarily captured. If the output vector supplied to
525        pcre_exec() was not big enough for this capture, the yield of the function
526        was still zero ("insufficient space for captured substrings"). This cannot
527        be totally fixed without adding another stack variable, which seems a lot
528        of expense for a edge case. However, I have improved the situation in cases
529        such as /(a)(b)x|abc/ matched against "abc", where the return code
530        indicates that fewer than the maximum number of slots in the ovector have
531        been set.
532    
533    3.  Related to (2) above: when there are more back references in a pattern than
534        slots in the output vector, pcre_exec() uses temporary memory during
535        matching, and copies in the captures as far as possible afterwards. It was
536        using the entire output vector, but this conflicts with the specification
537        that only 2/3 is used for passing back captured substrings. Now it uses
538        only the first 2/3, for compatibility. This is, of course, another edge
539        case.
540    
541    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
542        main code base, and can be used by building with --enable-jit. When this is
543        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
544        runtime --no-jit option is given.
545    
546    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
547        ovector, the return from the function was zero, implying that there were
548        other matches that did not fit. The correct "exactly full" value is now
549        returned.
550    
551    6.  If a subpattern that was called recursively or as a subroutine contained
552        (*PRUNE) or any other control that caused it to give a non-standard return,
553        invalid errors such as "Error -26 (nested recursion at the same subject
554        position)" or even infinite loops could occur.
555    
556    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
557        computing the minimum length on reaching *ACCEPT, and so ended up with the
558        wrong value of 1 rather than 0. Further investigation indicates that
559        computing a minimum subject length in the presence of *ACCEPT is difficult
560        (think back references, subroutine calls), and so I have changed the code
561        so that no minimum is registered for a pattern that contains *ACCEPT.
562    
563    8.  If (*THEN) was present in the first (true) branch of a conditional group,
564        it was not handled as intended. [But see 16 below.]
565    
566    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
567        Sheri Pierce.
568    
569    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
570        the first byte in a match must be "a".
571    
572    11. Change 17 for 8.13 increased the recursion depth for patterns like
573        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
574        pattern contains any instances of (*THEN). If it does not, the old
575        optimizations are restored. It would be nice to do this on a per-group
576        basis, but at the moment that is not feasible.
577    
578    12. In some environments, the output of pcretest -C is CRLF terminated. This
579        broke RunTest's code that checks for the link size. A single white space
580        character after the value is now allowed for.
581    
582    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
583        For "fr", it uses the Windows-specific input and output files.
584    
585    14. If (*THEN) appeared in a group that was called recursively or as a
586        subroutine, it did not work as intended. [But see next item.]
587    
588    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
589        pattern fragments (but not containing any | characters). If A and B are
590        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
591        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
592        D. In other words, Perl considers parentheses that do not contain any |
593        characters to be part of a surrounding alternative, whereas PCRE was
594        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
595        differently. PCRE now behaves in the same way as Perl, except in the case
596        of subroutine/recursion calls such as (?1) which have in any case always
597        been different (but PCRE had them first :-).
598    
599    16. Related to 15 above: Perl does not treat the | in a conditional group as
600        creating alternatives. Such a group is treated in the same way as an
601        ordinary group without any | characters when processing (*THEN). PCRE has
602        been changed to match Perl's behaviour.
603    
604    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
605        RunGrepTest script failed.
606    
607    18. Change 22 for version 13 caused atomic groups to use more stack. This is
608        inevitable for groups that contain captures, but it can lead to a lot of
609        stack use in large patterns. The old behaviour has been restored for atomic
610        groups that do not contain any capturing parentheses.
611    
612    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
613        suppress the check for a minimum subject length at run time. (If it was
614        given to pcre_exec() or pcre_dfa_exec() it did work.)
615    
616    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
617        fail to work when decoding hex characters in data strings in EBCDIC
618        environments.
619    
620    21. It appears that in at least one Mac OS environment, the isxdigit() function
621        is implemented as a macro that evaluates to its argument more than once,
622        contravening the C 90 Standard (I haven't checked a later standard). There
623        was an instance in pcretest which caused it to go wrong when processing
624        \x{...} escapes in subject strings. The has been rewritten to avoid using
625        things like p++ in the argument of isxdigit().
626    
627    
628    Version 8.13 16-Aug-2011
629    ------------------------
630    
631    1.  The Unicode data tables have been updated to Unicode 6.0.0.
632    
633    2.  Two minor typos in pcre_internal.h have been fixed.
634    
635    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
636        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
637        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
638    
639    4.  There were a number of related bugs in the code for matching backrefences
640        caselessly in UTF-8 mode when codes for the characters concerned were
641        different numbers of bytes. For example, U+023A and U+2C65 are an upper
642        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
643        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
644        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
645        2-byte code at the end of the subject (it thought there wasn't enough data
646        left).
647    
648    5.  Comprehensive information about what went wrong is now returned by
649        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
650        as the output vector has at least 2 elements. The offset of the start of
651        the failing character and a reason code are placed in the vector.
652    
653    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
654        now returned is for the first byte of the failing character, instead of the
655        last byte inspected. This is an incompatible change, but I hope it is small
656        enough not to be a problem. It makes the returned offset consistent with
657        pcre_exec() and pcre_dfa_exec().
658    
659    7.  pcretest now gives a text phrase as well as the error number when
660        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
661        failure, the offset and reason code are output.
662    
663    8.  When \R was used with a maximizing quantifier it failed to skip backwards
664        over a \r\n pair if the subsequent match failed. Instead, it just skipped
665        back over a single character (\n). This seems wrong (because it treated the
666        two characters as a single entity when going forwards), conflicts with the
667        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
668        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
669        has been changed.
670    
671    9.  Some internal refactoring has changed the processing so that the handling
672        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
673        time (the PCRE_DOTALL option was changed this way some time ago: version
674        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
675        which was always a bit of a fudge. It also means that there is one less
676        argument for the match() function, which reduces its stack requirements
677        slightly. This change also fixes an incompatibility with Perl: the pattern
678        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
679    
680    10. More internal refactoring has drastically reduced the number of recursive
681        calls to match() for possessively repeated groups such as (abc)++ when
682        using pcre_exec().
683    
684    11. While implementing 10, a number of bugs in the handling of groups were
685        discovered and fixed:
686    
687        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
688        (a|)*(?1) gave a compile-time internal error.
689        ((a|)+)+  did not notice that the outer group could match an empty string.
690        (^a|^)+   was not marked as anchored.
691        (.*a|.*)+ was not marked as matching at start or after a newline.
692    
693    12. Yet more internal refactoring has removed another argument from the match()
694        function. Special calls to this function are now indicated by setting a
695        value in a variable in the "match data" data block.
696    
697    13. Be more explicit in pcre_study() instead of relying on "default" for
698        opcodes that mean there is no starting character; this means that when new
699        ones are added and accidentally left out of pcre_study(), testing should
700        pick them up.
701    
702    14. The -s option of pcretest has been documented for ages as being an old
703        synonym of -m (show memory usage). I have changed it to mean "force study
704        for every regex", that is, assume /S for every regex. This is similar to -i
705        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
706        using it. It makes it easier to run collections of tests with and without
707        study enabled, and thereby test pcre_study() more easily. All the standard
708        tests are now run with and without -s (but some patterns can be marked as
709        "never study" - see 20 below).
710    
711    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
712        restoration of the capturing data to the outer values was not happening
713        correctly.
714    
715    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
716        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
717        pattern had matched an empty string, and so incorrectly returned a no
718        match.
719    
720    17. There was optimizing code for the last branch of non-capturing parentheses,
721        and also for the obeyed branch of a conditional subexpression, which used
722        tail recursion to cut down on stack usage. Unfortunately, now that there is
723        the possibility of (*THEN) occurring in these branches, tail recursion is
724        no longer possible because the return has to be checked for (*THEN). These
725        two optimizations have therefore been removed. [But see 8.20/11 above.]
726    
727    18. If a pattern containing \R was studied, it was assumed that \R always
728        matched two bytes, thus causing the minimum subject length to be
729        incorrectly computed because \R can also match just one byte.
730    
731    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
732        was incorrectly computed.
733    
734    20. If /S is present twice on a test pattern in pcretest input, it now
735        *disables* studying, thereby overriding the use of -s on the command line
736        (see 14 above). This is necessary for one or two tests to keep the output
737        identical in both cases.
738    
739    21. When (*ACCEPT) was used in an assertion that matched an empty string and
740        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
741    
742    22. When an atomic group that contained a capturing parenthesis was
743        successfully matched, but the branch in which it appeared failed, the
744        capturing was not being forgotten if a higher numbered group was later
745        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
746        group 1 to "a", when in fact it should be unset. This applied to multi-
747        branched capturing and non-capturing groups, repeated or not, and also to
748        positive assertions (capturing in negative assertions does not happen
749        in PCRE) and also to nested atomic groups.
750    
751    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
752        subject after a captured substring, to make it easier to tell which of a
753        number of identical substrings has been captured.
754    
755    24. The way atomic groups are processed by pcre_exec() has been changed so that
756        if they are repeated, backtracking one repetition now resets captured
757        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
758        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
759        "aaa". Previously, it would have been "a". As part of this code
760        refactoring, the way recursive calls are handled has also been changed.
761    
762    25. If an assertion condition captured any substrings, they were not passed
763        back unless some other capturing happened later. For example, if
764        (?(?=(a))a) was matched against "a", no capturing was returned.
765    
766    26. When studying a pattern that contained subroutine calls or assertions,
767        the code for finding the minimum length of a possible match was handling
768        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
769        group 1 called group 2 while simultaneously a separate group 2 called group
770        1). A stack overflow occurred in this case. I have fixed this by limiting
771        the recursion depth to 10.
772    
773    27. Updated RunTest.bat in the distribution to the version supplied by Tom
774        Fortmann. This supports explicit test numbers on the command line, and has
775        argument validation and error reporting.
776    
777    28. An instance of \X with an unlimited repeat could fail if at any point the
778        first character it looked at was a mark character.
779    
780    29. Some minor code refactoring concerning Unicode properties and scripts
781        should reduce the stack requirement of match() slightly.
782    
783    30. Added the '=' option to pcretest to check the setting of unused capturing
784        slots at the end of the pattern, which are documented as being -1, but are
785        not included in the return count.
786    
787    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
788        compiled something random. Now it gives a compile-time error (as does
789        Perl).
790    
791    32. A *MARK encountered during the processing of a positive assertion is now
792        recorded and passed back (compatible with Perl).
793    
794    33. If --only-matching or --colour was set on a pcregrep call whose pattern
795        had alternative anchored branches, the search for a second match in a line
796        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
797        matched the line "0102" twice. The same bug affected patterns that started
798        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
799        twice.
800    
801    34. Previously, PCRE did not allow quantification of assertions. However, Perl
802        does, and because of capturing effects, quantifying parenthesized
803        assertions may at times be useful. Quantifiers are now allowed for
804        parenthesized assertions.
805    
806    35. A minor code tidy in pcre_compile() when checking options for \R usage.
807    
808    36. \g was being checked for fancy things in a character class, when it should
809        just be a literal "g".
810    
811    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
812        appearance of a nested POSIX class supersedes an apparent external class.
813        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
814        unescaped square brackets may also appear as part of class names. For
815        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
816        more like Perl. (But see 8.20/1 above.)
817    
818    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
819        was because it thought it was \N{name}, which is not supported).
820    
821    39. Add minix to OS list not supporting the -S option in pcretest.
822    
823    40. PCRE tries to detect cases of infinite recursion at compile time, but it
824        cannot analyze patterns in sufficient detail to catch mutual recursions
825        such as ((?1))((?2)). There is now a runtime test that gives an error if a
826        subgroup is called recursively as a subpattern for a second time at the
827        same position in the subject string. In previous releases this might have
828        been caught by the recursion limit, or it might have run out of stack.
829    
830    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
831        happen only once. PCRE was, however incorrectly giving a compile time error
832        "recursive call could loop indefinitely" because it cannot analyze the
833        pattern in sufficient detail. The compile time test no longer happens when
834        PCRE is compiling a conditional subpattern, but actual runaway loops are
835        now caught at runtime (see 40 above).
836    
837    42. It seems that Perl allows any characters other than a closing parenthesis
838        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
839        has been changed to be the same.
840    
841    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
842        as not to get warnings when autogen.sh is called. Also changed
843        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
844    
845    44. To help people who use pcregrep to scan files containing exceedingly long
846        lines, the following changes have been made:
847    
848        (a) The default value of the buffer size parameter has been increased from
849            8K to 20K. (The actual buffer used is three times this size.)
850    
851        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
852            PCRE is built.
853    
854        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
855            to be set at run time.
856    
857        (d) Numerical values in pcregrep options can be followed by K or M, for
858            example --buffer-size=50K.
859    
860        (e) If a line being scanned overflows pcregrep's buffer, an error is now
861            given and the return code is set to 2.
862    
863    45. Add a pointer to the latest mark to the callout data block.
864    
865    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
866        partial match of an empty string instead of no match. This was specific to
867        the use of ".".
868    
869    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
870        complete match instead of a partial match. This bug was dependent on both
871        the PCRE_UTF8 and PCRE_DOTALL options being set.
872    
873    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
874        starting byte set, because \b was not being ignored.
875    
876    
877    Version 8.12 15-Jan-2011
878    ------------------------
879    
880    1.  Fixed some typos in the markup of the man pages, and wrote a script that
881        checks for such things as part of the documentation building process.
882    
883    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
884        --match-limit and --recursion-limit options (added for 8.11). In
885        particular, this made one of the standard tests fail. (The integer value
886        went into the wrong half of a long int.)
887    
888    3.  If the --colour option was given to pcregrep with -v (invert match), it
889        did strange things, either producing crazy output, or crashing. It should,
890        of course, ignore a request for colour when reporting lines that do not
891        match.
892    
893    4.  Another pcregrep bug caused similar problems if --colour was specified with
894        -M (multiline) and the pattern match finished with a line ending.
895    
896    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
897        matched in multiline mode, the following line was shown as part of the
898        match. This seems wrong, so I have changed it.
899    
900    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
901        the check for further matches in the same line (so they could be coloured)
902        to overrun the end of the current line. If another match was found, it was
903        incorrectly shown (and then shown again when found in the next line).
904    
905    7.  If pcregrep was compiled under Windows, there was a reference to the
906        function pcregrep_exit() before it was defined. I am assuming this was
907        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
908        reported by a user. I've moved the definition above the reference.
909    
910    
911    Version 8.11 10-Dec-2010
912    ------------------------
913    
914    1.  (*THEN) was not working properly if there were untried alternatives prior
915        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
916        backtracked to try for "b" instead of moving to the next alternative branch
917        at the same level (in this case, to look for "c"). The Perl documentation
918        is clear that when (*THEN) is backtracked onto, it goes to the "next
919        alternative in the innermost enclosing group".
920    
921    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
922        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
923        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
924        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
925        (*THEN).
926    
927    3.  If \s appeared in a character class, it removed the VT character from
928        the class, even if it had been included by some previous item, for example
929        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
930        of \s, but is part of the POSIX "space" class.)
931    
932    4.  A partial match never returns an empty string (because you can always
933        match an empty string at the end of the subject); however the checking for
934        an empty string was starting at the "start of match" point. This has been
935        changed to the "earliest inspected character" point, because the returned
936        data for a partial match starts at this character. This means that, for
937        example, /(?<=abc)def/ gives a partial match for the subject "abc"
938        (previously it gave "no match").
939    
940    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
941        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
942        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
943        has an implication that the given string is incomplete (because a partial
944        match is preferred over a full match). For this reason, these items now
945        give a partial match in this situation. [Aside: previously, the one case
946        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
947        match rather than a full match, which was wrong by the old rules, but is
948        now correct.]
949    
950    6.  There was a bug in the handling of #-introduced comments, recognized when
951        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
952        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
953        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
954        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
955        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
956        places in pcre_compile().
957    
958    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
959        comments when looking ahead for named forward references to subpatterns,
960        the only newline sequence it recognized was NL. It now handles newlines
961        according to the set newline convention.
962    
963    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
964        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
965        cater for a lack of strerror(). These oversights have been fixed.
966    
967    9.  Added --match-limit and --recursion-limit to pcregrep.
968    
969    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
970    
971    11. When the -o option was used, pcregrep was setting a return code of 1, even
972        when matches were found, and --line-buffered was not being honoured.
973    
974    12. Added an optional parentheses number to the -o and --only-matching options
975        of pcregrep.
976    
977    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
978        can match an empty string. The code to do it in pcretest and pcredemo
979        needed fixing:
980    
981        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
982            only one byte after an empty string match just before CRLF (this case
983            just got forgotten; "any" and "anycrlf" were OK).
984    
985        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
986            mode when an empty string match preceded an ASCII character followed by
987            a non-ASCII character. (The code for advancing by one character rather
988            than one byte was nonsense.)
989    
990        (c) The pcredemo.c sample program did not have any code at all to handle
991            the cases when CRLF is a valid newline sequence.
992    
993    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
994        as a starting offset was within the subject string. There is now a new
995        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
996        negative or greater than the length of the string. In order to test this,
997        pcretest is extended to allow the setting of negative starting offsets.
998    
999    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1000        starting offset points to the beginning of a UTF-8 character was
1001        unnecessarily clumsy. I tidied it up.
1002    
1003    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1004        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1005    
1006    17. Nobody had reported that the --include_dir option, which was added in
1007        release 7.7 should have been called --include-dir (hyphen, not underscore)
1008        for compatibility with GNU grep. I have changed it to --include-dir, but
1009        left --include_dir as an undocumented synonym, and the same for
1010        --exclude-dir, though that is not available in GNU grep, at least as of
1011        release 2.5.4.
1012    
1013    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1014        characters from a string of bytes) have been redefined so as not to use
1015        loops, in order to improve performance in some environments. At the same
1016        time, I abstracted some of the common code into auxiliary macros to save
1017        repetition (this should not affect the compiled code).
1018    
1019    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1020        compile-time error is now given if \c is not followed by an ASCII
1021        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1022        different, and any byte value is allowed.)
1023    
1024    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1025        START_OPTIMIZE option, which is now allowed at compile time - but just
1026        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1027        to pcregrep and other applications that have no direct access to PCRE
1028        options. The new /Y option in pcretest sets this option when calling
1029        pcre_compile().
1030    
1031    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1032        back references. Groups containing recursive back references were forced to
1033        be atomic by that change, but in the case of named groups, the amount of
1034        memory required was incorrectly computed, leading to "Failed: internal
1035        error: code overflow". This has been fixed.
1036    
1037    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1038        pcretest.c, to avoid build problems in some Borland environments.
1039    
1040    
1041    Version 8.10 25-Jun-2010
1042    ------------------------
1043    
1044    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1045        THEN.
1046    
1047    2.  (*ACCEPT) was not working when inside an atomic group.
1048    
1049    3.  Inside a character class, \B is treated as a literal by default, but
1050        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1051        causes the error). The code is unchanged, but I tidied the documentation.
1052    
1053    4.  Inside a character class, PCRE always treated \R and \X as literals,
1054        whereas Perl faults them if its -w option is set. I have changed PCRE so
1055        that it faults them when PCRE_EXTRA is set.
1056    
1057    5.  Added support for \N, which always matches any character other than
1058        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1059    
1060    6.  When compiling pcregrep with newer versions of gcc which may have
1061        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1062        declared with attribute warn_unused_result" were given. Just casting the
1063        result to (void) does not stop the warnings; a more elaborate fudge is
1064        needed. I've used a macro to implement this.
1065    
1066    7.  Minor change to pcretest.c to avoid a compiler warning.
1067    
1068    8.  Added four artifical Unicode properties to help with an option to make
1069        \s etc use properties (see next item). The new properties are: Xan
1070        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1071    
1072    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1073        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1074        this option. Modified pcretest to add /W to test this facility. Added
1075        REG_UCP to make it available via the POSIX interface.
1076    
1077    10. Added --line-buffered to pcregrep.
1078    
1079    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1080        studied, and the match started with a letter with a code point greater than
1081        127 whose first byte was different to the first byte of the other case of
1082        the letter, the other case of this starting letter was not recognized
1083        (#976).
1084    
1085    12. If a pattern that was studied started with a repeated Unicode property
1086        test, for example, \p{Nd}+, there was the theoretical possibility of
1087        setting up an incorrect bitmap of starting bytes, but fortunately it could
1088        not have actually happened in practice until change 8 above was made (it
1089        added property types that matched character-matching opcodes).
1090    
1091    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1092        possible starting bytes for non-anchored patterns.
1093    
1094    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1095        \R, and also a number of cases that involve Unicode properties, both
1096        explicit and implicit when PCRE_UCP is set.
1097    
1098    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1099        input, it could crash or give wrong results if characters with values
1100        greater than 0xc0 were present in the subject string. (Detail: it assumed
1101        UTF-8 input when processing these items.)
1102    
1103    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1104        size_t is 64-bit (#991).
1105    
1106    17. Added a check for running out of memory when PCRE is compiled with
1107        --disable-stack-for-recursion (#990).
1108    
1109    18. If the last data line in a file for pcretest does not have a newline on
1110        the end, a newline was missing in the output.
1111    
1112    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1113        less than 128) in its various bitmaps. However, there is a facility for
1114        generating tables according to the current locale when PCRE is compiled. It
1115        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1116        characters, are recognized by isspace() and therefore were getting set in
1117        these tables, and indeed these tables seem to approximate to ISO 8859. This
1118        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1119        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1120        which of course cannot start UTF-8 characters. I have changed the code so
1121        that only real ASCII characters (less than 128) and the correct starting
1122        bytes for UTF-8 encodings are set for characters greater than 127 when in
1123        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1124        altogether.)
1125    
1126    20. Added the /T option to pcretest so as to be able to run tests with non-
1127        standard character tables, thus making it possible to include the tests
1128        used for 19 above in the standard set of tests.
1129    
1130    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1131        reference to a subpattern the other side of a comment that contains an
1132        opening parenthesis caused either an internal compiling error, or a
1133        reference to the wrong subpattern.
1134    
1135    
1136    Version 8.02 19-Mar-2010
1137    ------------------------
1138    
1139    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1140    
1141    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1142        configured.
1143    
1144    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1145        original author of that file, following a query about its status.
1146    
1147    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1148        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1149    
1150    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1151        quantifier applied to a forward-referencing subroutine call, could compile
1152        incorrect code or give the error "internal error: previously-checked
1153        referenced subpattern not found".
1154    
1155    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1156        variables to point to external functions. For these systems, therefore,
1157        pcre_malloc etc. are now initialized to local functions that call the
1158        relevant global functions.
1159    
1160    7.  There were two entries missing in the vectors called coptable and poptable
1161        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1162        I've fixed the data, and added a kludgy way of testing at compile time that
1163        the lengths are correct (equal to the number of opcodes).
1164    
1165    8.  Following on from 7, I added a similar kludge to check the length of the
1166        eint vector in pcreposix.c.
1167    
1168    9.  Error texts for pcre_compile() are held as one long string to avoid too
1169        much relocation at load time. To find a text, the string is searched,
1170        counting zeros. There was no check for running off the end of the string,
1171        which could happen if a new error number was added without updating the
1172        string.
1173    
1174    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1175    
1176    11. \K was not working if it appeared in an atomic group or in a group that
1177        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1178        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1179        the assertion is positive, but not if it is negative.
1180    
1181    12. Change 11 fortuitously reduced the size of the stack frame used in the
1182        "match()" function of pcre_exec.c by one pointer. Forthcoming
1183        implementation of support for (*MARK) will need an extra pointer on the
1184        stack; I have reserved it now, so that the stack frame size does not
1185        decrease.
1186    
1187    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1188        item in branch that calls a recursion is a subroutine call - as in the
1189        second branch in the above example - was incorrectly given the compile-
1190        time error "recursive call could loop indefinitely" because pcre_compile()
1191        was not correctly checking the subroutine for matching a non-empty string.
1192    
1193    14. The checks for overrunning compiling workspace could trigger after an
1194        overrun had occurred. This is a "should never occur" error, but it can be
1195        triggered by pathological patterns such as hundreds of nested parentheses.
1196        The checks now trigger 100 bytes before the end of the workspace.
1197    
1198    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1199    
1200    
1201    Version 8.01 19-Jan-2010
1202    ------------------------
1203    
1204    1.  If a pattern contained a conditional subpattern with only one branch (in
1205        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1206        computed the wrong minimum data length (which is of course zero for such
1207        subpatterns). This could cause incorrect "no match" results.
1208    
1209    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1210        the pattern is reset in the first branch, pcre_compile() failed with
1211        "internal error: code overflow at offset...". This happened only when
1212        the reset was to the original external option setting. (An optimization
1213        abstracts leading options settings into an external setting, which was the
1214        cause of this.)
1215    
1216    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1217        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1218        assertion pattern did not match (meaning that the assertion was true), it
1219        was incorrectly treated as false if the SKIP had been reached during the
1220        matching. This also applied to assertions used as conditions.
1221    
1222    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1223        assertion subpattern, including such a pattern used as a condition,
1224        unpredictable results occurred, instead of the error return
1225        PCRE_ERROR_DFA_UITEM.
1226    
1227    5.  The C++ GlobalReplace function was not working like Perl for the special
1228        situation when an empty string is matched. It now does the fancy magic
1229        stuff that is necessary.
1230    
1231    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1232        removed. (These were left over from very, very early versions of PCRE.)
1233    
1234    7.  Some cosmetic changes to the code to make life easier when compiling it
1235        as part of something else:
1236    
1237        (a) Change DEBUG to PCRE_DEBUG.
1238    
1239        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1240            called "current" as "current_branch", to prevent a collision with the
1241            Linux macro when compiled as a kernel module.
1242    
1243        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1244            prevent a collision with the Linux macro when compiled as a kernel
1245            module.
1246    
1247    8.  In pcre_compile() there are some checks for integer overflows that used to
1248        cast potentially large values to (double). This has been changed to that
1249        when building, a check for int64_t is made, and if it is found, it is used
1250        instead, thus avoiding the use of floating point arithmetic. (There is no
1251        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1252        double.
1253    
1254    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1255        2005 (difference between two addresses compared to an unsigned value).
1256    
1257    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1258        custom one, because of the following reported problem in Windows:
1259    
1260          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1261              under Win32.
1262          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1263              therefore missing the function definition.
1264          - The compiler thus generates a "C" signature for the test function.
1265          - The linker fails to find the "C" function.
1266          - PCRE fails to configure if asked to do so against libbz2.
1267    
1268    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1269        messages were output:
1270    
1271          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1272          rerunning libtoolize, to keep the correct libtool macros in-tree.
1273          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1274    
1275        I have done both of these things.
1276    
1277    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1278        most of the time, it *can* run out if it is given a pattern that contains a
1279        runaway infinite recursion. I updated the discussion in the pcrestack man
1280        page.
1281    
1282    13. Now that we have gone to the x.xx style of version numbers, the minor
1283        version may start with zero. Using 08 or 09 is a bad idea because users
1284        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1285        interpreted as invalid octal numbers. I've updated the previous comment in
1286        configure.ac, and also added a check that gives an error if 08 or 09 are
1287        used.
1288    
1289    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1290        causing partial matching to fail when the end of the subject matched \W
1291        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1292    
1293    15. There were some discrepancies between the declarations in pcre_internal.h
1294        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1295        their definitions. The declarations used "const uschar *" and the
1296        definitions used USPTR. Even though USPTR is normally defined as "const
1297        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1298        reported that: "This difference in casting confuses some C++ compilers, for
1299        example, SunCC recognizes above declarations as different functions and
1300        generates broken code for hbpcre." I have changed the declarations to use
1301        USPTR.
1302    
1303    16. GNU libtool is named differently on some systems. The autogen.sh script now
1304        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1305        (FreeBSD).
1306    
1307    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1308        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1309        comment: "Figure out how to create a longlong from a string: strtoll and
1310        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1311        instance, but it only takes 2 args instead of 3!"
1312    
1313    18. A subtle bug concerned with back references has been fixed by a change of
1314        specification, with a corresponding code fix. A pattern such as
1315        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1316        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1317        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1318        same bug. Such groups have to be quantified to be useful, or contained
1319        inside another quantified group. (If there's no repetition, the reference
1320        can never match.) The problem arises because, having left the group and
1321        moved on to the rest of the pattern, a later failure that backtracks into
1322        the group uses the captured value from the final iteration of the group
1323        rather than the correct earlier one. I have fixed this in PCRE by forcing
1324        any group that contains a reference to itself to be an atomic group; that
1325        is, there cannot be any backtracking into it once it has completed. This is
1326        similar to recursive and subroutine calls.
1327    
1328    
1329    Version 8.00 19-Oct-09
1330    ----------------------
1331    
1332    1.  The table for translating pcre_compile() error codes into POSIX error codes
1333        was out-of-date, and there was no check on the pcre_compile() error code
1334        being within the table. This could lead to an OK return being given in
1335        error.
1336    
1337    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1338        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1339        in a Windows environment.
1340    
1341    3.  The pcregrep --count option prints the count for each file even when it is
1342        zero, as does GNU grep. However, pcregrep was also printing all files when
1343        --files-with-matches was added. Now, when both options are given, it prints
1344        counts only for those files that have at least one match. (GNU grep just
1345        prints the file name in this circumstance, but including the count seems
1346        more useful - otherwise, why use --count?) Also ensured that the
1347        combination -clh just lists non-zero counts, with no names.
1348    
1349    4.  The long form of the pcregrep -F option was incorrectly implemented as
1350        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1351        but it seems right to fix it, and I didn't think it was worth preserving
1352        the old behaviour.
1353    
1354    5.  The command line items --regex=pattern and --regexp=pattern were not
1355        recognized by pcregrep, which required --regex pattern or --regexp pattern
1356        (with a space rather than an '='). The man page documented the '=' forms,
1357        which are compatible with GNU grep; these now work.
1358    
1359    6.  No libpcreposix.pc file was created for pkg-config; there was just
1360        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1361    
1362    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1363        when UCP support is not needed, by modifying the Python script that
1364        generates it from Unicode data files. This should not matter if the module
1365        is correctly used as a library, but I received one complaint about 50K of
1366        unwanted data. My guess is that the person linked everything into his
1367        program rather than using a library. Anyway, it does no harm.
1368    
1369    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1370        was a minimum greater than 1 for a wide character in a possessive
1371        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1372        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1373        character. Chaos in the form of incorrect output or a compiling loop could
1374        result.
1375    
1376    9.  The restrictions on what a pattern can contain when partial matching is
1377        requested for pcre_exec() have been removed. All patterns can now be
1378        partially matched by this function. In addition, if there are at least two
1379        slots in the offset vector, the offset of the earliest inspected character
1380        for the match and the offset of the end of the subject are set in them when
1381        PCRE_ERROR_PARTIAL is returned.
1382    
1383    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1384        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1385        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1386        and may be more useful for multi-segment matching.
1387    
1388    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1389        used to be given if ever the end of the subject was reached; now it is
1390        given only if matching could not proceed because another character was
1391        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1392        string "Z", which now yields "no match" instead of "partial match". In the
1393        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1394        final character ended with (*FAIL).
1395    
1396    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1397        if the pattern had a "must contain" character that was already found in the
1398        earlier partial match, unless partial matching was again requested. For
1399        example, with the pattern /dog.(body)?/, the "must contain" character is
1400        "g". If the first part-match was for the string "dog", restarting with
1401        "sbody" failed. This bug has been fixed.
1402    
1403    13. The string returned by pcre_dfa_exec() after a partial match has been
1404        changed so that it starts at the first inspected character rather than the
1405        first character of the match. This makes a difference only if the pattern
1406        starts with a lookbehind assertion or \b or \B (\K is not supported by
1407        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1408        matching functions compatible, and I think it's the right thing to do.
1409    
1410    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1411        so that the demonstration program is easily available in environments where
1412        PCRE has not been installed from source.
1413    
1414    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1415        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1416        library.
1417    
1418    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1419        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1420        is not the first non-POSIX option to be added. Clearly some people find
1421        these options useful.
1422    
1423    17. If a caller to the POSIX matching function regexec() passes a non-zero
1424        value for nmatch with a NULL value for pmatch, the value of
1425        nmatch is forced to zero.
1426    
1427    18. RunGrepTest did not have a test for the availability of the -u option of
1428        the diff command, as RunTest does. It now checks in the same way as
1429        RunTest, and also checks for the -b option.
1430    
1431    19. If an odd number of negated classes containing just a single character
1432        interposed, within parentheses, between a forward reference to a named
1433        subpattern and the definition of the subpattern, compilation crashed with
1434        an internal error, complaining that it could not find the referenced
1435        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1436        [The bug was that it was starting one character too far in when skipping
1437        over the character class, thus treating the ] as data rather than
1438        terminating the class. This meant it could skip too much.]
1439    
1440    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1441        /g option in pcretest when the pattern contains \K, which makes it possible
1442        to have an empty string match not at the start, even when the pattern is
1443        anchored. Updated pcretest and pcredemo to use this option.
1444    
1445    21. If the maximum number of capturing subpatterns in a recursion was greater
1446        than the maximum at the outer level, the higher number was returned, but
1447        with unset values at the outer level. The correct (outer level) value is
1448        now given.
1449    
1450    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1451        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1452        make it do so. The string so far is captured, making this feature
1453        compatible with Perl.
1454    
1455    23. The tests have been re-organized, adding tests 11 and 12, to make it
1456        possible to check the Perl 5.10 features against Perl 5.10.
1457    
1458    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1459        pattern matches a fixed length string. PCRE did not allow this; now it
1460        does. Neither allows recursion.
1461    
1462    25. I finally figured out how to implement a request to provide the minimum
1463        length of subject string that was needed in order to match a given pattern.
1464        (It was back references and recursion that I had previously got hung up
1465        on.) This code has now been added to pcre_study(); it finds a lower bound
1466        to the length of subject needed. It is not necessarily the greatest lower
1467        bound, but using it to avoid searching strings that are too short does give
1468        some useful speed-ups. The value is available to calling programs via
1469        pcre_fullinfo().
1470    
1471    26. While implementing 25, I discovered to my embarrassment that pcretest had
1472        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1473        study optimizations had never been tested with that matching function.
1474        Oops. What is worse, even when it was passed study data, there was a bug in
1475        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1476        were also very few tests of studied patterns with pcre_dfa_exec().
1477    
1478    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1479        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1480        on the other side of the coin, they are no longer allowed to have different
1481        names, because these cannot be distinguished in PCRE, and this has caused
1482        confusion. (This is a difference from Perl.)
1483    
1484    28. When duplicate subpattern names are present (necessarily with different
1485        numbers, as required by 27 above), and a test is made by name in a
1486        conditional pattern, either for a subpattern having been matched, or for
1487        recursion in such a pattern, all the associated numbered subpatterns are
1488        tested, and the overall condition is true if the condition is true for any
1489        one of them. This is the way Perl works, and is also more like the way
1490        testing by number works.
1491    
1492    
1493    Version 7.9 11-Apr-09
1494    ---------------------
1495    
1496    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1497        (pcretest), all targets were linked against these libraries. This included
1498        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1499        libraries. This caused unwanted dependencies to be created. This problem
1500        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1501        pcretest is linked with readline.
1502    
1503    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1504        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1505        moved outside it again, because FALSE and TRUE are already defined in AIX,
1506        but BOOL is not.
1507    
1508    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1509        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1510    
1511    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1512        hyphens) following file names and line numbers when outputting matching
1513        lines. This is not true; no spaces are inserted. I have also clarified the
1514        wording for the --colour (or --color) option.
1515    
1516    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1517        was not coloured; this is different to GNU grep, so I have changed it to be
1518        the same.
1519    
1520    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1521        each matching line was coloured. Now it goes on to look for further matches
1522        of any of the test patterns, which is the same behaviour as GNU grep.
1523    
1524    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1525        doesn't make sense to accept an empty string match in pcregrep, so I have
1526        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1527        seems to be how GNU grep behaves.
1528    
1529    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1530        start or after a newline", because the conditional assertion was not being
1531        correctly handled. The rule now is that both the assertion and what follows
1532        in the first alternative must satisfy the test.
1533    
1534    9.  If auto-callout was enabled in a pattern with a conditional group whose
1535        condition was an assertion, PCRE could crash during matching, both with
1536        pcre_exec() and pcre_dfa_exec().
1537    
1538    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1539        used for matching.
1540    
1541    11. Unicode property support in character classes was not working for
1542        characters (bytes) greater than 127 when not in UTF-8 mode.
1543    
1544    12. Added the -M command line option to pcretest.
1545    
1546    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1547    
1548    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1549    
1550    16. Added comments and documentation about mis-use of no_arg in the C++
1551        wrapper.
1552    
1553    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1554        from Martin Jerabek that uses macro names for all relevant character and
1555        string constants.
1556    
1557    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1558        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1559        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1560        these, but not everybody uses configure.
1561    
1562    19. A conditional group that had only one branch was not being correctly
1563        recognized as an item that could match an empty string. This meant that an
1564        enclosing group might also not be so recognized, causing infinite looping
1565        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1566        with the subject "ab", where knowledge that the repeated group can match
1567        nothing is needed in order to break the loop.
1568    
1569    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1570        exec(), but without supplying a callout function, matching went wrong.
1571    
1572    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1573        leak if the size of the offset vector was greater than 30. When the vector
1574        is smaller, the saved offsets during recursion go onto a local stack
1575        vector, but for larger vectors malloc() is used. It was failing to free
1576        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1577        error, in fact).
1578    
1579    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1580        heapframe that is used only when UTF-8 support is enabled. This caused no
1581        problem, but was untidy.
1582    
1583    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1584        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1585        included within another project.
1586    
1587    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1588        slightly modified by me:
1589    
1590          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1591              not building pcregrep.
1592    
1593          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1594              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1595    
1596    25. Forward references, both numeric and by name, in patterns that made use of
1597        duplicate group numbers, could behave incorrectly or give incorrect errors,
1598        because when scanning forward to find the reference group, PCRE was not
1599        taking into account the duplicate group numbers. A pattern such as
1600        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1601    
1602    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1603        the feature of a custom pointer more persuasive (as requested by a user).
1604    
1605    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1606        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1607        pre-defined.
1608    
1609    28. Added support for (*UTF8) at the start of a pattern.
1610    
1611    29. Arrange for flags added by the "release type" setting in CMake to be shown
1612        in the configuration summary.
1613    
1614    
1615    Version 7.8 05-Sep-08
1616    ---------------------
1617    
1618    1.  Replaced UCP searching code with optimized version as implemented for Ad
1619        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1620        stage table and inline lookup instead of a function, giving speed ups of 2
1621        to 5 times on some simple patterns that I tested. Permission was given to
1622        distribute the MultiStage2.py script that generates the tables (it's not in
1623        the tarball, but is in the Subversion repository).
1624    
1625    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1626        scripts.
1627    
1628    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1629        a group with a zero qualifier. The result of the study could be incorrect,
1630        or the function might crash, depending on the pattern.
1631    
1632    4.  Caseless matching was not working for non-ASCII characters in back
1633        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1634        It now works when Unicode Property Support is available.
1635    
1636    5.  In pcretest, an escape such as \x{de} in the data was always generating
1637        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1638        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1639        truncation.
1640    
1641    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1642    
1643    7.  Added two (int) casts to pcregrep when printing the difference of two
1644        pointers, in case they are 64-bit values.
1645    
1646    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1647        test 2 if it fails.
1648    
1649    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1650        and a #define of that name to empty if it is not externally set. This is to
1651        allow users of MSVC to set it if necessary.
1652    
1653    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1654        the convenience functions in the pcre_get.c source file.
1655    
1656    11. An option change at the start of a pattern that had top-level alternatives
1657        could cause overwriting and/or a crash. This command provoked a crash in
1658        some environments:
1659    
1660          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1661    
1662        This potential security problem was recorded as CVE-2008-2371.
1663    
1664    12. For a pattern where the match had to start at the beginning or immediately
1665        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1666        pcre_dfa_exec() could read past the end of the passed subject if there was
1667        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1668        pcretest so that it places the subject at the end of its malloc-ed buffer.
1669    
1670    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1671        exec() might read past the end of the data buffer in UTF-8 mode.
1672    
1673    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1674        the data contained the byte 0x85 as part of a UTF-8 character within its
1675        first line. This applied both to normal and DFA matching.
1676    
1677    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1678        /^[^d]*?$/8 failed to match "abc".
1679    
1680    16. Added a missing copyright notice to pcrecpp_internal.h.
1681    
1682    17. Make it more clear in the documentation that values returned from
1683        pcre_exec() in ovector are byte offsets, not character counts.
1684    
1685    18. Tidied a few places to stop certain compilers from issuing warnings.
1686    
1687    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1688        supplied by Stefan Weber. I made a further small update for 7.8 because
1689        there is a change of source arrangements: the pcre_searchfuncs.c module is
1690        replaced by pcre_ucd.c.
1691    
1692    
1693    Version 7.7 07-May-08
1694    ---------------------
1695    
1696    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1697        a string to a long long, pretend we don't even have a long long." This is
1698        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1699    
1700    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1701        pre-7.6 versions, which defined a global no_arg variable instead of putting
1702        it in the RE class. (See also #8 below.)
1703    
1704    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1705        Lopes.
1706    
1707    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1708    
1709        (1) The include/exclude patterns were being applied to the whole pathnames
1710            of files, instead of just to the final components.
1711    
1712        (2) If there was more than one level of directory, the subdirectories were
1713            skipped unless they satisfied the include/exclude conditions. This is
1714            inconsistent with GNU grep (and could even be seen as contrary to the
1715            pcregrep specification - which I improved to make it absolutely clear).
1716            The action now is always to scan all levels of directory, and just
1717            apply the include/exclude patterns to regular files.
1718    
1719    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1720        --exclude_dir in the tests to avoid scanning .svn directories.
1721    
1722    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1723        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1724        doesn't support NULs in patterns.
1725    
1726    7.  Added some missing "const"s to declarations of static tables in
1727        pcre_compile.c and pcre_dfa_exec.c.
1728    
1729    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1730        caused by fix #2  above. (Subsequently also a second patch to fix the
1731        first patch. And a third patch - this was a messy problem.)
1732    
1733    9.  Applied Craig's patch to remove the use of push_back().
1734    
1735    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1736        matching function regexec().
1737    
1738    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1739        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1740        references. PCRE supports relative numbers with this syntax (I don't think
1741        Oniguruma does).
1742    
1743    12. Previously, a group with a zero repeat such as (...){0} was completely
1744        omitted from the compiled regex. However, this means that if the group
1745        was called as a subroutine from elsewhere in the pattern, things went wrong
1746        (an internal error was given). Such groups are now left in the compiled
1747        pattern, with a new opcode that causes them to be skipped at execution
1748        time.
1749    
1750    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1751        to the way PCRE behaves:
1752    
1753        (a) A lone ] character is dis-allowed (Perl treats it as data).
1754    
1755        (b) A back reference to an unmatched subpattern matches an empty string
1756            (Perl fails the current match path).
1757    
1758        (c) A data ] in a character class must be notated as \] because if the
1759            first data character in a class is ], it defines an empty class. (In
1760            Perl it is not possible to have an empty class.) The empty class []
1761            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1762            The negative empty class [^] matches any one character, independently
1763            of the DOTALL setting.
1764    
1765    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1766        non-existent subpattern following a character class starting with ']' and
1767        containing () gave an internal compiling error instead of "reference to
1768        non-existent subpattern". Fortunately, when the pattern did exist, the
1769        compiled code was correct. (When scanning forwards to check for the
1770        existencd of the subpattern, it was treating the data ']' as terminating
1771        the class, so got the count wrong. When actually compiling, the reference
1772        was subsequently set up correctly.)
1773    
1774    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1775        it was being rejected as not supported by pcre_dfa_exec(), even though
1776        other assertions are supported. I have made pcre_dfa_exec() support
1777        (*FAIL).
1778    
1779    16. The implementation of 13c above involved the invention of a new opcode,
1780        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1781        cannot be changed at match time, I realized I could make a small
1782        improvement to matching performance by compiling OP_ALLANY instead of
1783        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1784        on the OP_ANY path.
1785    
1786    17. Compiling pcretest on Windows with readline support failed without the
1787        following two fixes: (1) Make the unistd.h include conditional on
1788        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1789    
1790    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1791        ncurses library to be included for pcretest when ReadLine support is
1792        requested, but also to allow for it to be overridden. This patch came from
1793        Daniel Bergström.
1794    
1795    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1796        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1797        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1798        spotting this.
1799    
1800    
1801    Version 7.6 28-Jan-08
1802    ---------------------
1803    
1804    1.  A character class containing a very large number of characters with
1805        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1806        overflow.
1807    
1808    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1809        HAVE_LONG_LONG is not defined.
1810    
1811    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1812        bring them up to date and include new features. This patch includes:
1813    
1814        - Fixed PH's badly added libz and libbz2 support.
1815        - Fixed a problem with static linking.
1816        - Added pcredemo. [But later removed - see 7 below.]
1817        - Fixed dftables problem and added an option.
1818        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1819            HAVE_LONG_LONG.
1820        - Added readline support for pcretest.
1821        - Added an listing of the option settings after cmake has run.
1822    
1823    4.  A user submitted a patch to Makefile that makes it easy to create
1824        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1825        Makefile.am that cause it to include this special target, without
1826        affecting anything else. Note that the same mingw target plus all
1827        the other distribution libraries and programs are now supported
1828        when configuring with CMake (see 6 below) instead of with
1829        Configure/Make.
1830    
1831    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1832        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1833        exported in the Windows port". It has not yet been confirmed that the patch
1834        solves the problem, but it does no harm.
1835    
1836    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1837        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1838        with CMake, and also correct the comment about stack recursion.
1839    
1840    7.  Remove the automatic building of pcredemo from the ./configure system and
1841        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1842        of a program that users should build themselves after PCRE is installed, so
1843        building it automatically is not really right. What is more, it gave
1844        trouble in some build environments.
1845    
1846    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1847    
1848    
1849    Version 7.5 10-Jan-08
1850    ---------------------
1851    
1852    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1853        values in parens when parsing an RE using the C++ wrapper."
1854    
1855    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1856        Characters greater than 255 were excluded from the class instead of being
1857        included.
1858    
1859    3.  The same bug as (2) above applied to negated POSIX classes such as
1860        [:^space:].
1861    
1862    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1863        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1864        I have changed it.
1865    
1866    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1867        first named subpattern) and a construct such as (?&a) would reference the
1868        first named subpattern whose name started with "a" (in other words, the
1869        length check was missing). Both these problems are fixed. "Subpattern name
1870        expected" is now given for (?&) (a zero-length name), and this patch also
1871        makes it give the same error for \k'' (previously it complained that that
1872        was a reference to a non-existent subpattern).
1873    
1874    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1875        this is right because (?- can be followed by option settings as well as by
1876        digits. I have, however, made the messages clearer.
1877    
1878    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1879        than the number used in the conditional) now cause a compile-time error.
1880        This is actually not compatible with Perl, which accepts such patterns, but
1881        treats the conditional as always being FALSE (as PCRE used to), but it
1882        seems to me that giving a diagnostic is better.
1883    
1884    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1885        and messages.
1886    
1887    9.  Fix two occurrences of "backslash" in comments that should have been
1888        "backspace".
1889    
1890    10. Remove two redundant lines of code that can never be obeyed (their function
1891        was moved elsewhere).
1892    
1893    11. The program that makes PCRE's Unicode character property table had a bug
1894        which caused it to generate incorrect table entries for sequences of
1895        characters that have the same character type, but are in different scripts.
1896        It amalgamated them into a single range, with the script of the first of
1897        them. In other words, some characters were in the wrong script. There were
1898        thirteen such cases, affecting characters in the following ranges:
1899    
1900          U+002b0 - U+002c1
1901          U+0060c - U+0060d
1902          U+0061e - U+00612
1903          U+0064b - U+0065e
1904          U+0074d - U+0076d
1905          U+01800 - U+01805
1906          U+01d00 - U+01d77
1907          U+01d9b - U+01dbf
1908          U+0200b - U+0200f
1909          U+030fc - U+030fe
1910          U+03260 - U+0327f
1911          U+0fb46 - U+0fbb1
1912          U+10450 - U+1049d
1913    
1914    12. The -o option (show only the matching part of a line) for pcregrep was not
1915        compatible with GNU grep in that, if there was more than one match in a
1916        line, it showed only the first of them. It now behaves in the same way as
1917        GNU grep.
1918    
1919    13. If the -o and -v options were combined for pcregrep, it printed a blank
1920        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1921        does the same. The return code can be used to tell if there were any
1922        non-matching lines.
1923    
1924    14. Added --file-offsets and --line-offsets to pcregrep.
1925    
1926    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1927        infinitely looping recursion. The bug was that positive lookaheads were not
1928        being skipped when checking for a possible empty match (negative lookaheads
1929        and both kinds of lookbehind were skipped).
1930    
1931    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1932        inclusion of <windows.h> to before rather than after the definition of
1933        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1934    
1935    17. Specifying a possessive quantifier with a specific limit for a Unicode
1936        character property caused pcre_compile() to compile bad code, which led at
1937        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1938        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1939        caused the error; without that there was no problem.
1940    
1941    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1942    
1943    19. Added --enable-pcretest-libreadline.
1944    
1945    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1946        RE::GlobalReplace(). As a result, the number of replacements returned was
1947        double what it should be. I removed one of the increments, but Craig sent a
1948        later patch that removed the other one (the right fix) and added unit tests
1949        that check the return values (which was not done before).
1950    
1951    21. Several CMake things:
1952    
1953        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1954            the names libpcre and libpcreposix, not just pcre and pcreposix.
1955    
1956        (2) The above change means that pcretest and pcregrep are now correctly
1957            linked with the newly-built libraries, not previously installed ones.
1958    
1959        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1960    
1961    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1962        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1963        UTF-8 newline character). The key issue is that the pattern starts .*;
1964        this means that the match must be either at the beginning, or after a
1965        newline. The bug was in the code for advancing after a failed match and
1966        checking that the new position followed a newline. It was not taking
1967        account of UTF-8 characters correctly.
1968    
1969    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1970        character classes. PCRE was not treating the sequence [:...:] as a
1971        character class unless the ... were all letters. Perl, however, seems to
1972        allow any characters between [: and :], though of course it rejects as
1973        unknown any "names" that contain non-letters, because all the known class
1974        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1975        for example, whereas PCRE did not - it did not recognize a POSIX character
1976        class. This seemed a bit dangerous, so the code has been changed to be
1977        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1978        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1979        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1980        Perl does, and where it didn't before.
1981    
1982    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1983        Windows environments %n is disabled by default.
1984    
1985    
1986    Version 7.4 21-Sep-07
1987    ---------------------
1988    
1989    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1990        means that a class such as [\s] counted as "explicit reference to CR or
1991        LF". That isn't really right - the whole point of the change was to try to
1992        help when there was an actual mention of one of the two characters. So now
1993        the change happens only if \r or \n (or a literal CR or LF) character is
1994        encountered.
1995    
1996    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1997        of both had grown to the point where there were only 3 bits left.
1998        Fortunately, there was spare space in the data structure, and so I have
1999        moved the internal flags into a new 16-bit field to free up more option
2000        bits.
2001    
2002    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2003        but did not set the internal JCHANGED flag - either of these is enough to
2004        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2005        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2006        start sets both bits.
2007    
2008    4.  Added options (at build time, compile time, exec time) to change \R from
2009        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2010    
2011    5.  doc/pcresyntax.html was missing from the distribution.
2012    
2013    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2014        compatibility, even though it is no longer used.
2015    
2016    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2017        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2018        windows.h file is present (where different names are used). [This was
2019        reversed later after testing - see 16 below.]
2020    
2021    8.  Changed all #include <config.h> to #include "config.h". There were also
2022        some further <pcre.h> cases that I changed to "pcre.h".
2023    
2024    9.  When pcregrep was used with the --colour option, it missed the line ending
2025        sequence off the lines that it output.
2026    
2027    10. It was pointed out to me that arrays of string pointers cause lots of
2028        relocations when a shared library is dynamically loaded. A technique of
2029        using a single long string with a table of offsets can drastically reduce
2030        these. I have refactored PCRE in four places to do this. The result is
2031        dramatic:
2032    
2033          Originally:                          290
2034          After changing UCP table:            187
2035          After changing error message table:   43
2036          After changing table of "verbs"       36
2037          After changing table of Posix names   22
2038    
2039        Thanks to the folks working on Gregex for glib for this insight.
2040    
2041    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2042        unicode-properties was also set.
2043    
2044    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2045    
2046    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2047        checked only for CRLF.
2048    
2049    14. Added casts to pcretest.c to avoid compiler warnings.
2050    
2051    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2052    
2053    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2054        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2055        entirely. This removes changes made in 7 above.
2056    
2057    17. The CMake files have been updated, and there is now more information about
2058        building with CMake in the NON-UNIX-USE document.
2059    
2060    
2061    Version 7.3 28-Aug-07
2062    ---------------------
2063    
2064     1. In the rejigging of the build system that eventually resulted in 7.1, the
2065        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2066        brackets there is not right, since it causes compilers to look for an
2067        installed pcre.h, not the version that is in the source that is being
2068        compiled (which of course may be different). I have changed it back to:
2069    
2070          #include "pcre.h"
2071    
2072        I have a vague recollection that the change was concerned with compiling in
2073        different directories, but in the new build system, that is taken care of
2074        by the VPATH setting the Makefile.
2075    
2076     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2077        when the subject happened to end in the byte 0x85 (e.g. if the last
2078        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2079        characters but of course it shouldn't be taken as a newline when it is part
2080        of another character. The bug was that, for an unlimited repeat of . in
2081        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2082        characters when looking for a newline.
2083    
2084     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2085    
2086     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2087        in debug output.
2088    
2089     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2090        long printing in the pcrecpp unittest when running under MinGW.
2091    
2092     6. ESC_K was left out of the EBCDIC table.
2093    
2094     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2095        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2096        limit also applies to "virtual nesting" when a pattern is recursive, and in
2097        this case 1000 isn't so big. I have been able to remove this limit at the
2098        expense of backing off one optimization in certain circumstances. Normally,
2099        when pcre_exec() would call its internal match() function recursively and
2100        immediately return the result unconditionally, it uses a "tail recursion"
2101        feature to save stack. However, when a subpattern that can match an empty
2102        string has an unlimited repetition quantifier, it no longer makes this
2103        optimization. That gives it a stack frame in which to save the data for
2104        checking that an empty string has been matched. Previously this was taken
2105        from the 1000-entry workspace that had been reserved. So now there is no
2106        explicit limit, but more stack is used.
2107    
2108     8. Applied Daniel's patches to solve problems with the import/export magic
2109        syntax that is required for Windows, and which was going wrong for the
2110        pcreposix and pcrecpp parts of the library. These were overlooked when this
2111        problem was solved for the main library.
2112    
2113     9. There were some crude static tests to avoid integer overflow when computing
2114        the size of patterns that contain repeated groups with explicit upper
2115        limits. As the maximum quantifier is 65535, the maximum group length was
2116        set at 30,000 so that the product of these two numbers did not overflow a
2117        32-bit integer. However, it turns out that people want to use groups that
2118        are longer than 30,000 bytes (though not repeat them that many times).
2119        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2120        made it possible to implement the integer overflow checks in a much more
2121        dynamic way, which I have now done. The artificial limitation on group
2122        length has been removed - we now have only the limit on the total length of
2123        the compiled pattern, which depends on the LINK_SIZE setting.
2124    
2125    10. Fixed a bug in the documentation for get/copy named substring when
2126        duplicate names are permitted. If none of the named substrings are set, the
2127        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2128        empty string.
2129    
2130    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2131        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2132        because the ] is interpreted as the first data character and the
2133        terminating ] is not found. PCRE has been made compatible with Perl in this
2134        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2135        cause memory overwriting.
2136    
2137    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2138        string has been matched (to stop an infinite loop). It was not recognizing
2139        a conditional subpattern that could match an empty string if that
2140        subpattern was within another subpattern. For example, it looped when
2141        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2142        condition was not nested. This bug has been fixed.
2143    
2144    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2145        past the start of the subject in the presence of bytes with the top bit
2146        set, for example "\x8aBCD".
2147    
2148    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2149        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2150    
2151    14. Optimized (?!) to (*FAIL).
2152    
2153    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2154        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2155        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2156        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2157        does: it's just the validity check that is more restrictive.
2158    
2159    16. Inserted checks for integer overflows during escape sequence (backslash)
2160        processing, and also fixed erroneous offset values for syntax errors during
2161        backslash processing.
2162    
2163    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2164        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2165    
2166    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2167        caused an overrun.
2168    
2169    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2170        something other than just ASCII characters) inside a group that had an
2171        unlimited repeat caused a loop at compile time (while checking to see
2172        whether the group could match an empty string).
2173    
2174    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2175        [\P{Any}] did so. (Error in the code for printing property names.)
2176    
2177    21. An orphan \E inside a character class could cause a crash.
2178    
2179    22. A repeated capturing bracket such as (A)? could cause a wild memory
2180        reference during compilation.
2181    
2182    23. There are several functions in pcre_compile() that scan along a compiled
2183        expression for various reasons (e.g. to see if it's fixed length for look
2184        behind). There were bugs in these functions when a repeated \p or \P was
2185        present in the pattern. These operators have additional parameters compared
2186        with \d, etc, and these were not being taken into account when moving along
2187        the compiled data. Specifically:
2188    
2189        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2190            length.
2191    
2192        (b) An item such as \pL+ within a repeated group could cause crashes or
2193            loops.
2194    
2195        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2196            "reference to non-existent subpattern" error.
2197    
2198        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2199    
2200    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2201        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2202    
2203    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2204    
2205    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2206        character were causing crashes (broken optimization).
2207    
2208    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2209        \p or \P) caused a compile-time loop.
2210    
2211    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2212        break. For example, the unstudied pattern [\r\n]A does not match the string
2213        "\r\nA" because change 7.0/46 below moves the current point on by two
2214        characters after failing to match at the start. However, the pattern \nA
2215        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2216        the same is true. There doesn't seem any very clean way out of this, but
2217        what I have chosen to do makes the common cases work: PCRE now takes note
2218        of whether there can be an explicit match for \r or \n anywhere in the
2219        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2220        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2221        pattern has explicit CR or LF references.
2222    
2223    29. Added (*CR) etc for changing newline setting at start of pattern.
2224    
2225    
2226    Version 7.2 19-Jun-07
2227    ---------------------
2228    
2229     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2230        which is apparently normally available under Windows.
2231    
2232     2. Re-jig the pcregrep tests with different newline settings in an attempt
2233        to make them independent of the local environment's newline setting.
2234    
2235     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2236    
2237     4. Some of the "internals" tests were previously cut out when the link size
2238        was not 2, because the output contained actual offsets. The recent new
2239        "Z" feature of pcretest means that these can be cut out, making the tests
2240        usable with all link sizes.
2241    
2242     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2243        stack recursion. This gives a massive performance boost under BSD, but just
2244        a small improvement under Linux. However, it saves one field in the frame
2245        in all cases.
2246    
2247     6. Added more features from the forthcoming Perl 5.10:
2248    
2249        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2250            recursion call. It refers to the nth most recently opened parentheses.
2251    
2252        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2253            to be opened parentheses.
2254    
2255        (c) Conditions that refer to capturing parentheses can be specified
2256            relatively, for example, (?(-2)... or (?(+3)...
2257    
2258        (d) \K resets the start of the current match so that everything before
2259            is not part of it.
2260    
2261        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2262    
2263        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2264            reference syntax.
2265    
2266        (g) (?| introduces a group in which the numbering of parentheses in each
2267            alternative starts with the same number.
2268    
2269        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2270    
2271     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2272        PCRE_INFO_JCHANGED.
2273    
2274     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2275        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2276        for detecting groups that can match an empty string.
2277    
2278     9. A pattern with a very large number of alternatives (more than several
2279        hundred) was running out of internal workspace during the pre-compile
2280        phase, where pcre_compile() figures out how much memory will be needed. A
2281        bit of new cunning has reduced the workspace needed for groups with
2282        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2283        workspace instead of running out of the 4096 that are available.
2284    
2285    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2286    
2287    11. Applied patch from Google to remove an optimization that didn't quite work.
2288        The report of the bug said:
2289    
2290          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2291          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2292          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2293    
2294    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2295        it matched the wrong number of bytes.
2296    
2297    
2298    Version 7.1 24-Apr-07
2299  ---------------------  ---------------------
2300    
2301   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2302      that is more "standard", making use of automake and other autotools. There      that is more "standard", making use of automake and other Autotools. There
2303      is some re-arrangement of the files and adjustment of comments consequent      is some re-arrangement of the files and adjustment of comments consequent
2304      on this.      on this.
2305    
2306   2. Part of the patch fixed a problem with the pcregrep tests. The test of -r   2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2307      for recursive directory scanning broke on some systems because the files      for recursive directory scanning broke on some systems because the files
2308      are not scanned in any specific order and on different systems the order      are not scanned in any specific order and on different systems the order
2309      was different. A call to "sort" has been inserted into RunGrepTest for the      was different. A call to "sort" has been inserted into RunGrepTest for the
2310      approprate test as a short-term fix. In the longer term there may be an      approprate test as a short-term fix. In the longer term there may be an
2311      alternative.      alternative.
2312    
2313   3. I had an email from Eric Raymond about problems translating some of PCRE's   3. I had an email from Eric Raymond about problems translating some of PCRE's
2314      man pages to HTML (despite the fact that I distribute HTML pages, some      man pages to HTML (despite the fact that I distribute HTML pages, some
2315      people do their own conversions for various reasons). The problems      people do their own conversions for various reasons). The problems
2316      concerned the use of low-level troff macros .br and .in. I have therefore      concerned the use of low-level troff macros .br and .in. I have therefore
2317      removed all such uses from the man pages (some were redundant, some could      removed all such uses from the man pages (some were redundant, some could
2318      be replaced by .nf/.fi pairs). The maintain/132html script that I use to      be replaced by .nf/.fi pairs). The 132html script that I use to generate
2319      generate HTML has been updated to handle .nf/.fi and to complain if it      HTML has been updated to handle .nf/.fi and to complain if it encounters
2320      encounters .br or .in.      .br or .in.
2321    
2322   4. Updated comments in configure.ac that get placed in config.h.in and also   4. Updated comments in configure.ac that get placed in config.h.in and also
2323      arranged for config.h to be included in the distribution, for the benefit      arranged for config.h to be included in the distribution, with the name
2324      of those who have to compile without Autotools (compare pcre.h).      config.h.generic, for the benefit of those who have to compile without
2325        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2326   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan  
2327      Weber: (1) pcre_internal.h was missing some function renames; (2) updated   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2328      makevp.bat for the current PCRE, using the additional files !compile.txt,      Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2329      !linklib.txt, and pcregexp.pas.      makevp.bat for the current PCRE, using the additional files
2330        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2331   6. A Windows user reported a minor discrepancy with test 2, which turned out  
2332      to be caused by a trailing space on an input line that had got lost in his   6. A Windows user reported a minor discrepancy with test 2, which turned out
2333        to be caused by a trailing space on an input line that had got lost in his
2334      copy. The trailing space was an accident, so I've just removed it.      copy. The trailing space was an accident, so I've just removed it.
2335    
2336   7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told   7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2337      that is needed.      that is needed.
2338    
2339   8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)   8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2340      as "const" (a) because they are and (b) because it helps the PHP      as "const" (a) because they are and (b) because it helps the PHP
2341      maintainers who have recently made a script to detect big data structures      maintainers who have recently made a script to detect big data structures
2342      in the php code that should be moved to the .rodata section. I remembered      in the php code that should be moved to the .rodata section. I remembered
2343      to update Builducptable as well, so it won't revert if ucptable.h is ever      to update Builducptable as well, so it won't revert if ucptable.h is ever
2344      re-created.      re-created.
2345    
2346   9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,   9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2347      pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in      pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2348      order to be able to cut out the UTF-8 tables in the latter when UTF-8      order to be able to cut out the UTF-8 tables in the latter when UTF-8
2349      support is not required. This saves 1.5-2K of code, which is important in      support is not required. This saves 1.5-2K of code, which is important in
2350      some applications.      some applications.
2351    
2352  10. The emulated memmove() function in pcre_internal.h (provided for those      Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2353      environments that have neither memmove() nor bcopy()) was defining its      so as not to refer to the tables, even though these functions will never be
2354      arguments as char * instead of void *.      called when UTF-8 support is disabled. Otherwise there are problems with a
2355        shared library.
2356    
2357    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2358    
2359        (a) It was defining its arguments as char * instead of void *.
2360    
2361        (b) It was assuming that all moves were upwards in memory; this was true
2362            a long time ago when I wrote it, but is no longer the case.
2363    
2364        The emulated memove() is provided for those environments that have neither
2365        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2366        is clearly not the case, as these two bugs were recently reported.
2367    
2368    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2369        and Detrail to create the HTML documentation, the .txt form of the man
2370        pages, and it removes trailing spaces from listed files. It also creates
2371        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2372        case, it wraps all the #defines with #ifndefs. This script should be run
2373        before "make dist".
2374    
2375    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2376        with Unicode property support.
2377    
2378        (a) For a maximizing quantifier, if the two different cases of the
2379            character were of different lengths in their UTF-8 codings (there are
2380            some cases like this - I found 11), and the matching function had to
2381            back up over a mixture of the two cases, it incorrectly assumed they
2382            were both the same length.
2383    
2384        (b) When PCRE was configured to use the heap rather than the stack for
2385            recursion during matching, it was not correctly preserving the data for
2386            the other case of a UTF-8 character when checking ahead for a match
2387            while processing a minimizing repeat. If the check also involved
2388            matching a wide character, but failed, corruption could cause an
2389            erroneous result when trying to check for a repeat of the original
2390            character.
2391    
2392    13. Some tidying changes to the testing mechanism:
2393    
2394        (a) The RunTest script now detects the internal link size and whether there
2395            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2396            values substituted by "configure". (The RunGrepTest script already did
2397            this for UTF-8.) The configure.ac script no longer substitutes the
2398            relevant variables.
2399    
2400        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2401            with length and offset values. This means that the output is different
2402            for different internal link sizes. Test 2 is skipped for link sizes
2403            other than 2 because of this, bypassing the problem. Unfortunately,
2404            there was also a test in test 3 (the locale tests) that used /B and
2405            failed for link sizes other than 2. Rather than cut the whole test out,
2406            I have added a new /Z option to pcretest that replaces the length and
2407            offset values with spaces. This is now used to make test 3 independent
2408            of link size. (Test 2 will be tidied up later.)
2409    
2410    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2411        segmentation fault instead of returning the appropriate error message.
2412    
2413    15. In multiline mode when the newline sequence was set to "any", the pattern
2414        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2415        This doesn't seem right; it now treats the CRLF combination as the line
2416        ending, and so does not match in that case. It's only a pattern such as ^$
2417        that would hit this one: something like ^ABC$ would have failed after \r
2418        and then tried again after \r\n.
2419    
2420    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2421        in an attempt to make files that differ only in their line terminators
2422        compare equal. This works on Linux.
2423    
2424    17. Under certain error circumstances pcregrep might try to free random memory
2425        as it exited. This is now fixed, thanks to valgrind.
2426    
2427    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2428        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2429        was because its rules for how to advance for /g after matching an empty
2430        string at the end of a line did not allow for this case. They now check for
2431        it specially.
2432    
2433    20. pcretest is supposed to handle patterns and data of any length, by
2434        extending its buffers when necessary. It was getting this wrong when the
2435        buffer for a data line had to be extended.
2436    
2437    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2438        CRLF as a newline sequence.
2439    
2440    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2441        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2442        I have nevertheless tidied it up.
2443    
2444    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2445    
2446    24. Added a man page for pcre-config.
2447    
2448    
2449  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06
2450  ---------------------  ---------------------

Legend:
Removed from v.108  
changed lines
  Added in v.1235

  ViewVC Help
Powered by ViewVC 1.1.5