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

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 1219 by ph10, Sun Nov 11 18:04:37 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  Version 8.32 11-November-2012
5    -----------------------------
6    
7    1.  Improved JIT compiler optimizations for first character search and single
8        character iterators.
9    
10    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11        Patch by Daniel Richard G.
12    
13    3.  Single character iterator optimizations in the JIT compiler.
14    
15    4.  Improved JIT compiler optimizations for character ranges.
16    
17    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18        Reported by Giuseppe D'Angelo.
19    
20    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21        of a line, was being set incorrectly in cases where .* appeared inside
22        atomic brackets at the start of a pattern, or where there was a subsequent
23        *PRUNE or *SKIP.
24    
25    7.  Improved instruction cache flush for POWER/PowerPC.
26        Patch by Daniel Richard G.
27    
28    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
29        grep:
30    
31        (a) There is now no limit to the number of patterns to be matched.
32    
33        (b) An error is given if a pattern is too long.
34    
35        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36            are now supported.
37    
38        (d) --exclude-from and --include-from (multiple use) have been added.
39    
40        (e) Exclusions and inclusions now apply to all files and directories, not
41            just to those obtained from scanning a directory recursively.
42    
43        (f) Multiple uses of -f and --file-list are now supported.
44    
45        (g) In a Windows environment, the default for -d has been changed from
46            "read" (the GNU grep default) to "skip", because otherwise the presence
47            of a directory in the file list provokes an error.
48    
49        (h) The documentation has been revised and clarified in places.
50    
51    9.  Improve the matching speed of capturing brackets.
52    
53    10. Changed the meaning of \X so that it now matches a Unicode extended
54        grapheme cluster.
55    
56    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57        out POSIX threads when JIT support is configured.
58    
59    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
60    
61    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62        structure could go wrong in environments where size_t is not the same size
63        as int.
64    
65    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
66        set.
67    
68    15. The EBCDIC support had decayed; later updates to the code had included
69        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
70        general tidy up of EBCDIC-related issues, and the documentation was also
71        not quite right. There is now a test that can be run on ASCII systems to
72        check some of the EBCDIC-related things (but is it not a full test).
73    
74    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
75        in a small tidy to the code.
76    
77    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
78    
79    18. If the --only-matching (-o) option in pcregrep is specified multiple
80        times, each one causes appropriate output. For example, -o1 -o2 outputs the
81        substrings matched by the 1st and 2nd capturing parentheses. A separating
82        string can be specified by --om-separator (default empty).
83    
84    19. Improving the first n character searches.
85    
86    20. Turn case lists for horizontal and vertical white space into macros so that
87        they are defined only once.
88    
89    21. This set of changes together give more compatible Unicode case-folding
90        behaviour for characters that have more than one other case when UCP
91        support is available.
92    
93        (a) The Unicode property table now has offsets into a new table of sets of
94            three or more characters that are case-equivalent. The MultiStage2.py
95            script that generates these tables (the pcre_ucd.c file) now scans
96            CaseFolding.txt instead of UnicodeData.txt for character case
97            information.
98    
99        (b) The code for adding characters or ranges of characters to a character
100            class has been abstracted into a generalized function that also handles
101            case-independence. In UTF-mode with UCP support, this uses the new data
102            to handle characters with more than one other case.
103    
104        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
105            whose other case is greater than 256 are now correctly matched
106            caselessly. Previously, the high codepoint matched the low one, but not
107            vice versa.
108    
109        (d) The processing of \h, \H, \v, and \ in character classes now makes use
110            of the new class addition function, using character lists defined as
111            macros alongside the case definitions of 20 above.
112    
113        (e) Caseless back references now work with characters that have more than
114            one other case.
115    
116        (f) General caseless matching of characters with more than one other case
117            is supported.
118    
119    22. Unicode character properties were updated from Unicode 6.2.0
120    
121    23. Improved CMake support under Windows. Patch by Daniel Richard G.
122    
123    24. Add support for 32-bit character strings, and UTF-32
124    
125    25. Major JIT compiler update (code refactoring and bugfixing).
126        Experimental Sparc 32 support is added.
127    
128    26. Applied a modified version of Daniel Richard G's patch to create
129        pcre.h.generic and config.h.generic by "make" instead of in the
130        PrepareRelease script.
131    
132    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
133        pcre_compile.c when checking for a zero character.
134    
135    28. Introducing a native interface for JIT. Through this interface, the compiled
136        machine code can be directly executed. The purpose of this interface is to
137        provide fast pattern matching, so several sanity checks are not performed.
138        However, feature tests are still performed. The new interface provides
139        1.4x speedup compared to the old one.
140    
141    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
142        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
143        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
144    
145    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
146        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
147        the "old" RFC 2279). Instead, it ended up passing a negative length to
148        pcre_exec().
149    
150    31. Add support for GCC's visibility feature to hide internal functions.
151    
152    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
153        "unknown -C option" after outputting 0 or 1.
154    
155    33. There is now support for generating a code coverage report for the test
156        suite in environments where gcc is the compiler and lcov is installed. This
157        is mainly for the benefit of the developers.
158    
159    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
160        unaddressable using valgrind annotations, allowing valgrind to detect
161        invalid memory accesses. This is mainly for the benefit of the developers.
162    
163    25. (*UTF) can now be used to start a pattern in any of the three libraries.
164    
165    
166    Version 8.31 06-July-2012
167    -------------------------
168    
169    1.  Fixing a wrong JIT test case and some compiler warnings.
170    
171    2.  Removed a bashism from the RunTest script.
172    
173    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
174        to unsigned type, result still unsigned" that was given by an MS compiler
175        on encountering the code "-sizeof(xxx)".
176    
177    4.  Partial matching support is added to the JIT compiler.
178    
179    5.  Fixed several bugs concerned with partial matching of items that consist
180        of more than one character:
181    
182        (a) /^(..)\1/ did not partially match "aba" because checking references was
183            done on an "all or nothing" basis. This also applied to repeated
184            references.
185    
186        (b) \R did not give a hard partial match if \r was found at the end of the
187            subject.
188    
189        (c) \X did not give a hard partial match after matching one or more
190            characters at the end of the subject.
191    
192        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
193            a partial match for the string "\r".
194    
195        (e) When newline was set to CRLF, the metacharacter "." did not recognize
196            a partial match for a CR character at the end of the subject string.
197    
198    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
199        running pcretest, the text "(JIT)" added to the output whenever JIT is
200        actually used to run the match.
201    
202    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
203        or /S+[+] with a digit between 1 and 7.
204    
205    8.  OP_NOT now supports any UTF character not just single-byte ones.
206    
207    9.  (*MARK) control verb is now supported by the JIT compiler.
208    
209    10. The command "./RunTest list" lists the available tests without actually
210        running any of them. (Because I keep forgetting what they all are.)
211    
212    11. Add PCRE_INFO_MAXLOOKBEHIND.
213    
214    12. Applied a (slightly modified) user-supplied patch that improves performance
215        when the heap is used for recursion (compiled with --disable-stack-for-
216        recursion). Instead of malloc and free for each heap frame each time a
217        logical recursion happens, frames are retained on a chain and re-used where
218        possible. This sometimes gives as much as 30% improvement.
219    
220    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
221        call.
222    
223    14. As documented, (*COMMIT) is now confined to within a positive assertion.
224    
225    15. It is now possible to link pcretest with libedit as an alternative to
226        libreadline.
227    
228    16. (*COMMIT) control verb is now supported by the JIT compiler.
229    
230    17. The Unicode data tables have been updated to Unicode 6.1.0.
231    
232    18. Added --file-list option to pcregrep.
233    
234    19. Added binary file support to pcregrep, including the -a, --binary-files,
235        -I, and --text options.
236    
237    20. The madvise function is renamed for posix_madvise for QNX compatibility
238        reasons. Fixed by Giuseppe D'Angelo.
239    
240    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
241    
242    22. Changed the option for creating symbolic links for 16-bit man pages from
243        -s to -sf so that re-installing does not cause issues.
244    
245    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
246    
247    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
248        matches in certain environments (the workspace was not being correctly
249        retained). Also added to pcre_dfa_exec() a simple plausibility check on
250        some of the workspace data at the beginning of a restart.
251    
252    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
253        was not doing so when it should - probably a typo introduced by SVN 528
254        (change 8.10/14).
255    
256    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
257        \w+ when the character tables indicated that \x{c4} was a word character.
258        There were several related cases, all because the tests for doing a table
259        lookup were testing for characters less than 127 instead of 255.
260    
261    27. If a pattern contains capturing parentheses that are not used in a match,
262        their slots in the ovector are set to -1. For those that are higher than
263        any matched groups, this happens at the end of processing. In the case when
264        there were back references that the ovector was too small to contain
265        (causing temporary malloc'd memory to be used during matching), and the
266        highest capturing number was not used, memory off the end of the ovector
267        was incorrectly being set to -1. (It was using the size of the temporary
268        memory instead of the true size.)
269    
270    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
271        ovector size, it uses memory at the end of the block that it has got.
272    
273    29. Check for an overlong MARK name and give an error at compile time. The
274        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
275    
276    30. JIT compiler update.
277    
278    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
279        Rill for the patch.
280    
281    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
282    
283    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
284    
285    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
286        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
287        was enabled.
288    
289    35. Improve JIT code generation for greedy plus quantifier.
290    
291    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
292        1 to "aa" instead of to an empty string. The bug affected repeated groups
293        that could potentially match an empty string.
294    
295    37. Optimizing single character iterators in JIT.
296    
297    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
298        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
299        codepoints that are too big for the mode are faulted, and in a UTF mode,
300        disallowed codepoints are also faulted.
301    
302    39. If PCRE was compiled with UTF support, in three places in the DFA
303        matcher there was code that should only have been obeyed in UTF mode, but
304        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
305        processing when bytes with values greater than 127 were present. In 16-bit
306        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
307        both cases the values are those that cannot be the first data item in a UTF
308        character. The three items that might have provoked this were recursions,
309        possessively repeated groups, and atomic groups.
310    
311    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
312        and pcregrep, because some OS require shared objects to be explicitly
313        passed to ld, causing the link step to fail if they are not.
314    
315    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
316        mode, patterns that started with \h* or \R* might be incorrectly matched.
317    
318    
319    Version 8.30 04-February-2012
320    -----------------------------
321    
322    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
323        name is defined in ctype.h.
324    
325    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
326        only in quite long subpatterns.
327    
328    3.  Removed the function pcre_info(), which has been obsolete and deprecated
329        since it was replaced by pcre_fullinfo() in February 2000.
330    
331    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
332        match a (*MARK), and the match failed at the start of the subject, a
333        reference to memory before the start of the subject could occur. This bug
334        was introduced by fix 17 of release 8.21.
335    
336    5.  A reference to an unset group with zero minimum repetition was giving
337        totally wrong answers (in non-JavaScript-compatibility mode). For example,
338        /(another)?(\1?)test/ matched against "hello world test". This bug was
339        introduced in release 8.13.
340    
341    6.  Add support for 16-bit character strings (a large amount of work involving
342        many changes and refactorings).
343    
344    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
345        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
346        from a file.
347    
348    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
349        rounding is not applied in this particular case).
350    
351    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
352        if they appear, or are escaped, in patterns.
353    
354    10. Get rid of a number of -Wunused-but-set-variable warnings.
355    
356    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
357        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
358        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
359        also returns the mark "x". This bug applied to capturing parentheses,
360        non-capturing parentheses, and atomic parentheses. It also applied to some
361        assertions.
362    
363    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
364        information out of configure.ac instead of relying on pcre.h.generic, which
365        is not stored in the repository.
366    
367    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
368        -lreadline.
369    
370    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
371    
372    15. Applied Graycode's patch to put the top-level frame on the stack rather
373        than the heap when not using the stack for recursion. This gives a
374        performance improvement in many cases when recursion is not deep.
375    
376    16. Experimental code added to "pcretest -C" to output the stack frame size.
377    
378    
379    Version 8.21 12-Dec-2011
380    ------------------------
381    
382    1.  Updating the JIT compiler.
383    
384    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
385        are added as well.
386    
387    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
388        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
389        calling _pcre_jit_exec. Some extra comments are added.
390    
391    4.  (*MARK) settings inside atomic groups that do not contain any capturing
392        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
393        was introduced by change 18 for 8.20.
394    
395    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
396        ECMA-262 standard.
397    
398    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
399        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
400        This bug was probably introduced by change 9 of 8.13.
401    
402    7.  While fixing 6 above, I noticed that a number of other items were being
403        incorrectly rejected as "not fixed length". This arose partly because newer
404        opcodes had not been added to the fixed-length checking code. I have (a)
405        corrected the bug and added tests for these items, and (b) arranged for an
406        error to occur if an unknown opcode is encountered while checking for fixed
407        length instead of just assuming "not fixed length". The items that were
408        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
409        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
410        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
411    
412    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
413        being incorrectly compiled and would have given unpredicatble results.
414    
415    9.  A possessively repeated subpattern with minimum repeat count greater than
416        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
417        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
418        the first (A) could occur when it should not.
419    
420    10. Add a cast and remove a redundant test from the code.
421    
422    11. JIT should use pcre_malloc/pcre_free for allocation.
423    
424    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
425        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
426        is anything other than /usr, -L is still shown).
427    
428    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
429    
430    14. Perl does not support \N without a following name in a [] class; PCRE now
431        also gives an error.
432    
433    15. If a forward reference was repeated with an upper limit of around 2000,
434        it caused the error "internal error: overran compiling workspace". The
435        maximum number of forward references (including repeats) was limited by the
436        internal workspace, and dependent on the LINK_SIZE. The code has been
437        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
438        the default depends on LINK_SIZE. There is a new upper limit (for safety)
439        of around 200,000 forward references. While doing this, I also speeded up
440        the filling in of repeated forward references.
441    
442    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
443        incorrectly expecting the subject to contain another "a" after the start.
444    
445    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
446        in the match, the SKIP should be ignored. This was not happening; instead
447        the SKIP was being treated as NOMATCH. For patterns such as
448        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
449        tested.
450    
451    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
452        now much more compatible with Perl, in particular in cases where the result
453        is a non-match for a non-anchored pattern. For example, if
454        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
455        "m", where previously it did not return a name. A side effect of this
456        change is that for partial matches, the last encountered mark name is
457        returned, as for non matches. A number of tests that were previously not
458        Perl-compatible have been moved into the Perl-compatible test files. The
459        refactoring has had the pleasing side effect of removing one argument from
460        the match() function, thus reducing its stack requirements.
461    
462    19. If the /S+ option was used in pcretest to study a pattern using JIT,
463        subsequent uses of /S (without +) incorrectly behaved like /S+.
464    
465    21. Retrieve executable code size support for the JIT compiler and fixing
466        some warnings.
467    
468    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
469        not work when the shorter character appeared right at the end of the
470        subject string.
471    
472    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
473        systems.
474    
475    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
476        output it when the /M option is used in pcretest.
477    
478    25. The CheckMan script was not being included in the distribution. Also, added
479        an explicit "perl" to run Perl scripts from the PrepareRelease script
480        because this is reportedly needed in Windows.
481    
482    26. If study data was being save in a file and studying had not found a set of
483        "starts with" bytes for the pattern, the data written to the file (though
484        never used) was taken from uninitialized memory and so caused valgrind to
485        complain.
486    
487    27. Updated RunTest.bat as provided by Sheri Pierce.
488    
489    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
490    
491    29. Computation of memory usage for the table of capturing group names was
492        giving an unnecessarily large value.
493    
494    
495    Version 8.20 21-Oct-2011
496    ------------------------
497    
498    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
499        a POSIX class. After further experiments with Perl, which convinced me that
500        Perl has bugs and confusions, a closing square bracket is no longer allowed
501        in a POSIX name. This bug also affected patterns with classes that started
502        with full stops.
503    
504    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
505        captured substring, but while checking the failing first alternative,
506        substring 1 is temporarily captured. If the output vector supplied to
507        pcre_exec() was not big enough for this capture, the yield of the function
508        was still zero ("insufficient space for captured substrings"). This cannot
509        be totally fixed without adding another stack variable, which seems a lot
510        of expense for a edge case. However, I have improved the situation in cases
511        such as /(a)(b)x|abc/ matched against "abc", where the return code
512        indicates that fewer than the maximum number of slots in the ovector have
513        been set.
514    
515    3.  Related to (2) above: when there are more back references in a pattern than
516        slots in the output vector, pcre_exec() uses temporary memory during
517        matching, and copies in the captures as far as possible afterwards. It was
518        using the entire output vector, but this conflicts with the specification
519        that only 2/3 is used for passing back captured substrings. Now it uses
520        only the first 2/3, for compatibility. This is, of course, another edge
521        case.
522    
523    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
524        main code base, and can be used by building with --enable-jit. When this is
525        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
526        runtime --no-jit option is given.
527    
528    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
529        ovector, the return from the function was zero, implying that there were
530        other matches that did not fit. The correct "exactly full" value is now
531        returned.
532    
533    6.  If a subpattern that was called recursively or as a subroutine contained
534        (*PRUNE) or any other control that caused it to give a non-standard return,
535        invalid errors such as "Error -26 (nested recursion at the same subject
536        position)" or even infinite loops could occur.
537    
538    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
539        computing the minimum length on reaching *ACCEPT, and so ended up with the
540        wrong value of 1 rather than 0. Further investigation indicates that
541        computing a minimum subject length in the presence of *ACCEPT is difficult
542        (think back references, subroutine calls), and so I have changed the code
543        so that no minimum is registered for a pattern that contains *ACCEPT.
544    
545    8.  If (*THEN) was present in the first (true) branch of a conditional group,
546        it was not handled as intended. [But see 16 below.]
547    
548    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
549        Sheri Pierce.
550    
551    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
552        the first byte in a match must be "a".
553    
554    11. Change 17 for 8.13 increased the recursion depth for patterns like
555        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
556        pattern contains any instances of (*THEN). If it does not, the old
557        optimizations are restored. It would be nice to do this on a per-group
558        basis, but at the moment that is not feasible.
559    
560    12. In some environments, the output of pcretest -C is CRLF terminated. This
561        broke RunTest's code that checks for the link size. A single white space
562        character after the value is now allowed for.
563    
564    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
565        For "fr", it uses the Windows-specific input and output files.
566    
567    14. If (*THEN) appeared in a group that was called recursively or as a
568        subroutine, it did not work as intended. [But see next item.]
569    
570    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
571        pattern fragments (but not containing any | characters). If A and B are
572        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
573        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
574        D. In other words, Perl considers parentheses that do not contain any |
575        characters to be part of a surrounding alternative, whereas PCRE was
576        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
577        differently. PCRE now behaves in the same way as Perl, except in the case
578        of subroutine/recursion calls such as (?1) which have in any case always
579        been different (but PCRE had them first :-).
580    
581    16. Related to 15 above: Perl does not treat the | in a conditional group as
582        creating alternatives. Such a group is treated in the same way as an
583        ordinary group without any | characters when processing (*THEN). PCRE has
584        been changed to match Perl's behaviour.
585    
586    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
587        RunGrepTest script failed.
588    
589    18. Change 22 for version 13 caused atomic groups to use more stack. This is
590        inevitable for groups that contain captures, but it can lead to a lot of
591        stack use in large patterns. The old behaviour has been restored for atomic
592        groups that do not contain any capturing parentheses.
593    
594    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
595        suppress the check for a minimum subject length at run time. (If it was
596        given to pcre_exec() or pcre_dfa_exec() it did work.)
597    
598    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
599        fail to work when decoding hex characters in data strings in EBCDIC
600        environments.
601    
602    21. It appears that in at least one Mac OS environment, the isxdigit() function
603        is implemented as a macro that evaluates to its argument more than once,
604        contravening the C 90 Standard (I haven't checked a later standard). There
605        was an instance in pcretest which caused it to go wrong when processing
606        \x{...} escapes in subject strings. The has been rewritten to avoid using
607        things like p++ in the argument of isxdigit().
608    
609    
610    Version 8.13 16-Aug-2011
611    ------------------------
612    
613    1.  The Unicode data tables have been updated to Unicode 6.0.0.
614    
615    2.  Two minor typos in pcre_internal.h have been fixed.
616    
617    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
618        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
619        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
620    
621    4.  There were a number of related bugs in the code for matching backrefences
622        caselessly in UTF-8 mode when codes for the characters concerned were
623        different numbers of bytes. For example, U+023A and U+2C65 are an upper
624        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
625        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
626        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
627        2-byte code at the end of the subject (it thought there wasn't enough data
628        left).
629    
630    5.  Comprehensive information about what went wrong is now returned by
631        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
632        as the output vector has at least 2 elements. The offset of the start of
633        the failing character and a reason code are placed in the vector.
634    
635    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
636        now returned is for the first byte of the failing character, instead of the
637        last byte inspected. This is an incompatible change, but I hope it is small
638        enough not to be a problem. It makes the returned offset consistent with
639        pcre_exec() and pcre_dfa_exec().
640    
641    7.  pcretest now gives a text phrase as well as the error number when
642        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
643        failure, the offset and reason code are output.
644    
645    8.  When \R was used with a maximizing quantifier it failed to skip backwards
646        over a \r\n pair if the subsequent match failed. Instead, it just skipped
647        back over a single character (\n). This seems wrong (because it treated the
648        two characters as a single entity when going forwards), conflicts with the
649        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
650        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
651        has been changed.
652    
653    9.  Some internal refactoring has changed the processing so that the handling
654        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
655        time (the PCRE_DOTALL option was changed this way some time ago: version
656        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
657        which was always a bit of a fudge. It also means that there is one less
658        argument for the match() function, which reduces its stack requirements
659        slightly. This change also fixes an incompatibility with Perl: the pattern
660        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
661    
662    10. More internal refactoring has drastically reduced the number of recursive
663        calls to match() for possessively repeated groups such as (abc)++ when
664        using pcre_exec().
665    
666    11. While implementing 10, a number of bugs in the handling of groups were
667        discovered and fixed:
668    
669        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
670        (a|)*(?1) gave a compile-time internal error.
671        ((a|)+)+  did not notice that the outer group could match an empty string.
672        (^a|^)+   was not marked as anchored.
673        (.*a|.*)+ was not marked as matching at start or after a newline.
674    
675    12. Yet more internal refactoring has removed another argument from the match()
676        function. Special calls to this function are now indicated by setting a
677        value in a variable in the "match data" data block.
678    
679    13. Be more explicit in pcre_study() instead of relying on "default" for
680        opcodes that mean there is no starting character; this means that when new
681        ones are added and accidentally left out of pcre_study(), testing should
682        pick them up.
683    
684    14. The -s option of pcretest has been documented for ages as being an old
685        synonym of -m (show memory usage). I have changed it to mean "force study
686        for every regex", that is, assume /S for every regex. This is similar to -i
687        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
688        using it. It makes it easier to run collections of tests with and without
689        study enabled, and thereby test pcre_study() more easily. All the standard
690        tests are now run with and without -s (but some patterns can be marked as
691        "never study" - see 20 below).
692    
693    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
694        restoration of the capturing data to the outer values was not happening
695        correctly.
696    
697    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
698        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
699        pattern had matched an empty string, and so incorrectly returned a no
700        match.
701    
702    17. There was optimizing code for the last branch of non-capturing parentheses,
703        and also for the obeyed branch of a conditional subexpression, which used
704        tail recursion to cut down on stack usage. Unfortunately, now that there is
705        the possibility of (*THEN) occurring in these branches, tail recursion is
706        no longer possible because the return has to be checked for (*THEN). These
707        two optimizations have therefore been removed. [But see 8.20/11 above.]
708    
709    18. If a pattern containing \R was studied, it was assumed that \R always
710        matched two bytes, thus causing the minimum subject length to be
711        incorrectly computed because \R can also match just one byte.
712    
713    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
714        was incorrectly computed.
715    
716    20. If /S is present twice on a test pattern in pcretest input, it now
717        *disables* studying, thereby overriding the use of -s on the command line
718        (see 14 above). This is necessary for one or two tests to keep the output
719        identical in both cases.
720    
721    21. When (*ACCEPT) was used in an assertion that matched an empty string and
722        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
723    
724    22. When an atomic group that contained a capturing parenthesis was
725        successfully matched, but the branch in which it appeared failed, the
726        capturing was not being forgotten if a higher numbered group was later
727        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
728        group 1 to "a", when in fact it should be unset. This applied to multi-
729        branched capturing and non-capturing groups, repeated or not, and also to
730        positive assertions (capturing in negative assertions does not happen
731        in PCRE) and also to nested atomic groups.
732    
733    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
734        subject after a captured substring, to make it easier to tell which of a
735        number of identical substrings has been captured.
736    
737    24. The way atomic groups are processed by pcre_exec() has been changed so that
738        if they are repeated, backtracking one repetition now resets captured
739        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
740        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
741        "aaa". Previously, it would have been "a". As part of this code
742        refactoring, the way recursive calls are handled has also been changed.
743    
744    25. If an assertion condition captured any substrings, they were not passed
745        back unless some other capturing happened later. For example, if
746        (?(?=(a))a) was matched against "a", no capturing was returned.
747    
748    26. When studying a pattern that contained subroutine calls or assertions,
749        the code for finding the minimum length of a possible match was handling
750        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
751        group 1 called group 2 while simultaneously a separate group 2 called group
752        1). A stack overflow occurred in this case. I have fixed this by limiting
753        the recursion depth to 10.
754    
755    27. Updated RunTest.bat in the distribution to the version supplied by Tom
756        Fortmann. This supports explicit test numbers on the command line, and has
757        argument validation and error reporting.
758    
759    28. An instance of \X with an unlimited repeat could fail if at any point the
760        first character it looked at was a mark character.
761    
762    29. Some minor code refactoring concerning Unicode properties and scripts
763        should reduce the stack requirement of match() slightly.
764    
765    30. Added the '=' option to pcretest to check the setting of unused capturing
766        slots at the end of the pattern, which are documented as being -1, but are
767        not included in the return count.
768    
769    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
770        compiled something random. Now it gives a compile-time error (as does
771        Perl).
772    
773    32. A *MARK encountered during the processing of a positive assertion is now
774        recorded and passed back (compatible with Perl).
775    
776    33. If --only-matching or --colour was set on a pcregrep call whose pattern
777        had alternative anchored branches, the search for a second match in a line
778        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
779        matched the line "0102" twice. The same bug affected patterns that started
780        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
781        twice.
782    
783    34. Previously, PCRE did not allow quantification of assertions. However, Perl
784        does, and because of capturing effects, quantifying parenthesized
785        assertions may at times be useful. Quantifiers are now allowed for
786        parenthesized assertions.
787    
788    35. A minor code tidy in pcre_compile() when checking options for \R usage.
789    
790    36. \g was being checked for fancy things in a character class, when it should
791        just be a literal "g".
792    
793    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
794        appearance of a nested POSIX class supersedes an apparent external class.
795        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
796        unescaped square brackets may also appear as part of class names. For
797        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
798        more like Perl. (But see 8.20/1 above.)
799    
800    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
801        was because it thought it was \N{name}, which is not supported).
802    
803    39. Add minix to OS list not supporting the -S option in pcretest.
804    
805    40. PCRE tries to detect cases of infinite recursion at compile time, but it
806        cannot analyze patterns in sufficient detail to catch mutual recursions
807        such as ((?1))((?2)). There is now a runtime test that gives an error if a
808        subgroup is called recursively as a subpattern for a second time at the
809        same position in the subject string. In previous releases this might have
810        been caught by the recursion limit, or it might have run out of stack.
811    
812    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
813        happen only once. PCRE was, however incorrectly giving a compile time error
814        "recursive call could loop indefinitely" because it cannot analyze the
815        pattern in sufficient detail. The compile time test no longer happens when
816        PCRE is compiling a conditional subpattern, but actual runaway loops are
817        now caught at runtime (see 40 above).
818    
819    42. It seems that Perl allows any characters other than a closing parenthesis
820        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
821        has been changed to be the same.
822    
823    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
824        as not to get warnings when autogen.sh is called. Also changed
825        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
826    
827    44. To help people who use pcregrep to scan files containing exceedingly long
828        lines, the following changes have been made:
829    
830        (a) The default value of the buffer size parameter has been increased from
831            8K to 20K. (The actual buffer used is three times this size.)
832    
833        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
834            PCRE is built.
835    
836        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
837            to be set at run time.
838    
839        (d) Numerical values in pcregrep options can be followed by K or M, for
840            example --buffer-size=50K.
841    
842        (e) If a line being scanned overflows pcregrep's buffer, an error is now
843            given and the return code is set to 2.
844    
845    45. Add a pointer to the latest mark to the callout data block.
846    
847    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
848        partial match of an empty string instead of no match. This was specific to
849        the use of ".".
850    
851    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
852        complete match instead of a partial match. This bug was dependent on both
853        the PCRE_UTF8 and PCRE_DOTALL options being set.
854    
855    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
856        starting byte set, because \b was not being ignored.
857    
858    
859    Version 8.12 15-Jan-2011
860    ------------------------
861    
862    1.  Fixed some typos in the markup of the man pages, and wrote a script that
863        checks for such things as part of the documentation building process.
864    
865    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
866        --match-limit and --recursion-limit options (added for 8.11). In
867        particular, this made one of the standard tests fail. (The integer value
868        went into the wrong half of a long int.)
869    
870    3.  If the --colour option was given to pcregrep with -v (invert match), it
871        did strange things, either producing crazy output, or crashing. It should,
872        of course, ignore a request for colour when reporting lines that do not
873        match.
874    
875    4.  Another pcregrep bug caused similar problems if --colour was specified with
876        -M (multiline) and the pattern match finished with a line ending.
877    
878    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
879        matched in multiline mode, the following line was shown as part of the
880        match. This seems wrong, so I have changed it.
881    
882    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
883        the check for further matches in the same line (so they could be coloured)
884        to overrun the end of the current line. If another match was found, it was
885        incorrectly shown (and then shown again when found in the next line).
886    
887    7.  If pcregrep was compiled under Windows, there was a reference to the
888        function pcregrep_exit() before it was defined. I am assuming this was
889        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
890        reported by a user. I've moved the definition above the reference.
891    
892    
893    Version 8.11 10-Dec-2010
894    ------------------------
895    
896    1.  (*THEN) was not working properly if there were untried alternatives prior
897        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
898        backtracked to try for "b" instead of moving to the next alternative branch
899        at the same level (in this case, to look for "c"). The Perl documentation
900        is clear that when (*THEN) is backtracked onto, it goes to the "next
901        alternative in the innermost enclosing group".
902    
903    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
904        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
905        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
906        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
907        (*THEN).
908    
909    3.  If \s appeared in a character class, it removed the VT character from
910        the class, even if it had been included by some previous item, for example
911        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
912        of \s, but is part of the POSIX "space" class.)
913    
914    4.  A partial match never returns an empty string (because you can always
915        match an empty string at the end of the subject); however the checking for
916        an empty string was starting at the "start of match" point. This has been
917        changed to the "earliest inspected character" point, because the returned
918        data for a partial match starts at this character. This means that, for
919        example, /(?<=abc)def/ gives a partial match for the subject "abc"
920        (previously it gave "no match").
921    
922    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
923        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
924        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
925        has an implication that the given string is incomplete (because a partial
926        match is preferred over a full match). For this reason, these items now
927        give a partial match in this situation. [Aside: previously, the one case
928        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
929        match rather than a full match, which was wrong by the old rules, but is
930        now correct.]
931    
932    6.  There was a bug in the handling of #-introduced comments, recognized when
933        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
934        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
935        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
936        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
937        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
938        places in pcre_compile().
939    
940    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
941        comments when looking ahead for named forward references to subpatterns,
942        the only newline sequence it recognized was NL. It now handles newlines
943        according to the set newline convention.
944    
945    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
946        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
947        cater for a lack of strerror(). These oversights have been fixed.
948    
949    9.  Added --match-limit and --recursion-limit to pcregrep.
950    
951    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
952    
953    11. When the -o option was used, pcregrep was setting a return code of 1, even
954        when matches were found, and --line-buffered was not being honoured.
955    
956    12. Added an optional parentheses number to the -o and --only-matching options
957        of pcregrep.
958    
959    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
960        can match an empty string. The code to do it in pcretest and pcredemo
961        needed fixing:
962    
963        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
964            only one byte after an empty string match just before CRLF (this case
965            just got forgotten; "any" and "anycrlf" were OK).
966    
967        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
968            mode when an empty string match preceded an ASCII character followed by
969            a non-ASCII character. (The code for advancing by one character rather
970            than one byte was nonsense.)
971    
972        (c) The pcredemo.c sample program did not have any code at all to handle
973            the cases when CRLF is a valid newline sequence.
974    
975    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
976        as a starting offset was within the subject string. There is now a new
977        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
978        negative or greater than the length of the string. In order to test this,
979        pcretest is extended to allow the setting of negative starting offsets.
980    
981    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
982        starting offset points to the beginning of a UTF-8 character was
983        unnecessarily clumsy. I tidied it up.
984    
985    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
986        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
987    
988    17. Nobody had reported that the --include_dir option, which was added in
989        release 7.7 should have been called --include-dir (hyphen, not underscore)
990        for compatibility with GNU grep. I have changed it to --include-dir, but
991        left --include_dir as an undocumented synonym, and the same for
992        --exclude-dir, though that is not available in GNU grep, at least as of
993        release 2.5.4.
994    
995    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
996        characters from a string of bytes) have been redefined so as not to use
997        loops, in order to improve performance in some environments. At the same
998        time, I abstracted some of the common code into auxiliary macros to save
999        repetition (this should not affect the compiled code).
1000    
1001    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1002        compile-time error is now given if \c is not followed by an ASCII
1003        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1004        different, and any byte value is allowed.)
1005    
1006    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1007        START_OPTIMIZE option, which is now allowed at compile time - but just
1008        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1009        to pcregrep and other applications that have no direct access to PCRE
1010        options. The new /Y option in pcretest sets this option when calling
1011        pcre_compile().
1012    
1013    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1014        back references. Groups containing recursive back references were forced to
1015        be atomic by that change, but in the case of named groups, the amount of
1016        memory required was incorrectly computed, leading to "Failed: internal
1017        error: code overflow". This has been fixed.
1018    
1019    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1020        pcretest.c, to avoid build problems in some Borland environments.
1021    
1022    
1023    Version 8.10 25-Jun-2010
1024    ------------------------
1025    
1026    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1027        THEN.
1028    
1029    2.  (*ACCEPT) was not working when inside an atomic group.
1030    
1031    3.  Inside a character class, \B is treated as a literal by default, but
1032        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1033        causes the error). The code is unchanged, but I tidied the documentation.
1034    
1035    4.  Inside a character class, PCRE always treated \R and \X as literals,
1036        whereas Perl faults them if its -w option is set. I have changed PCRE so
1037        that it faults them when PCRE_EXTRA is set.
1038    
1039    5.  Added support for \N, which always matches any character other than
1040        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1041    
1042    6.  When compiling pcregrep with newer versions of gcc which may have
1043        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1044        declared with attribute warn_unused_result" were given. Just casting the
1045        result to (void) does not stop the warnings; a more elaborate fudge is
1046        needed. I've used a macro to implement this.
1047    
1048    7.  Minor change to pcretest.c to avoid a compiler warning.
1049    
1050    8.  Added four artifical Unicode properties to help with an option to make
1051        \s etc use properties (see next item). The new properties are: Xan
1052        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1053    
1054    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1055        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1056        this option. Modified pcretest to add /W to test this facility. Added
1057        REG_UCP to make it available via the POSIX interface.
1058    
1059    10. Added --line-buffered to pcregrep.
1060    
1061    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1062        studied, and the match started with a letter with a code point greater than
1063        127 whose first byte was different to the first byte of the other case of
1064        the letter, the other case of this starting letter was not recognized
1065        (#976).
1066    
1067    12. If a pattern that was studied started with a repeated Unicode property
1068        test, for example, \p{Nd}+, there was the theoretical possibility of
1069        setting up an incorrect bitmap of starting bytes, but fortunately it could
1070        not have actually happened in practice until change 8 above was made (it
1071        added property types that matched character-matching opcodes).
1072    
1073    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1074        possible starting bytes for non-anchored patterns.
1075    
1076    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1077        \R, and also a number of cases that involve Unicode properties, both
1078        explicit and implicit when PCRE_UCP is set.
1079    
1080    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1081        input, it could crash or give wrong results if characters with values
1082        greater than 0xc0 were present in the subject string. (Detail: it assumed
1083        UTF-8 input when processing these items.)
1084    
1085    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1086        size_t is 64-bit (#991).
1087    
1088    17. Added a check for running out of memory when PCRE is compiled with
1089        --disable-stack-for-recursion (#990).
1090    
1091    18. If the last data line in a file for pcretest does not have a newline on
1092        the end, a newline was missing in the output.
1093    
1094    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1095        less than 128) in its various bitmaps. However, there is a facility for
1096        generating tables according to the current locale when PCRE is compiled. It
1097        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1098        characters, are recognized by isspace() and therefore were getting set in
1099        these tables, and indeed these tables seem to approximate to ISO 8859. This
1100        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1101        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1102        which of course cannot start UTF-8 characters. I have changed the code so
1103        that only real ASCII characters (less than 128) and the correct starting
1104        bytes for UTF-8 encodings are set for characters greater than 127 when in
1105        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1106        altogether.)
1107    
1108    20. Added the /T option to pcretest so as to be able to run tests with non-
1109        standard character tables, thus making it possible to include the tests
1110        used for 19 above in the standard set of tests.
1111    
1112    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1113        reference to a subpattern the other side of a comment that contains an
1114        opening parenthesis caused either an internal compiling error, or a
1115        reference to the wrong subpattern.
1116    
1117    
1118    Version 8.02 19-Mar-2010
1119    ------------------------
1120    
1121    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1122    
1123    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1124        configured.
1125    
1126    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1127        original author of that file, following a query about its status.
1128    
1129    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1130        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1131    
1132    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1133        quantifier applied to a forward-referencing subroutine call, could compile
1134        incorrect code or give the error "internal error: previously-checked
1135        referenced subpattern not found".
1136    
1137    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1138        variables to point to external functions. For these systems, therefore,
1139        pcre_malloc etc. are now initialized to local functions that call the
1140        relevant global functions.
1141    
1142    7.  There were two entries missing in the vectors called coptable and poptable
1143        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1144        I've fixed the data, and added a kludgy way of testing at compile time that
1145        the lengths are correct (equal to the number of opcodes).
1146    
1147    8.  Following on from 7, I added a similar kludge to check the length of the
1148        eint vector in pcreposix.c.
1149    
1150    9.  Error texts for pcre_compile() are held as one long string to avoid too
1151        much relocation at load time. To find a text, the string is searched,
1152        counting zeros. There was no check for running off the end of the string,
1153        which could happen if a new error number was added without updating the
1154        string.
1155    
1156    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1157    
1158    11. \K was not working if it appeared in an atomic group or in a group that
1159        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1160        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1161        the assertion is positive, but not if it is negative.
1162    
1163    12. Change 11 fortuitously reduced the size of the stack frame used in the
1164        "match()" function of pcre_exec.c by one pointer. Forthcoming
1165        implementation of support for (*MARK) will need an extra pointer on the
1166        stack; I have reserved it now, so that the stack frame size does not
1167        decrease.
1168    
1169    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1170        item in branch that calls a recursion is a subroutine call - as in the
1171        second branch in the above example - was incorrectly given the compile-
1172        time error "recursive call could loop indefinitely" because pcre_compile()
1173        was not correctly checking the subroutine for matching a non-empty string.
1174    
1175    14. The checks for overrunning compiling workspace could trigger after an
1176        overrun had occurred. This is a "should never occur" error, but it can be
1177        triggered by pathological patterns such as hundreds of nested parentheses.
1178        The checks now trigger 100 bytes before the end of the workspace.
1179    
1180    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1181    
1182    
1183    Version 8.01 19-Jan-2010
1184    ------------------------
1185    
1186    1.  If a pattern contained a conditional subpattern with only one branch (in
1187        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1188        computed the wrong minimum data length (which is of course zero for such
1189        subpatterns). This could cause incorrect "no match" results.
1190    
1191    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1192        the pattern is reset in the first branch, pcre_compile() failed with
1193        "internal error: code overflow at offset...". This happened only when
1194        the reset was to the original external option setting. (An optimization
1195        abstracts leading options settings into an external setting, which was the
1196        cause of this.)
1197    
1198    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1199        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1200        assertion pattern did not match (meaning that the assertion was true), it
1201        was incorrectly treated as false if the SKIP had been reached during the
1202        matching. This also applied to assertions used as conditions.
1203    
1204    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1205        assertion subpattern, including such a pattern used as a condition,
1206        unpredictable results occurred, instead of the error return
1207        PCRE_ERROR_DFA_UITEM.
1208    
1209    5.  The C++ GlobalReplace function was not working like Perl for the special
1210        situation when an empty string is matched. It now does the fancy magic
1211        stuff that is necessary.
1212    
1213    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1214        removed. (These were left over from very, very early versions of PCRE.)
1215    
1216    7.  Some cosmetic changes to the code to make life easier when compiling it
1217        as part of something else:
1218    
1219        (a) Change DEBUG to PCRE_DEBUG.
1220    
1221        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1222            called "current" as "current_branch", to prevent a collision with the
1223            Linux macro when compiled as a kernel module.
1224    
1225        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1226            prevent a collision with the Linux macro when compiled as a kernel
1227            module.
1228    
1229    8.  In pcre_compile() there are some checks for integer overflows that used to
1230        cast potentially large values to (double). This has been changed to that
1231        when building, a check for int64_t is made, and if it is found, it is used
1232        instead, thus avoiding the use of floating point arithmetic. (There is no
1233        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1234        double.
1235    
1236    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1237        2005 (difference between two addresses compared to an unsigned value).
1238    
1239    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1240        custom one, because of the following reported problem in Windows:
1241    
1242          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1243              under Win32.
1244          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1245              therefore missing the function definition.
1246          - The compiler thus generates a "C" signature for the test function.
1247          - The linker fails to find the "C" function.
1248          - PCRE fails to configure if asked to do so against libbz2.
1249    
1250    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1251        messages were output:
1252    
1253          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1254          rerunning libtoolize, to keep the correct libtool macros in-tree.
1255          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1256    
1257        I have done both of these things.
1258    
1259    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1260        most of the time, it *can* run out if it is given a pattern that contains a
1261        runaway infinite recursion. I updated the discussion in the pcrestack man
1262        page.
1263    
1264    13. Now that we have gone to the x.xx style of version numbers, the minor
1265        version may start with zero. Using 08 or 09 is a bad idea because users
1266        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1267        interpreted as invalid octal numbers. I've updated the previous comment in
1268        configure.ac, and also added a check that gives an error if 08 or 09 are
1269        used.
1270    
1271    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1272        causing partial matching to fail when the end of the subject matched \W
1273        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1274    
1275    15. There were some discrepancies between the declarations in pcre_internal.h
1276        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1277        their definitions. The declarations used "const uschar *" and the
1278        definitions used USPTR. Even though USPTR is normally defined as "const
1279        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1280        reported that: "This difference in casting confuses some C++ compilers, for
1281        example, SunCC recognizes above declarations as different functions and
1282        generates broken code for hbpcre." I have changed the declarations to use
1283        USPTR.
1284    
1285    16. GNU libtool is named differently on some systems. The autogen.sh script now
1286        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1287        (FreeBSD).
1288    
1289    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1290        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1291        comment: "Figure out how to create a longlong from a string: strtoll and
1292        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1293        instance, but it only takes 2 args instead of 3!"
1294    
1295    18. A subtle bug concerned with back references has been fixed by a change of
1296        specification, with a corresponding code fix. A pattern such as
1297        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1298        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1299        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1300        same bug. Such groups have to be quantified to be useful, or contained
1301        inside another quantified group. (If there's no repetition, the reference
1302        can never match.) The problem arises because, having left the group and
1303        moved on to the rest of the pattern, a later failure that backtracks into
1304        the group uses the captured value from the final iteration of the group
1305        rather than the correct earlier one. I have fixed this in PCRE by forcing
1306        any group that contains a reference to itself to be an atomic group; that
1307        is, there cannot be any backtracking into it once it has completed. This is
1308        similar to recursive and subroutine calls.
1309    
1310    
1311    Version 8.00 19-Oct-09
1312    ----------------------
1313    
1314    1.  The table for translating pcre_compile() error codes into POSIX error codes
1315        was out-of-date, and there was no check on the pcre_compile() error code
1316        being within the table. This could lead to an OK return being given in
1317        error.
1318    
1319    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1320        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1321        in a Windows environment.
1322    
1323    3.  The pcregrep --count option prints the count for each file even when it is
1324        zero, as does GNU grep. However, pcregrep was also printing all files when
1325        --files-with-matches was added. Now, when both options are given, it prints
1326        counts only for those files that have at least one match. (GNU grep just
1327        prints the file name in this circumstance, but including the count seems
1328        more useful - otherwise, why use --count?) Also ensured that the
1329        combination -clh just lists non-zero counts, with no names.
1330    
1331    4.  The long form of the pcregrep -F option was incorrectly implemented as
1332        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1333        but it seems right to fix it, and I didn't think it was worth preserving
1334        the old behaviour.
1335    
1336    5.  The command line items --regex=pattern and --regexp=pattern were not
1337        recognized by pcregrep, which required --regex pattern or --regexp pattern
1338        (with a space rather than an '='). The man page documented the '=' forms,
1339        which are compatible with GNU grep; these now work.
1340    
1341    6.  No libpcreposix.pc file was created for pkg-config; there was just
1342        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1343    
1344    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1345        when UCP support is not needed, by modifying the Python script that
1346        generates it from Unicode data files. This should not matter if the module
1347        is correctly used as a library, but I received one complaint about 50K of
1348        unwanted data. My guess is that the person linked everything into his
1349        program rather than using a library. Anyway, it does no harm.
1350    
1351    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1352        was a minimum greater than 1 for a wide character in a possessive
1353        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1354        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1355        character. Chaos in the form of incorrect output or a compiling loop could
1356        result.
1357    
1358    9.  The restrictions on what a pattern can contain when partial matching is
1359        requested for pcre_exec() have been removed. All patterns can now be
1360        partially matched by this function. In addition, if there are at least two
1361        slots in the offset vector, the offset of the earliest inspected character
1362        for the match and the offset of the end of the subject are set in them when
1363        PCRE_ERROR_PARTIAL is returned.
1364    
1365    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1366        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1367        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1368        and may be more useful for multi-segment matching.
1369    
1370    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1371        used to be given if ever the end of the subject was reached; now it is
1372        given only if matching could not proceed because another character was
1373        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1374        string "Z", which now yields "no match" instead of "partial match". In the
1375        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1376        final character ended with (*FAIL).
1377    
1378    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1379        if the pattern had a "must contain" character that was already found in the
1380        earlier partial match, unless partial matching was again requested. For
1381        example, with the pattern /dog.(body)?/, the "must contain" character is
1382        "g". If the first part-match was for the string "dog", restarting with
1383        "sbody" failed. This bug has been fixed.
1384    
1385    13. The string returned by pcre_dfa_exec() after a partial match has been
1386        changed so that it starts at the first inspected character rather than the
1387        first character of the match. This makes a difference only if the pattern
1388        starts with a lookbehind assertion or \b or \B (\K is not supported by
1389        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1390        matching functions compatible, and I think it's the right thing to do.
1391    
1392    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1393        so that the demonstration program is easily available in environments where
1394        PCRE has not been installed from source.
1395    
1396    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1397        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1398        library.
1399    
1400    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1401        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1402        is not the first non-POSIX option to be added. Clearly some people find
1403        these options useful.
1404    
1405    17. If a caller to the POSIX matching function regexec() passes a non-zero
1406        value for nmatch with a NULL value for pmatch, the value of
1407        nmatch is forced to zero.
1408    
1409    18. RunGrepTest did not have a test for the availability of the -u option of
1410        the diff command, as RunTest does. It now checks in the same way as
1411        RunTest, and also checks for the -b option.
1412    
1413    19. If an odd number of negated classes containing just a single character
1414        interposed, within parentheses, between a forward reference to a named
1415        subpattern and the definition of the subpattern, compilation crashed with
1416        an internal error, complaining that it could not find the referenced
1417        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1418        [The bug was that it was starting one character too far in when skipping
1419        over the character class, thus treating the ] as data rather than
1420        terminating the class. This meant it could skip too much.]
1421    
1422    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1423        /g option in pcretest when the pattern contains \K, which makes it possible
1424        to have an empty string match not at the start, even when the pattern is
1425        anchored. Updated pcretest and pcredemo to use this option.
1426    
1427    21. If the maximum number of capturing subpatterns in a recursion was greater
1428        than the maximum at the outer level, the higher number was returned, but
1429        with unset values at the outer level. The correct (outer level) value is
1430        now given.
1431    
1432    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1433        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1434        make it do so. The string so far is captured, making this feature
1435        compatible with Perl.
1436    
1437    23. The tests have been re-organized, adding tests 11 and 12, to make it
1438        possible to check the Perl 5.10 features against Perl 5.10.
1439    
1440    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1441        pattern matches a fixed length string. PCRE did not allow this; now it
1442        does. Neither allows recursion.
1443    
1444    25. I finally figured out how to implement a request to provide the minimum
1445        length of subject string that was needed in order to match a given pattern.
1446        (It was back references and recursion that I had previously got hung up
1447        on.) This code has now been added to pcre_study(); it finds a lower bound
1448        to the length of subject needed. It is not necessarily the greatest lower
1449        bound, but using it to avoid searching strings that are too short does give
1450        some useful speed-ups. The value is available to calling programs via
1451        pcre_fullinfo().
1452    
1453    26. While implementing 25, I discovered to my embarrassment that pcretest had
1454        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1455        study optimizations had never been tested with that matching function.
1456        Oops. What is worse, even when it was passed study data, there was a bug in
1457        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1458        were also very few tests of studied patterns with pcre_dfa_exec().
1459    
1460    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1461        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1462        on the other side of the coin, they are no longer allowed to have different
1463        names, because these cannot be distinguished in PCRE, and this has caused
1464        confusion. (This is a difference from Perl.)
1465    
1466    28. When duplicate subpattern names are present (necessarily with different
1467        numbers, as required by 27 above), and a test is made by name in a
1468        conditional pattern, either for a subpattern having been matched, or for
1469        recursion in such a pattern, all the associated numbered subpatterns are
1470        tested, and the overall condition is true if the condition is true for any
1471        one of them. This is the way Perl works, and is also more like the way
1472        testing by number works.
1473    
1474    
1475    Version 7.9 11-Apr-09
1476    ---------------------
1477    
1478    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1479        (pcretest), all targets were linked against these libraries. This included
1480        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1481        libraries. This caused unwanted dependencies to be created. This problem
1482        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1483        pcretest is linked with readline.
1484    
1485    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1486        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1487        moved outside it again, because FALSE and TRUE are already defined in AIX,
1488        but BOOL is not.
1489    
1490    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1491        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1492    
1493    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1494        hyphens) following file names and line numbers when outputting matching
1495        lines. This is not true; no spaces are inserted. I have also clarified the
1496        wording for the --colour (or --color) option.
1497    
1498    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1499        was not coloured; this is different to GNU grep, so I have changed it to be
1500        the same.
1501    
1502    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1503        each matching line was coloured. Now it goes on to look for further matches
1504        of any of the test patterns, which is the same behaviour as GNU grep.
1505    
1506    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1507        doesn't make sense to accept an empty string match in pcregrep, so I have
1508        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1509        seems to be how GNU grep behaves.
1510    
1511    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1512        start or after a newline", because the conditional assertion was not being
1513        correctly handled. The rule now is that both the assertion and what follows
1514        in the first alternative must satisfy the test.
1515    
1516    9.  If auto-callout was enabled in a pattern with a conditional group whose
1517        condition was an assertion, PCRE could crash during matching, both with
1518        pcre_exec() and pcre_dfa_exec().
1519    
1520    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1521        used for matching.
1522    
1523    11. Unicode property support in character classes was not working for
1524        characters (bytes) greater than 127 when not in UTF-8 mode.
1525    
1526    12. Added the -M command line option to pcretest.
1527    
1528    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1529    
1530    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1531    
1532    16. Added comments and documentation about mis-use of no_arg in the C++
1533        wrapper.
1534    
1535    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1536        from Martin Jerabek that uses macro names for all relevant character and
1537        string constants.
1538    
1539    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1540        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1541        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1542        these, but not everybody uses configure.
1543    
1544    19. A conditional group that had only one branch was not being correctly
1545        recognized as an item that could match an empty string. This meant that an
1546        enclosing group might also not be so recognized, causing infinite looping
1547        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1548        with the subject "ab", where knowledge that the repeated group can match
1549        nothing is needed in order to break the loop.
1550    
1551    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1552        exec(), but without supplying a callout function, matching went wrong.
1553    
1554    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1555        leak if the size of the offset vector was greater than 30. When the vector
1556        is smaller, the saved offsets during recursion go onto a local stack
1557        vector, but for larger vectors malloc() is used. It was failing to free
1558        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1559        error, in fact).
1560    
1561    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1562        heapframe that is used only when UTF-8 support is enabled. This caused no
1563        problem, but was untidy.
1564    
1565    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1566        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1567        included within another project.
1568    
1569    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1570        slightly modified by me:
1571    
1572          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1573              not building pcregrep.
1574    
1575          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1576              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1577    
1578    25. Forward references, both numeric and by name, in patterns that made use of
1579        duplicate group numbers, could behave incorrectly or give incorrect errors,
1580        because when scanning forward to find the reference group, PCRE was not
1581        taking into account the duplicate group numbers. A pattern such as
1582        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1583    
1584    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1585        the feature of a custom pointer more persuasive (as requested by a user).
1586    
1587    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1588        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1589        pre-defined.
1590    
1591    28. Added support for (*UTF8) at the start of a pattern.
1592    
1593    29. Arrange for flags added by the "release type" setting in CMake to be shown
1594        in the configuration summary.
1595    
1596    
1597    Version 7.8 05-Sep-08
1598    ---------------------
1599    
1600    1.  Replaced UCP searching code with optimized version as implemented for Ad
1601        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1602        stage table and inline lookup instead of a function, giving speed ups of 2
1603        to 5 times on some simple patterns that I tested. Permission was given to
1604        distribute the MultiStage2.py script that generates the tables (it's not in
1605        the tarball, but is in the Subversion repository).
1606    
1607    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1608        scripts.
1609    
1610    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1611        a group with a zero qualifier. The result of the study could be incorrect,
1612        or the function might crash, depending on the pattern.
1613    
1614    4.  Caseless matching was not working for non-ASCII characters in back
1615        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1616        It now works when Unicode Property Support is available.
1617    
1618    5.  In pcretest, an escape such as \x{de} in the data was always generating
1619        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1620        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1621        truncation.
1622    
1623    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1624    
1625    7.  Added two (int) casts to pcregrep when printing the difference of two
1626        pointers, in case they are 64-bit values.
1627    
1628    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1629        test 2 if it fails.
1630    
1631    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1632        and a #define of that name to empty if it is not externally set. This is to
1633        allow users of MSVC to set it if necessary.
1634    
1635    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1636        the convenience functions in the pcre_get.c source file.
1637    
1638    11. An option change at the start of a pattern that had top-level alternatives
1639        could cause overwriting and/or a crash. This command provoked a crash in
1640        some environments:
1641    
1642          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1643    
1644        This potential security problem was recorded as CVE-2008-2371.
1645    
1646    12. For a pattern where the match had to start at the beginning or immediately
1647        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1648        pcre_dfa_exec() could read past the end of the passed subject if there was
1649        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1650        pcretest so that it places the subject at the end of its malloc-ed buffer.
1651    
1652    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1653        exec() might read past the end of the data buffer in UTF-8 mode.
1654    
1655    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1656        the data contained the byte 0x85 as part of a UTF-8 character within its
1657        first line. This applied both to normal and DFA matching.
1658    
1659    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1660        /^[^d]*?$/8 failed to match "abc".
1661    
1662    16. Added a missing copyright notice to pcrecpp_internal.h.
1663    
1664    17. Make it more clear in the documentation that values returned from
1665        pcre_exec() in ovector are byte offsets, not character counts.
1666    
1667    18. Tidied a few places to stop certain compilers from issuing warnings.
1668    
1669    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1670        supplied by Stefan Weber. I made a further small update for 7.8 because
1671        there is a change of source arrangements: the pcre_searchfuncs.c module is
1672        replaced by pcre_ucd.c.
1673    
1674    
1675    Version 7.7 07-May-08
1676    ---------------------
1677    
1678    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1679        a string to a long long, pretend we don't even have a long long." This is
1680        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1681    
1682    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1683        pre-7.6 versions, which defined a global no_arg variable instead of putting
1684        it in the RE class. (See also #8 below.)
1685    
1686    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1687        Lopes.
1688    
1689    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1690    
1691        (1) The include/exclude patterns were being applied to the whole pathnames
1692            of files, instead of just to the final components.
1693    
1694        (2) If there was more than one level of directory, the subdirectories were
1695            skipped unless they satisfied the include/exclude conditions. This is
1696            inconsistent with GNU grep (and could even be seen as contrary to the
1697            pcregrep specification - which I improved to make it absolutely clear).
1698            The action now is always to scan all levels of directory, and just
1699            apply the include/exclude patterns to regular files.
1700    
1701    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1702        --exclude_dir in the tests to avoid scanning .svn directories.
1703    
1704    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1705        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1706        doesn't support NULs in patterns.
1707    
1708    7.  Added some missing "const"s to declarations of static tables in
1709        pcre_compile.c and pcre_dfa_exec.c.
1710    
1711    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1712        caused by fix #2  above. (Subsequently also a second patch to fix the
1713        first patch. And a third patch - this was a messy problem.)
1714    
1715    9.  Applied Craig's patch to remove the use of push_back().
1716    
1717    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1718        matching function regexec().
1719    
1720    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1721        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1722        references. PCRE supports relative numbers with this syntax (I don't think
1723        Oniguruma does).
1724    
1725    12. Previously, a group with a zero repeat such as (...){0} was completely
1726        omitted from the compiled regex. However, this means that if the group
1727        was called as a subroutine from elsewhere in the pattern, things went wrong
1728        (an internal error was given). Such groups are now left in the compiled
1729        pattern, with a new opcode that causes them to be skipped at execution
1730        time.
1731    
1732    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1733        to the way PCRE behaves:
1734    
1735        (a) A lone ] character is dis-allowed (Perl treats it as data).
1736    
1737        (b) A back reference to an unmatched subpattern matches an empty string
1738            (Perl fails the current match path).
1739    
1740        (c) A data ] in a character class must be notated as \] because if the
1741            first data character in a class is ], it defines an empty class. (In
1742            Perl it is not possible to have an empty class.) The empty class []
1743            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1744            The negative empty class [^] matches any one character, independently
1745            of the DOTALL setting.
1746    
1747    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1748        non-existent subpattern following a character class starting with ']' and
1749        containing () gave an internal compiling error instead of "reference to
1750        non-existent subpattern". Fortunately, when the pattern did exist, the
1751        compiled code was correct. (When scanning forwards to check for the
1752        existencd of the subpattern, it was treating the data ']' as terminating
1753        the class, so got the count wrong. When actually compiling, the reference
1754        was subsequently set up correctly.)
1755    
1756    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1757        it was being rejected as not supported by pcre_dfa_exec(), even though
1758        other assertions are supported. I have made pcre_dfa_exec() support
1759        (*FAIL).
1760    
1761    16. The implementation of 13c above involved the invention of a new opcode,
1762        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1763        cannot be changed at match time, I realized I could make a small
1764        improvement to matching performance by compiling OP_ALLANY instead of
1765        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1766        on the OP_ANY path.
1767    
1768    17. Compiling pcretest on Windows with readline support failed without the
1769        following two fixes: (1) Make the unistd.h include conditional on
1770        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1771    
1772    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1773        ncurses library to be included for pcretest when ReadLine support is
1774        requested, but also to allow for it to be overridden. This patch came from
1775        Daniel Bergström.
1776    
1777    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1778        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1779        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1780        spotting this.
1781    
1782    
1783    Version 7.6 28-Jan-08
1784    ---------------------
1785    
1786    1.  A character class containing a very large number of characters with
1787        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1788        overflow.
1789    
1790    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1791        HAVE_LONG_LONG is not defined.
1792    
1793    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1794        bring them up to date and include new features. This patch includes:
1795    
1796        - Fixed PH's badly added libz and libbz2 support.
1797        - Fixed a problem with static linking.
1798        - Added pcredemo. [But later removed - see 7 below.]
1799        - Fixed dftables problem and added an option.
1800        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1801            HAVE_LONG_LONG.
1802        - Added readline support for pcretest.
1803        - Added an listing of the option settings after cmake has run.
1804    
1805    4.  A user submitted a patch to Makefile that makes it easy to create
1806        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1807        Makefile.am that cause it to include this special target, without
1808        affecting anything else. Note that the same mingw target plus all
1809        the other distribution libraries and programs are now supported
1810        when configuring with CMake (see 6 below) instead of with
1811        Configure/Make.
1812    
1813    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1814        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1815        exported in the Windows port". It has not yet been confirmed that the patch
1816        solves the problem, but it does no harm.
1817    
1818    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1819        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1820        with CMake, and also correct the comment about stack recursion.
1821    
1822    7.  Remove the automatic building of pcredemo from the ./configure system and
1823        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1824        of a program that users should build themselves after PCRE is installed, so
1825        building it automatically is not really right. What is more, it gave
1826        trouble in some build environments.
1827    
1828    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1829    
1830    
1831    Version 7.5 10-Jan-08
1832    ---------------------
1833    
1834    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1835        values in parens when parsing an RE using the C++ wrapper."
1836    
1837    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1838        Characters greater than 255 were excluded from the class instead of being
1839        included.
1840    
1841    3.  The same bug as (2) above applied to negated POSIX classes such as
1842        [:^space:].
1843    
1844    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1845        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1846        I have changed it.
1847    
1848    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1849        first named subpattern) and a construct such as (?&a) would reference the
1850        first named subpattern whose name started with "a" (in other words, the
1851        length check was missing). Both these problems are fixed. "Subpattern name
1852        expected" is now given for (?&) (a zero-length name), and this patch also
1853        makes it give the same error for \k'' (previously it complained that that
1854        was a reference to a non-existent subpattern).
1855    
1856    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1857        this is right because (?- can be followed by option settings as well as by
1858        digits. I have, however, made the messages clearer.
1859    
1860    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1861        than the number used in the conditional) now cause a compile-time error.
1862        This is actually not compatible with Perl, which accepts such patterns, but
1863        treats the conditional as always being FALSE (as PCRE used to), but it
1864        seems to me that giving a diagnostic is better.
1865    
1866    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1867        and messages.
1868    
1869    9.  Fix two occurrences of "backslash" in comments that should have been
1870        "backspace".
1871    
1872    10. Remove two redundant lines of code that can never be obeyed (their function
1873        was moved elsewhere).
1874    
1875    11. The program that makes PCRE's Unicode character property table had a bug
1876        which caused it to generate incorrect table entries for sequences of
1877        characters that have the same character type, but are in different scripts.
1878        It amalgamated them into a single range, with the script of the first of
1879        them. In other words, some characters were in the wrong script. There were
1880        thirteen such cases, affecting characters in the following ranges:
1881    
1882          U+002b0 - U+002c1
1883          U+0060c - U+0060d
1884          U+0061e - U+00612
1885          U+0064b - U+0065e
1886          U+0074d - U+0076d
1887          U+01800 - U+01805
1888          U+01d00 - U+01d77
1889          U+01d9b - U+01dbf
1890          U+0200b - U+0200f
1891          U+030fc - U+030fe
1892          U+03260 - U+0327f
1893          U+0fb46 - U+0fbb1
1894          U+10450 - U+1049d
1895    
1896    12. The -o option (show only the matching part of a line) for pcregrep was not
1897        compatible with GNU grep in that, if there was more than one match in a
1898        line, it showed only the first of them. It now behaves in the same way as
1899        GNU grep.
1900    
1901    13. If the -o and -v options were combined for pcregrep, it printed a blank
1902        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1903        does the same. The return code can be used to tell if there were any
1904        non-matching lines.
1905    
1906    14. Added --file-offsets and --line-offsets to pcregrep.
1907    
1908    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1909        infinitely looping recursion. The bug was that positive lookaheads were not
1910        being skipped when checking for a possible empty match (negative lookaheads
1911        and both kinds of lookbehind were skipped).
1912    
1913    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1914        inclusion of <windows.h> to before rather than after the definition of
1915        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1916    
1917    17. Specifying a possessive quantifier with a specific limit for a Unicode
1918        character property caused pcre_compile() to compile bad code, which led at
1919        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1920        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1921        caused the error; without that there was no problem.
1922    
1923    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1924    
1925    19. Added --enable-pcretest-libreadline.
1926    
1927    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1928        RE::GlobalReplace(). As a result, the number of replacements returned was
1929        double what it should be. I removed one of the increments, but Craig sent a
1930        later patch that removed the other one (the right fix) and added unit tests
1931        that check the return values (which was not done before).
1932    
1933    21. Several CMake things:
1934    
1935        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1936            the names libpcre and libpcreposix, not just pcre and pcreposix.
1937    
1938        (2) The above change means that pcretest and pcregrep are now correctly
1939            linked with the newly-built libraries, not previously installed ones.
1940    
1941        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1942    
1943    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1944        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1945        UTF-8 newline character). The key issue is that the pattern starts .*;
1946        this means that the match must be either at the beginning, or after a
1947        newline. The bug was in the code for advancing after a failed match and
1948        checking that the new position followed a newline. It was not taking
1949        account of UTF-8 characters correctly.
1950    
1951    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1952        character classes. PCRE was not treating the sequence [:...:] as a
1953        character class unless the ... were all letters. Perl, however, seems to
1954        allow any characters between [: and :], though of course it rejects as
1955        unknown any "names" that contain non-letters, because all the known class
1956        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1957        for example, whereas PCRE did not - it did not recognize a POSIX character
1958        class. This seemed a bit dangerous, so the code has been changed to be
1959        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1960        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1961        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1962        Perl does, and where it didn't before.
1963    
1964    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1965        Windows environments %n is disabled by default.
1966    
1967    
1968    Version 7.4 21-Sep-07
1969    ---------------------
1970    
1971    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1972        means that a class such as [\s] counted as "explicit reference to CR or
1973        LF". That isn't really right - the whole point of the change was to try to
1974        help when there was an actual mention of one of the two characters. So now
1975        the change happens only if \r or \n (or a literal CR or LF) character is
1976        encountered.
1977    
1978    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1979        of both had grown to the point where there were only 3 bits left.
1980        Fortunately, there was spare space in the data structure, and so I have
1981        moved the internal flags into a new 16-bit field to free up more option
1982        bits.
1983    
1984    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1985        but did not set the internal JCHANGED flag - either of these is enough to
1986        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1987        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1988        start sets both bits.
1989    
1990    4.  Added options (at build time, compile time, exec time) to change \R from
1991        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1992    
1993    5.  doc/pcresyntax.html was missing from the distribution.
1994    
1995    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1996        compatibility, even though it is no longer used.
1997    
1998    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1999        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2000        windows.h file is present (where different names are used). [This was
2001        reversed later after testing - see 16 below.]
2002    
2003    8.  Changed all #include <config.h> to #include "config.h". There were also
2004        some further <pcre.h> cases that I changed to "pcre.h".
2005    
2006    9.  When pcregrep was used with the --colour option, it missed the line ending
2007        sequence off the lines that it output.
2008    
2009    10. It was pointed out to me that arrays of string pointers cause lots of
2010        relocations when a shared library is dynamically loaded. A technique of
2011        using a single long string with a table of offsets can drastically reduce
2012        these. I have refactored PCRE in four places to do this. The result is
2013        dramatic:
2014    
2015          Originally:                          290
2016          After changing UCP table:            187
2017          After changing error message table:   43
2018          After changing table of "verbs"       36
2019          After changing table of Posix names   22
2020    
2021        Thanks to the folks working on Gregex for glib for this insight.
2022    
2023    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2024        unicode-properties was also set.
2025    
2026    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2027    
2028    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2029        checked only for CRLF.
2030    
2031    14. Added casts to pcretest.c to avoid compiler warnings.
2032    
2033    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2034    
2035    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2036        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2037        entirely. This removes changes made in 7 above.
2038    
2039    17. The CMake files have been updated, and there is now more information about
2040        building with CMake in the NON-UNIX-USE document.
2041    
2042    
2043    Version 7.3 28-Aug-07
2044    ---------------------
2045    
2046     1. In the rejigging of the build system that eventually resulted in 7.1, the
2047        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2048        brackets there is not right, since it causes compilers to look for an
2049        installed pcre.h, not the version that is in the source that is being
2050        compiled (which of course may be different). I have changed it back to:
2051    
2052          #include "pcre.h"
2053    
2054        I have a vague recollection that the change was concerned with compiling in
2055        different directories, but in the new build system, that is taken care of
2056        by the VPATH setting the Makefile.
2057    
2058     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2059        when the subject happened to end in the byte 0x85 (e.g. if the last
2060        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2061        characters but of course it shouldn't be taken as a newline when it is part
2062        of another character. The bug was that, for an unlimited repeat of . in
2063        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2064        characters when looking for a newline.
2065    
2066     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2067    
2068     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2069        in debug output.
2070    
2071     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2072        long printing in the pcrecpp unittest when running under MinGW.
2073    
2074     6. ESC_K was left out of the EBCDIC table.
2075    
2076     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2077        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2078        limit also applies to "virtual nesting" when a pattern is recursive, and in
2079        this case 1000 isn't so big. I have been able to remove this limit at the
2080        expense of backing off one optimization in certain circumstances. Normally,
2081        when pcre_exec() would call its internal match() function recursively and
2082        immediately return the result unconditionally, it uses a "tail recursion"
2083        feature to save stack. However, when a subpattern that can match an empty
2084        string has an unlimited repetition quantifier, it no longer makes this
2085        optimization. That gives it a stack frame in which to save the data for
2086        checking that an empty string has been matched. Previously this was taken
2087        from the 1000-entry workspace that had been reserved. So now there is no
2088        explicit limit, but more stack is used.
2089    
2090     8. Applied Daniel's patches to solve problems with the import/export magic
2091        syntax that is required for Windows, and which was going wrong for the
2092        pcreposix and pcrecpp parts of the library. These were overlooked when this
2093        problem was solved for the main library.
2094    
2095     9. There were some crude static tests to avoid integer overflow when computing
2096        the size of patterns that contain repeated groups with explicit upper
2097        limits. As the maximum quantifier is 65535, the maximum group length was
2098        set at 30,000 so that the product of these two numbers did not overflow a
2099        32-bit integer. However, it turns out that people want to use groups that
2100        are longer than 30,000 bytes (though not repeat them that many times).
2101        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2102        made it possible to implement the integer overflow checks in a much more
2103        dynamic way, which I have now done. The artificial limitation on group
2104        length has been removed - we now have only the limit on the total length of
2105        the compiled pattern, which depends on the LINK_SIZE setting.
2106    
2107    10. Fixed a bug in the documentation for get/copy named substring when
2108        duplicate names are permitted. If none of the named substrings are set, the
2109        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2110        empty string.
2111    
2112    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2113        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2114        because the ] is interpreted as the first data character and the
2115        terminating ] is not found. PCRE has been made compatible with Perl in this
2116        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2117        cause memory overwriting.
2118    
2119    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2120        string has been matched (to stop an infinite loop). It was not recognizing
2121        a conditional subpattern that could match an empty string if that
2122        subpattern was within another subpattern. For example, it looped when
2123        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2124        condition was not nested. This bug has been fixed.
2125    
2126    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2127        past the start of the subject in the presence of bytes with the top bit
2128        set, for example "\x8aBCD".
2129    
2130    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2131        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2132    
2133    14. Optimized (?!) to (*FAIL).
2134    
2135    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2136        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2137        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2138        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2139        does: it's just the validity check that is more restrictive.
2140    
2141    16. Inserted checks for integer overflows during escape sequence (backslash)
2142        processing, and also fixed erroneous offset values for syntax errors during
2143        backslash processing.
2144    
2145    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2146        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2147    
2148    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2149        caused an overrun.
2150    
2151    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2152        something other than just ASCII characters) inside a group that had an
2153        unlimited repeat caused a loop at compile time (while checking to see
2154        whether the group could match an empty string).
2155    
2156    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2157        [\P{Any}] did so. (Error in the code for printing property names.)
2158    
2159    21. An orphan \E inside a character class could cause a crash.
2160    
2161    22. A repeated capturing bracket such as (A)? could cause a wild memory
2162        reference during compilation.
2163    
2164    23. There are several functions in pcre_compile() that scan along a compiled
2165        expression for various reasons (e.g. to see if it's fixed length for look
2166        behind). There were bugs in these functions when a repeated \p or \P was
2167        present in the pattern. These operators have additional parameters compared
2168        with \d, etc, and these were not being taken into account when moving along
2169        the compiled data. Specifically:
2170    
2171        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2172            length.
2173    
2174        (b) An item such as \pL+ within a repeated group could cause crashes or
2175            loops.
2176    
2177        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2178            "reference to non-existent subpattern" error.
2179    
2180        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2181    
2182    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2183        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2184    
2185    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2186    
2187    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2188        character were causing crashes (broken optimization).
2189    
2190    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2191        \p or \P) caused a compile-time loop.
2192    
2193    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2194        break. For example, the unstudied pattern [\r\n]A does not match the string
2195        "\r\nA" because change 7.0/46 below moves the current point on by two
2196        characters after failing to match at the start. However, the pattern \nA
2197        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2198        the same is true. There doesn't seem any very clean way out of this, but
2199        what I have chosen to do makes the common cases work: PCRE now takes note
2200        of whether there can be an explicit match for \r or \n anywhere in the
2201        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2202        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2203        pattern has explicit CR or LF references.
2204    
2205    29. Added (*CR) etc for changing newline setting at start of pattern.
2206    
2207    
2208    Version 7.2 19-Jun-07
2209    ---------------------
2210    
2211     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2212        which is apparently normally available under Windows.
2213    
2214     2. Re-jig the pcregrep tests with different newline settings in an attempt
2215        to make them independent of the local environment's newline setting.
2216    
2217     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2218    
2219     4. Some of the "internals" tests were previously cut out when the link size
2220        was not 2, because the output contained actual offsets. The recent new
2221        "Z" feature of pcretest means that these can be cut out, making the tests
2222        usable with all link sizes.
2223    
2224     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2225        stack recursion. This gives a massive performance boost under BSD, but just
2226        a small improvement under Linux. However, it saves one field in the frame
2227        in all cases.
2228    
2229     6. Added more features from the forthcoming Perl 5.10:
2230    
2231        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2232            recursion call. It refers to the nth most recently opened parentheses.
2233    
2234        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2235            to be opened parentheses.
2236    
2237        (c) Conditions that refer to capturing parentheses can be specified
2238            relatively, for example, (?(-2)... or (?(+3)...
2239    
2240        (d) \K resets the start of the current match so that everything before
2241            is not part of it.
2242    
2243        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2244    
2245        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2246            reference syntax.
2247    
2248        (g) (?| introduces a group in which the numbering of parentheses in each
2249            alternative starts with the same number.
2250    
2251        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2252    
2253     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2254        PCRE_INFO_JCHANGED.
2255    
2256     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2257        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2258        for detecting groups that can match an empty string.
2259    
2260     9. A pattern with a very large number of alternatives (more than several
2261        hundred) was running out of internal workspace during the pre-compile
2262        phase, where pcre_compile() figures out how much memory will be needed. A
2263        bit of new cunning has reduced the workspace needed for groups with
2264        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2265        workspace instead of running out of the 4096 that are available.
2266    
2267    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2268    
2269    11. Applied patch from Google to remove an optimization that didn't quite work.
2270        The report of the bug said:
2271    
2272          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2273          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2274          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2275    
2276    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2277        it matched the wrong number of bytes.
2278    
2279    
2280    Version 7.1 24-Apr-07
2281    ---------------------
2282    
2283     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2284        that is more "standard", making use of automake and other Autotools. There
2285        is some re-arrangement of the files and adjustment of comments consequent
2286        on this.
2287    
2288     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2289        for recursive directory scanning broke on some systems because the files
2290        are not scanned in any specific order and on different systems the order
2291        was different. A call to "sort" has been inserted into RunGrepTest for the
2292        approprate test as a short-term fix. In the longer term there may be an
2293        alternative.
2294    
2295     3. I had an email from Eric Raymond about problems translating some of PCRE's
2296        man pages to HTML (despite the fact that I distribute HTML pages, some
2297        people do their own conversions for various reasons). The problems
2298        concerned the use of low-level troff macros .br and .in. I have therefore
2299        removed all such uses from the man pages (some were redundant, some could
2300        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2301        HTML has been updated to handle .nf/.fi and to complain if it encounters
2302        .br or .in.
2303    
2304     4. Updated comments in configure.ac that get placed in config.h.in and also
2305        arranged for config.h to be included in the distribution, with the name
2306        config.h.generic, for the benefit of those who have to compile without
2307        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2308    
2309     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2310        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2311        makevp.bat for the current PCRE, using the additional files
2312        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2313    
2314     6. A Windows user reported a minor discrepancy with test 2, which turned out
2315        to be caused by a trailing space on an input line that had got lost in his
2316        copy. The trailing space was an accident, so I've just removed it.
2317    
2318     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2319        that is needed.
2320    
2321     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2322        as "const" (a) because they are and (b) because it helps the PHP
2323        maintainers who have recently made a script to detect big data structures
2324        in the php code that should be moved to the .rodata section. I remembered
2325        to update Builducptable as well, so it won't revert if ucptable.h is ever
2326        re-created.
2327    
2328     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2329        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2330        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2331        support is not required. This saves 1.5-2K of code, which is important in
2332        some applications.
2333    
2334        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2335        so as not to refer to the tables, even though these functions will never be
2336        called when UTF-8 support is disabled. Otherwise there are problems with a
2337        shared library.
2338    
2339    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2340    
2341        (a) It was defining its arguments as char * instead of void *.
2342    
2343        (b) It was assuming that all moves were upwards in memory; this was true
2344            a long time ago when I wrote it, but is no longer the case.
2345    
2346        The emulated memove() is provided for those environments that have neither
2347        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2348        is clearly not the case, as these two bugs were recently reported.
2349    
2350    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2351        and Detrail to create the HTML documentation, the .txt form of the man
2352        pages, and it removes trailing spaces from listed files. It also creates
2353        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2354        case, it wraps all the #defines with #ifndefs. This script should be run
2355        before "make dist".
2356    
2357    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2358        with Unicode property support.
2359    
2360        (a) For a maximizing quantifier, if the two different cases of the
2361            character were of different lengths in their UTF-8 codings (there are
2362            some cases like this - I found 11), and the matching function had to
2363            back up over a mixture of the two cases, it incorrectly assumed they
2364            were both the same length.
2365    
2366        (b) When PCRE was configured to use the heap rather than the stack for
2367            recursion during matching, it was not correctly preserving the data for
2368            the other case of a UTF-8 character when checking ahead for a match
2369            while processing a minimizing repeat. If the check also involved
2370            matching a wide character, but failed, corruption could cause an
2371            erroneous result when trying to check for a repeat of the original
2372            character.
2373    
2374    13. Some tidying changes to the testing mechanism:
2375    
2376        (a) The RunTest script now detects the internal link size and whether there
2377            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2378            values substituted by "configure". (The RunGrepTest script already did
2379            this for UTF-8.) The configure.ac script no longer substitutes the
2380            relevant variables.
2381    
2382        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2383            with length and offset values. This means that the output is different
2384            for different internal link sizes. Test 2 is skipped for link sizes
2385            other than 2 because of this, bypassing the problem. Unfortunately,
2386            there was also a test in test 3 (the locale tests) that used /B and
2387            failed for link sizes other than 2. Rather than cut the whole test out,
2388            I have added a new /Z option to pcretest that replaces the length and
2389            offset values with spaces. This is now used to make test 3 independent
2390            of link size. (Test 2 will be tidied up later.)
2391    
2392    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2393        segmentation fault instead of returning the appropriate error message.
2394    
2395    15. In multiline mode when the newline sequence was set to "any", the pattern
2396        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2397        This doesn't seem right; it now treats the CRLF combination as the line
2398        ending, and so does not match in that case. It's only a pattern such as ^$
2399        that would hit this one: something like ^ABC$ would have failed after \r
2400        and then tried again after \r\n.
2401    
2402    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2403        in an attempt to make files that differ only in their line terminators
2404        compare equal. This works on Linux.
2405    
2406    17. Under certain error circumstances pcregrep might try to free random memory
2407        as it exited. This is now fixed, thanks to valgrind.
2408    
2409    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2410        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2411        was because its rules for how to advance for /g after matching an empty
2412        string at the end of a line did not allow for this case. They now check for
2413        it specially.
2414    
2415    20. pcretest is supposed to handle patterns and data of any length, by
2416        extending its buffers when necessary. It was getting this wrong when the
2417        buffer for a data line had to be extended.
2418    
2419    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2420        CRLF as a newline sequence.
2421    
2422    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2423        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2424        I have nevertheless tidied it up.
2425    
2426    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2427    
2428    24. Added a man page for pcre-config.
2429    
2430    
2431    Version 7.0 19-Dec-06
2432    ---------------------
2433    
2434     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2435        moving to gcc 4.1.1.
2436    
2437     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2438        sys/time.h, which is documented as needed for this function. It doesn't
2439        seem to matter on Linux, but it showed up on some releases of OS X.
2440    
2441     3. It seems that there are systems where bytes whose values are greater than
2442        127 match isprint() in the "C" locale. The "C" locale should be the
2443        default when a C program starts up. In most systems, only ASCII printing
2444        characters match isprint(). This difference caused the output from pcretest
2445        to vary, making some of the tests fail. I have changed pcretest so that:
2446    
2447        (a) When it is outputting text in the compiled version of a pattern, bytes
2448            other than 32-126 are always shown as hex escapes.
2449    
2450        (b) When it is outputting text that is a matched part of a subject string,
2451            it does the same, unless a different locale has been set for the match
2452            (using the /L modifier). In this case, it uses isprint() to decide.
2453    
2454     4. Fixed a major bug that caused incorrect computation of the amount of memory
2455        required for a compiled pattern when options that changed within the
2456        pattern affected the logic of the preliminary scan that determines the
2457        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2458        that the computed length was too small. The symptoms of this bug were
2459        either the PCRE error "internal error: code overflow" from pcre_compile(),
2460        or a glibc crash with a message such as "pcretest: free(): invalid next
2461        size (fast)". Examples of patterns that provoked this bug (shown in
2462        pcretest format) are:
2463    
2464          /(?-x: )/x
2465          /(?x)(?-x: \s*#\s*)/
2466          /((?i)[\x{c0}])/8
2467          /(?i:[\x{c0}])/8
2468    
2469        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2470        is now done differently.
2471    
2472     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2473        wrapper classes; (b) implement a new function in the C++ scanner that is
2474        more efficient than the old way of doing things because it avoids levels of
2475        recursion in the regex matching; (c) add a paragraph to the documentation
2476        for the FullMatch() function.
2477    
2478     6. The escape sequence \n was being treated as whatever was defined as
2479        "newline". Not only was this contrary to the documentation, which states
2480        that \n is character 10 (hex 0A), but it also went horribly wrong when
2481        "newline" was defined as CRLF. This has been fixed.
2482    
2483     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2484        was being set to -1 for the "end of line" case (supposedly a value that no
2485        character can have). Though this value is never used (the check for end of
2486        line is "zero bytes in current character"), it caused compiler complaints.
2487        I've changed it to 0xffffffff.
2488    
2489     8. In pcre_version.c, the version string was being built by a sequence of
2490        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2491        string (as it is for production releases) called a macro with an empty
2492        argument. The C standard says the result of this is undefined. The gcc
2493        compiler treats it as an empty string (which was what was wanted) but it is
2494        reported that Visual C gives an error. The source has been hacked around to
2495        avoid this problem.
2496    
2497     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2498        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2499        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2500        of them did).
2501    
2502    10. Originally, pcretest opened its input and output without "b"; then I was
2503        told that "b" was needed in some environments, so it was added for release
2504        5.0 to both the input and output. (It makes no difference on Unix-like
2505        systems.) Later I was told that it is wrong for the input on Windows. I've
2506        now abstracted the modes into two macros, to make it easier to fiddle with
2507        them, and removed "b" from the input mode under Windows.
2508    
2509    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2510    
2511    12. Added -help and --help to pcretest as an official way of being reminded
2512        of the options.
2513    
2514    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2515        and pcrecpp.cc because they annoy compilers at high warning levels.
2516    
2517    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2518    
2519    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2520        scripts are not C programs :-) and which was not noticed because it works
2521        on Linux.
2522    
2523    16. pcretest is supposed to handle any length of pattern and data line (as one
2524        line or as a continued sequence of lines) by extending its input buffer if
2525        necessary. This feature was broken for very long pattern lines, leading to
2526        a string of junk being passed to pcre_compile() if the pattern was longer
2527        than about 50K.
2528    
2529    17. I have done a major re-factoring of the way pcre_compile() computes the
2530        amount of memory needed for a compiled pattern. Previously, there was code
2531        that made a preliminary scan of the pattern in order to do this. That was
2532        OK when PCRE was new, but as the facilities have expanded, it has become
2533        harder and harder to keep it in step with the real compile phase, and there
2534        have been a number of bugs (see for example, 4 above). I have now found a
2535        cunning way of running the real compile function in a "fake" mode that
2536        enables it to compute how much memory it would need, while actually only
2537        ever using a few hundred bytes of working memory and without too many
2538        tests of the mode. This should make future maintenance and development
2539        easier. A side effect of this work is that the limit of 200 on the nesting
2540        depth of parentheses has been removed (though this was never a serious
2541        limitation, I suspect). However, there is a downside: pcre_compile() now
2542        runs more slowly than before (30% or more, depending on the pattern). I
2543        hope this isn't a big issue. There is no effect on runtime performance.
2544    
2545    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2546        newline (only possible for the last line of a file) and it was a
2547        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2548    
2549    19. Added additional timing features to pcretest. (1) The -tm option now times
2550        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2551        separate command line item, by a number that specifies the number of
2552        repeats to use when timing. The default is 50000; this gives better
2553        precision, but takes uncomfortably long for very large patterns.
2554    
2555    20. Extended pcre_study() to be more clever in cases where a branch of a
2556        subpattern has no definite first character. For example, (a*|b*)[cd] would
2557        previously give no result from pcre_study(). Now it recognizes that the
2558        first character must be a, b, c, or d.
2559    
2560    21. There was an incorrect error "recursive call could loop indefinitely" if
2561        a subpattern (or the entire pattern) that was being tested for matching an
2562        empty string contained only one non-empty item after a nested subpattern.
2563        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2564        incorrectly, because the \d was being skipped in the check.
2565    
2566    22. The pcretest program now has a new pattern option /B and a command line
2567        option -b, which is equivalent to adding /B to every pattern. This causes
2568        it to show the compiled bytecode, without the additional information that
2569        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2570        is the same as /B/I).
2571    
2572    23. A new optimization is now able automatically to treat some sequences such
2573        as a*b as a*+b. More specifically, if something simple (such as a character
2574        or a simple class like \d) has an unlimited quantifier, and is followed by
2575        something that cannot possibly match the quantified thing, the quantifier
2576        is automatically "possessified".
2577    
2578    24. A recursive reference to a subpattern whose number was greater than 39
2579        went wrong under certain circumstances in UTF-8 mode. This bug could also
2580        have affected the operation of pcre_study().
2581    
2582    25. Realized that a little bit of performance could be had by replacing
2583        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2584    
2585    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2586    
2587    27. Possessive quantifiers such as a++ were previously implemented by turning
2588        them into atomic groups such as ($>a+). Now they have their own opcodes,
2589        which improves performance. This includes the automatically created ones
2590        from 23 above.
2591    
2592    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2593        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2594        the first matched character to be a colon. This applied both to named and
2595        numbered groups.
2596    
2597    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2598    
2599    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2600        building PCRE on Windows easier, so I have included it in the distribution.
2601    
2602    31. There is now a check in pcretest against a ridiculously large number being
2603        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2604        loop, the loop is abandoned.
2605    
2606    32. Forward references to subpatterns in conditions such as (?(2)...) where
2607        subpattern 2 is defined later cause pcre_compile() to search forwards in
2608        the pattern for the relevant set of parentheses. This search went wrong
2609        when there were unescaped parentheses in a character class, parentheses
2610        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2611    
2612    33. "Subroutine" calls and backreferences were previously restricted to
2613        referencing subpatterns earlier in the regex. This restriction has now
2614        been removed.
2615    
2616    34. Added a number of extra features that are going to be in Perl 5.10. On the
2617        whole, these are just syntactic alternatives for features that PCRE had
2618        previously implemented using the Python syntax or my own invention. The
2619        other formats are all retained for compatibility.
2620    
2621        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2622            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2623            also .NET compatible.
2624    
2625        (b) A recursion or subroutine call to a named group can now be defined as
2626            (?&name) as well as (?P>name).
2627    
2628        (c) A backreference to a named group can now be defined as \k<name> or
2629            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2630            5.10, are also .NET compatible.
2631    
2632        (d) A conditional reference to a named group can now use the syntax
2633            (?(<name>) or (?('name') as well as (?(name).
2634    
2635        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2636            groups (named and numbered) that are never evaluated inline, but can be
2637            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2638            is always false. There may be only one alternative in such a group.
2639    
2640        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2641            as the simple (?(R). The condition is true only if the most recent
2642            recursion is that of the given number or name. It does not search out
2643            through the entire recursion stack.
2644    
2645        (g) The escape \gN or \g{N} has been added, where N is a positive or
2646            negative number, specifying an absolute or relative reference.
2647    
2648    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2649        some "unreachable code" warnings.
2650    
2651    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2652        things, this adds five new scripts.
2653    
2654    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2655        There were also incompatibilities regarding the handling of \Q..\E inside
2656        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2657        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2658    
2659    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2660        matches an empty string, and forcibly breaks the loop. There were bugs in
2661        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2662        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2663        separate and independent bugs (that affected different cases) have been
2664        fixed.
2665    
2666    39. Refactored the code to abolish the use of different opcodes for small
2667        capturing bracket numbers. This is a tidy that I avoided doing when I
2668        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2669        The new approach is not only tidier, it makes it possible to reduce the
2670        memory needed to fix the previous bug (38).
2671    
2672    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2673        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2674        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2675        mode.
2676    
2677    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2678        report.
2679    
2680    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2681        copy construction and assignment in the C++ wrapper.
2682    
2683    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2684        couple of bugs that could have given wrong results in the "--newline=crlf"
2685        case.
2686    
2687    44. Added a number of casts and did some reorganization of signed/unsigned int
2688        variables following suggestions from Dair Grant. Also renamed the variable
2689        "this" as "item" because it is a C++ keyword.
2690    
2691    45. Arranged for dftables to add
2692    
2693          #include "pcre_internal.h"
2694    
2695        to pcre_chartables.c because without it, gcc 4.x may remove the array
2696        definition from the final binary if PCRE is built into a static library and
2697        dead code stripping is activated.
2698    
2699    46. For an unanchored pattern, if a match attempt fails at the start of a
2700        newline sequence, and the newline setting is CRLF or ANY, and the next two
2701        characters are CRLF, advance by two characters instead of one.
2702    
2703    
2704    Version 6.7 04-Jul-06
2705    ---------------------
2706    
2707     1. In order to handle tests when input lines are enormously long, pcretest has
2708        been re-factored so that it automatically extends its buffers when
2709        necessary. The code is crude, but this _is_ just a test program. The
2710        default size has been increased from 32K to 50K.
2711    
2712     2. The code in pcre_study() was using the value of the re argument before
2713        testing it for NULL. (Of course, in any sensible call of the function, it
2714        won't be NULL.)
2715    
2716     3. The memmove() emulation function in pcre_internal.h, which is used on
2717        systems that lack both memmove() and bcopy() - that is, hardly ever -
2718        was missing a "static" storage class specifier.
2719    
2720     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2721        containing an extended class (one that cannot be represented by a bitmap
2722        because it contains high-valued characters or Unicode property items, e.g.
2723        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2724        pattern, but PCRE should not loop if you do not (it no longer does).
2725        [Detail: two cases were found: (a) a repeated subpattern containing an
2726        extended class; (b) a recursive reference to a subpattern that followed a
2727        previous extended class. It wasn't skipping over the extended class
2728        correctly when UTF-8 mode was not set.]
2729    
2730     5. A negated single-character class was not being recognized as fixed-length
2731        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2732        compile error "lookbehind assertion is not fixed length".
2733    
2734     6. The RunPerlTest auxiliary script was showing an unexpected difference
2735        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2736        write a Perl script that can interpret lines of an input file either as
2737        byte characters or as UTF-8, which is what "perltest" was being required to
2738        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2739        can't do is switch easily at run time between having the "use utf8;" pragma
2740        or not. In the end, I fudged it by using the RunPerlTest script to insert
2741        "use utf8;" explicitly for the UTF-8 tests.
2742    
2743     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2744        the end of the subject string, contrary to the documentation and to what
2745        Perl does. This was true of both matching functions. Now it matches only at
2746        the start of the subject and immediately after *internal* newlines.
2747    
2748     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2749        a pointer to an int instead of a pointer to an unsigned long int. This
2750        caused problems on 64-bit systems.
2751    
2752     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2753        instance of the 'standard' template library not being so standard".
2754    
2755    10. There was no check on the number of named subpatterns nor the maximum
2756        length of a subpattern name. The product of these values is used to compute
2757        the size of the memory block for a compiled pattern. By supplying a very
2758        long subpattern name and a large number of named subpatterns, the size
2759        computation could be caused to overflow. This is now prevented by limiting
2760        the length of names to 32 characters, and the number of named subpatterns
2761        to 10,000.
2762    
2763    11. Subpatterns that are repeated with specific counts have to be replicated in
2764        the compiled pattern. The size of memory for this was computed from the
2765        length of the subpattern and the repeat count. The latter is limited to
2766        65535, but there was no limit on the former, meaning that integer overflow
2767        could in principle occur. The compiled length of a repeated subpattern is
2768        now limited to 30,000 bytes in order to prevent this.
2769    
2770    12. Added the optional facility to have named substrings with the same name.
2771    
2772    13. Added the ability to use a named substring as a condition, using the
2773        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2774        are numbers (not recommended). Forward references are permitted.
2775    
2776    14. Added forward references in named backreferences (if you see what I mean).
2777    
2778    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2779        pattern could run off the end of the subject. For example, the pattern
2780        "(?s)(.{1,5})"8 did this with the subject "ab".
2781    
2782    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2783        PCRE_CASELESS was set when matching characters that were quantified with ?
2784        or *.
2785    
2786    17. A character class other than a single negated character that had a minimum
2787        but no maximum quantifier - for example [ab]{6,} - was not handled
2788        correctly by pce_dfa_exec(). It would match only one character.
2789    
2790    18. A valid (though odd) pattern that looked like a POSIX character
2791        class but used an invalid character after [ (for example [[,abc,]]) caused
2792        pcre_compile() to give the error "Failed: internal error: code overflow" or
2793        in some cases to crash with a glibc free() error. This could even happen if
2794        the pattern terminated after [[ but there just happened to be a sequence of
2795        letters, a binary zero, and a closing ] in the memory that followed.
2796    
2797    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2798        over the years. Originally (before any Unicode support), just the bottom 8
2799        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2800        output from "man perlunicode" includes this:
2801    
2802          The regular expression compiler produces polymorphic opcodes.  That
2803          is, the pattern adapts to the data and automatically switches to
2804          the Unicode character scheme when presented with Unicode data--or
2805          instead uses a traditional byte scheme when presented with byte
2806          data.
2807    
2808        Sadly, a wide octal escape does not cause a switch, and in a string with
2809        no other multibyte characters, these octal escapes are treated as before.
2810        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2811        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2812        Unicode string.
2813    
2814        I have not perpetrated such confusion in PCRE. Up till now, it took just
2815        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2816        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2817        translate to the appropriate multibyte character.
2818    
2819    29. Applied some refactoring to reduce the number of warnings from Microsoft
2820        and Borland compilers. This has included removing the fudge introduced
2821        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2822        a warning about an unused variable.
2823    
2824    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2825        characters since release 4.0, because Perl (from release 5.004) does not.
2826        [Or at least, is documented not to: some releases seem to be in conflict
2827        with the documentation.] However, when a pattern was studied with
2828        pcre_study() and all its branches started with \s, PCRE still included VT
2829        as a possible starting character. Of course, this did no harm; it just
2830        caused an unnecessary match attempt.
2831    
2832    22. Removed a now-redundant internal flag bit that recorded the fact that case
2833        dependency changed within the pattern. This was once needed for "required
2834        byte" processing, but is no longer used. This recovers a now-scarce options
2835        bit. Also moved the least significant internal flag bit to the most-
2836        significant bit of the word, which was not previously used (hangover from
2837        the days when it was an int rather than a uint) to free up another bit for
2838        the future.
2839    
2840    23. Added support for CRLF line endings as well as CR and LF. As well as the
2841        default being selectable at build time, it can now be changed at runtime
2842        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2843        specify that it is scanning data with non-default line endings.
2844    
2845    24. Changed the definition of CXXLINK to make it agree with the definition of
2846        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2847    
2848    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2849        recursions. This makes a big different to stack usage for some patterns.
2850    
2851    26. If a subpattern containing a named recursion or subroutine reference such
2852        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2853        the space required for the compiled pattern went wrong and gave too small a
2854        value. Depending on the environment, this could lead to "Failed: internal
2855        error: code overflow at offset 49" or "glibc detected double free or
2856        corruption" errors.
2857    
2858    27. Applied patches from Google (a) to support the new newline modes and (b) to
2859        advance over multibyte UTF-8 characters in GlobalReplace.
2860    
2861    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2862        difference for some implementation of PCRE in some Windows version.
2863    
2864    29. Added some extra testing facilities to pcretest:
2865    
2866        \q<number>   in a data line sets the "match limit" value
2867        \Q<number>   in a data line sets the "match recursion limt" value
2868        -S <number>  sets the stack size, where <number> is in megabytes
2869    
2870        The -S option isn't available for Windows.
2871    
2872    
2873    Version 6.6 06-Feb-06
2874    ---------------------
2875    
2876     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2877        in pcreposix.h. I have copied the definition from pcre.h.
2878    
2879     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2880        because pcre.h is no longer a built file.
2881    
2882     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2883        not normally included in the compiled code.
2884    
2885    
2886    Version 6.5 01-Feb-06
2887    ---------------------
2888    
2889     1. When using the partial match feature with pcre_dfa_exec(), it was not
2890        anchoring the second and subsequent partial matches at the new starting
2891        point. This could lead to incorrect results. For example, with the pattern
2892        /1234/, partially matching against "123" and then "a4" gave a match.
2893    
2894     2. Changes to pcregrep:
2895    
2896        (a) All non-match returns from pcre_exec() were being treated as failures
2897            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2898            error message is output. Some extra information is given for the
2899            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2900            probably the only errors that are likely to be caused by users (by
2901            specifying a regex that has nested indefinite repeats, for instance).
2902            If there are more than 20 of these errors, pcregrep is abandoned.
2903    
2904        (b) A binary zero was treated as data while matching, but terminated the
2905            output line if it was written out. This has been fixed: binary zeroes
2906            are now no different to any other data bytes.
2907    
2908        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2909            used to set a locale for matching. The --locale=xxxx long option has
2910            been added (no short equivalent) to specify a locale explicitly on the
2911            pcregrep command, overriding the environment variables.
2912    
2913        (d) When -B was used with -n, some line numbers in the output were one less
2914            than they should have been.
2915    
2916        (e) Added the -o (--only-matching) option.
2917    
2918        (f) If -A or -C was used with -c (count only), some lines of context were
2919            accidentally printed for the final match.
2920    
2921        (g) Added the -H (--with-filename) option.
2922    
2923        (h) The combination of options -rh failed to suppress file names for files
2924            that were found from directory arguments.
2925    
2926        (i) Added the -D (--devices) and -d (--directories) options.
2927    
2928        (j) Added the -F (--fixed-strings) option.
2929    
2930        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2931    
2932        (l) Added the --colo(u)r option.
2933    
2934        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2935            is not present by default.
2936    
2937     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2938        items such as (?R) or (?1), when the recursion could match a number of
2939        alternatives. If it matched one of the alternatives, but subsequently,
2940        outside the recursion, there was a failure, the code tried to back up into
2941        the recursion. However, because of the way PCRE is implemented, this is not
2942        possible, and the result was an incorrect result from the match.
2943    
2944        In order to prevent this happening, the specification of recursion has
2945        been changed so that all such subpatterns are automatically treated as
2946        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2947    
2948     4. I had overlooked the fact that, in some locales, there are characters for
2949        which isalpha() is true but neither isupper() nor islower() are true. In
2950        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2951        and ordfeminine) are like this. This affected the treatment of \w and \W
2952        when they appeared in character classes, but not when they appeared outside
2953        a character class. The bit map for "word" characters is now created
2954        separately from the results of isalnum() instead of just taking it from the
2955        upper, lower, and digit maps. (Plus the underscore character, of course.)
2956    
2957     5. The above bug also affected the handling of POSIX character classes such as
2958        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2959        permanent tables. Instead, the bit maps for such a class were previously
2960        created as the appropriate unions of the upper, lower, and digit bitmaps.
2961        Now they are created by subtraction from the [[:word:]] class, which has
2962        its own bitmap.
2963    
2964     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2965        It is created by subtracting the vertical space characters (\x09, \x0a,
2966        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2967        subtraction was done in the overall bitmap for a character class, meaning
2968        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2969        be recognized. This bug has been fixed.
2970    
2971     7. Patches from the folks at Google:
2972    
2973          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2974          real life, but is still worth protecting against".
2975    
2976          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2977          regular expressions".
2978    
2979          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2980          have it.
2981    
2982          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2983          "configure" and the latter not, in order to fix a problem somebody had
2984          with compiling the Arg class on HP-UX.
2985    
2986          (e) Improve the error-handling of the C++ wrapper a little bit.
2987    
2988          (f) New tests for checking recursion limiting.
2989    
2990     8. The pcre_memmove() function, which is used only if the environment does not
2991        have a standard memmove() function (and is therefore rarely compiled),
2992        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2993        returning a result (though PCRE never actually uses the result).
2994    
2995     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2996        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2997        returned instead of calling malloc() with an overflowing number that would
2998        most likely cause subsequent chaos.
2999    
3000    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3001    
3002    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3003        with this option is matched, the nmatch and pmatch options of regexec() are
3004        ignored.
3005    
3006    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3007        provided in case anyone wants to the the POSIX interface with UTF-8
3008        strings.
3009    
3010    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3011        C++ linking (needed for some HP-UX environments).
3012    
3013    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3014        (unused parameter) and in the pcre_printint() function (omitted "default"
3015        switch label when the default is to do nothing).
3016    
3017    15. Added some code to make it possible, when PCRE is compiled as a C++
3018        library, to replace subject pointers for pcre_exec() with a smart pointer
3019        class, thus making it possible to process discontinuous strings.
3020    
3021    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3022        much the same function. They were added by different people who were trying
3023        to make PCRE easy to compile on non-Unix systems. It has been suggested
3024        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3025        for compiling on Windows systems. I have therefore replaced it with
3026        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3027        defaults to "extern" for C or "extern C" for C++, which works fine on
3028        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3029        SCOPE with something explicit in config.h. In addition:
3030    
3031        (a) pcreposix.h still had just "extern" instead of either of these macros;
3032            I have replaced it with PCRE_DATA_SCOPE.
3033    
3034        (b) Functions such as _pcre_xclass(), which are internal to the library,
3035            but external in the C sense, all had PCRE_EXPORT in their definitions.
3036            This is apparently wrong for the Windows case, so I have removed it.
3037            (It makes no difference on Unix-like systems.)
3038    
3039    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3040        of recursive calls to match(). This is different to MATCH_LIMIT because
3041        that limits the total number of calls to match(), not all of which increase
3042        the depth of recursion. Limiting the recursion depth limits the amount of
3043        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3044        when PCRE is compiled, and changed at run time. A patch from Google adds
3045        this functionality to the C++ interface.
3046    
3047    18. Changes to the handling of Unicode character properties:
3048    
3049        (a) Updated the table to Unicode 4.1.0.
3050    
3051        (b) Recognize characters that are not in the table as "Cn" (undefined).
3052    
3053        (c) I revised the way the table is implemented to a much improved format
3054            which includes recognition of ranges. It now supports the ranges that
3055            are defined in UnicodeData.txt, and it also amalgamates other
3056            characters into ranges. This has reduced the number of entries in the
3057            table from around 16,000 to around 3,000, thus reducing its size
3058            considerably. I realized I did not need to use a tree structure after
3059            all - a binary chop search is just as efficient. Having reduced the
3060            number of entries, I extended their size from 6 bytes to 8 bytes to
3061            allow for more data.
3062    
3063        (d) Added support for Unicode script names via properties such as \p{Han}.
3064    
3065    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3066        matching that character.
3067    
3068    20. When matching a repeated Unicode property with a minimum greater than zero,
3069        (for example \pL{2,}), PCRE could look past the end of the subject if it
3070        reached it while seeking the minimum number of characters. This could
3071        happen only if some of the characters were more than one byte long, because
3072        there is a check for at least the minimum number of bytes.
3073    
3074    21. Refactored the implementation of \p and \P so as to be more general, to
3075        allow for more different types of property in future. This has changed the
3076        compiled form incompatibly. Anybody with saved compiled patterns that use
3077        \p or \P will have to recompile them.
3078    
3079    22. Added "Any" and "L&" to the supported property types.
3080    
3081    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3082        but give a compile time error if the value is greater than 0xff.
3083    
3084    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3085        accidentally not being installed or uninstalled.
3086    
3087    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3088        made were to insert the current release number. This seemed silly, because
3089        it made things harder for people building PCRE on systems that don't run
3090        "configure". I have turned pcre.h into a distributed file, no longer built
3091        by "configure", with the version identification directly included. There is
3092        no longer a pcre.h.in file.
3093    
3094        However, this change necessitated a change to the pcre-config script as
3095        well. It is built from pcre-config.in, and one of the substitutions was the
3096        release number. I have updated configure.ac so that ./configure now finds
3097        the release number by grepping pcre.h.
3098    
3099    26. Added the ability to run the tests under valgrind.
3100    
3101    
3102    Version 6.4 05-Sep-05
3103    ---------------------
3104    
3105     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3106        "--" to be printed when multiple files were scanned, even when none of the
3107        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3108        consider it to be a bug, and have restored the previous behaviour.
3109    
3110     2. A couple of code tidies to get rid of compiler warnings.
3111    
3112     3. The pcretest program used to cheat by referring to symbols in the library
3113        whose names begin with _pcre_. These are internal symbols that are not
3114        really supposed to be visible externally, and in some environments it is
3115        possible to suppress them. The cheating is now confined to including
3116        certain files from the library's source, which is a bit cleaner.
3117    
3118     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3119        file's purpose clearer.
3120    
3121     5. Reorganized pcre_ucp_findchar().
3122    
3123    
3124    Version 6.3 15-Aug-05
3125    ---------------------
3126    
3127     1. The file libpcre.pc.in did not have general read permission in the tarball.
3128    
3129     2. There were some problems when building without C++ support:
3130    
3131        (a) If C++ support was not built, "make install" and "make test" still
3132            tried to test it.
3133    
3134        (b) There were problems when the value of CXX was explicitly set. Some
3135            changes have been made to try to fix these, and ...
3136    
3137        (c) --disable-cpp can now be used to explicitly disable C++ support.
3138    
3139        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3140            backslash in a target when C++ was disabled. This confuses some
3141            versions of "make", apparently. Using an intermediate variable solves
3142            this. (Same for CPP_LOBJ.)
3143    
3144     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3145        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3146        necessary on certain architectures.
3147    
3148     4. Added a setting of -export-symbols-regex to the link command to remove
3149        those symbols that are exported in the C sense, but actually are local
3150        within the library, and not documented. Their names all begin with
3151        "_pcre_". This is not a perfect job, because (a) we have to except some
3152        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3153        available (and never for static libraries). I have made a note to try to
3154        find a way round (a) in the future.
3155    
3156    
3157    Version 6.2 01-Aug-05
3158    ---------------------
3159    
3160     1. There was no test for integer overflow of quantifier values. A construction
3161        such as {1111111111111111} would give undefined results. What is worse, if
3162        a minimum quantifier for a parenthesized subpattern overflowed and became
3163        negative, the calculation of the memory size went wrong. This could have
3164        led to memory overwriting.
3165    
3166     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3167    
3168     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3169        operating environments where this matters.
3170    
3171     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3172        PCRE options from within the C++ wrapper.
3173    
3174     5. Named capturing subpatterns were not being correctly counted when a pattern
3175        was compiled. This caused two problems: (a) If there were more than 100
3176        such subpatterns, the calculation of the memory needed for the whole
3177        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3178        back references of the form \12, where the number was greater than 9, were
3179        not recognized as back references, even though there were sufficient
3180        previous subpatterns.
3181    
3182     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3183        versions of gcc, e.g. 2.95.4.
3184    
3185    
3186    Version 6.1 21-Jun-05
3187    ---------------------
3188    
3189     1. There was one reference to the variable "posix" in pcretest.c that was not
3190        surrounded by "#if !defined NOPOSIX".
3191    
3192     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3193        the cross-check on the old pcre_info() function, for the benefit of the
3194        cut-down version of PCRE that is currently imported into Exim.
3195    
3196     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3197        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3198        patterns that start (?i) but I don't think that matters. The use of (?i) is
3199        just an example; this all applies to the other options as well.
3200    
3201     4. Since libtool seems to echo the compile commands it is issuing, the output
3202        from "make" can be reduced a bit by putting "@" in front of each libtool
3203        compile command.
3204    
3205     5. Patch from the folks at Google for configure.in to be a bit more thorough
3206        in checking for a suitable C++ installation before trying to compile the
3207        C++ stuff. This should fix a reported problem when a compiler was present,
3208        but no suitable headers.
3209    
3210     6. The man pages all had just "PCRE" as their title. I have changed them to
3211        be the relevant file name. I have also arranged that these names are
3212        retained in the file doc/pcre.txt, which is a concatenation in text format
3213        of all the man pages except the little individual ones for each function.
3214    
3215     7. The NON-UNIX-USE file had not been updated for the different set of source
3216        files that come with release 6. I also added a few comments about the C++
3217        wrapper.
3218    
3219    
3220    Version 6.0 07-Jun-05
3221    ---------------------
3222    
3223     1. Some minor internal re-organization to help with my DFA experiments.
3224    
3225     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3226        didn't matter for the library itself when fully configured, but did matter
3227        when compiling without UCP support, or within Exim, where the ucp files are
3228        not imported.
3229    
3230     3. Refactoring of the library code to split up the various functions into
3231        different source modules. The addition of the new DFA matching code (see
3232        below) to a single monolithic source would have made it really too
3233        unwieldy, quite apart from causing all the code to be include in a
3234        statically linked application, when only some functions are used. This is
3235        relevant even without the DFA addition now that patterns can be compiled in
3236        one application and matched in another.
3237    
3238        The downside of splitting up is that there have to be some external
3239        functions and data tables that are used internally in different modules of
3240        the library but which are not part of the API. These have all had their
3241        names changed to start with "_pcre_" so that they are unlikely to clash
3242        with other external names.
3243    
3244     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3245        a different (DFA) algorithm. Although it is slower than the original
3246        function, it does have some advantages for certain types of matching
3247        problem.
3248    
3249     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3250        including restarting after a partial match.
3251    
3252     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3253        defined when compiling for Windows was sent to me. I have put it into the
3254        code, though I have no means of testing or verifying it.
3255    
3256     7. Added the pcre_refcount() auxiliary function.
3257    
3258     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3259        match before or at the first newline in the subject string. In pcretest,
3260        the /f option on a pattern can be used to set this.
3261    
3262     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3263        would behave wrongly. This has been present in PCRE since release 4.0.
3264    
3265    10. A number of changes to the pcregrep command:
3266    
3267        (a) Refactored how -x works; insert ^(...)$ instead of setting
3268            PCRE_ANCHORED and checking the length, in preparation for adding
3269            something similar for -w.
3270    
3271        (b) Added the -w (match as a word) option.
3272    
3273        (c) Refactored the way lines are read and buffered so as to have more
3274            than one at a time available.
3275    
3276        (d) Implemented a pcregrep test script.
3277    
3278        (e) Added the -M (multiline match) option. This allows patterns to match
3279            over several lines of the subject. The buffering ensures that at least
3280            8K, or the rest of the document (whichever is the shorter) is available
3281            for matching (and similarly the previous 8K for lookbehind assertions).
3282    
3283        (f) Changed the --help output so that it now says
3284    
3285              -w, --word-regex(p)
3286    
3287            instead of two lines, one with "regex" and the other with "regexp"
3288            because that confused at least one person since the short forms are the
3289            same. (This required a bit of code, as the output is generated
3290            automatically from a table. It wasn't just a text change.)
3291    
3292        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3293            option but starts with a hyphen. Could be a pattern or a path name
3294            starting with a hyphen, for instance.
3295    
3296        (h) "-" can be given as a file name to represent stdin.
3297    
3298        (i) When file names are being printed, "(standard input)" is used for
3299            the standard input, for compatibility with GNU grep. Previously
3300            "<stdin>" was used.
3301    
3302        (j) The option --label=xxx can be used to supply a name to be used for
3303            stdin when file names are being printed. There is no short form.
3304    
3305        (k) Re-factored the options decoding logic because we are going to add
3306            two more options that take data. Such options can now be given in four
3307            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3308    
3309        (l) Added the -A, -B, and -C options for requesting that lines of context
3310            around matches be printed.
3311    
3312        (m) Added the -L option to print the names of files that do not contain
3313            any matching lines, that is, the complement of -l.
3314    
3315        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3316            continue to scan other files.
3317    
3318        (o) The -s option was incorrectly implemented. For compatibility with other
3319            greps, it now suppresses the error message for a non-existent or non-
3320            accessible file (but not the return code). There is a new option called
3321            -q that suppresses the output of matching lines, which was what -s was
3322            previously doing.
3323    
3324        (p) Added --include and --exclude options to specify files for inclusion
3325            and exclusion when recursing.
3326    
3327    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3328        Hopefully, it now does.
3329    
3330    12. Missing cast in pcre_study().
3331    
3332    13. Added an "uninstall" target to the makefile.
3333    
3334    14. Replaced "extern" in the function prototypes in Makefile.in with
3335        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3336        world, but is set differently for Windows.
3337    
3338    15. Added a second compiling function called pcre_compile2(). The only
3339        difference is that it has an extra argument, which is a pointer to an
3340        integer error code. When there is a compile-time failure, this is set
3341        non-zero, in addition to the error test pointer being set to point to an
3342        error message. The new argument may be NULL if no error number is required
3343        (but then you may as well call pcre_compile(), which is now just a
3344        wrapper). This facility is provided because some applications need a
3345        numeric error indication, but it has also enabled me to tidy up the way
3346        compile-time errors are handled in the POSIX wrapper.
3347    
3348    16. Added VPATH=.libs to the makefile; this should help when building with one
3349        prefix path and installing with another. (Or so I'm told by someone who
3350        knows more about this stuff than I do.)
3351    
3352    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3353        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3354        match everything, including newlines. This is not POSIX-compatible, but
3355        somebody wanted the feature. From pcretest it can be activated by using
3356        both the P and the s flags.
3357    
3358    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3359    
3360    19. libpcre.pc was being incorrectly installed as executable.
3361    
3362    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3363        it now also looks for '\r' so that it will work unmodified on Windows.
3364    
3365    21. Added Google's contributed C++ wrapper to the distribution.
3366    
3367    22. Added some untidy missing memory free() calls in pcretest, to keep
3368        Electric Fence happy when testing.
3369    
3370    
3371    
3372    Version 5.0 13-Sep-04
3373    ---------------------
3374    
3375     1. Internal change: literal characters are no longer packed up into items
3376        containing multiple characters in a single byte-string. Each character
3377        is now matched using a separate opcode. However, there may be more than one
3378        byte in the character in UTF-8 mode.
3379    
3380     2. The pcre_callout_block structure has two new fields: pattern_position and
3381        next_item_length. These contain the offset in the pattern to the next match
3382        item, and its length, respectively.
3383    
3384     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3385        insertion of callouts before each pattern item. Added the /C option to
3386        pcretest to make use of this.
3387    
3388     4. On the advice of a Windows user, the lines
3389    
3390          #if defined(_WIN32) || defined(WIN32)
3391          _setmode( _fileno( stdout ), 0x8000 );
3392          #endif  /* defined(_WIN32) || defined(WIN32) */
3393    
3394        have been added to the source of pcretest. This apparently does useful
3395        magic in relation to line terminators.
3396    
3397     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3398        for the benefit of those environments where the "b" makes a difference.
3399    
3400     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3401        to know about it. I have put a hack into configure.in that adds in code
3402        to set GCC=yes if CC=icc. This seems to end up at a point in the
3403        generated configure script that is early enough to affect the setting of
3404        compiler options, which is what is needed, but I have no means of testing
3405        whether it really works. (The user who reported this had patched the
3406        generated configure script, which of course I cannot do.)
3407    
3408        LATER: After change 22 below (new libtool files), the configure script
3409        seems to know about icc (and also ecc). Therefore, I have commented out
3410        this hack in configure.in.
3411    
3412     7. Added support for pkg-config (2 patches were sent in).
3413    
3414     8. Negated POSIX character classes that used a combination of internal tables
3415        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3416        [[:^ascii]]. Typically, they would match almost any characters. The other
3417        POSIX classes were not broken in this way.
3418    
3419     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3420        to find the match, as PCRE was deluded into thinking that the match had to
3421        start at the start point or following a newline. The same bug applied to
3422        patterns with negative forward assertions or any backward assertions
3423        preceding ".*" at the start, unless the pattern required a fixed first
3424        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3425    
3426    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3427        starting at the last subject character, bytes beyond the end of the subject
3428        string were read.
3429    
3430    11. Renamed the variable "class" as "classbits" to make life easier for C++
3431        users. (Previously there was a macro definition, but it apparently wasn't
3432        enough.)
3433    
3434    12. Added the new field "tables" to the extra data so that tables can be passed
3435        in at exec time, or the internal tables can be re-selected. This allows
3436        a compiled regex to be saved and re-used at a later time by a different
3437        program that might have everything at different addresses.
3438    
3439    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3440        -R library as well as a -L library.
3441    
3442    14. The debugging options of pcretest (-d on the command line or D on a
3443        pattern) showed incorrect output for anything following an extended class
3444        that contained multibyte characters and which was followed by a quantifier.
3445    
3446    15. Added optional support for general category Unicode character properties
3447        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3448        support. It adds about 90K to the size of the library. The meanings of the
3449        inbuilt class escapes such as \d and \s have NOT been changed.
3450    
3451    16. Updated pcredemo.c to include calls to free() to release the memory for the
3452        compiled pattern.
3453    
3454    17. The generated file chartables.c was being created in the source directory
3455        instead of in the building directory. This caused the build to fail if the
3456        source directory was different from the building directory, and was
3457        read-only.
3458    
3459    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3460        file. No doubt somebody will tell me if they don't make sense... Also added
3461        Dan Mooney's comments about building on OpenVMS.
3462    
3463    19. Added support for partial matching via the PCRE_PARTIAL option for
3464        pcre_exec() and the \P data escape in pcretest.
3465    
3466    20. Extended pcretest with 3 new pattern features:
3467    
3468        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3469              write the compiled pattern to the file whose name is "rest-of-line".
3470              This is a straight binary dump of the data, with the saved pointer to
3471              the character tables forced to be NULL. The study data, if any, is
3472              written too. After writing, pcretest reads a new pattern.
3473    
3474        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3475              compiled pattern from the given file. There must not be any
3476              occurrences of "<" in the file name (pretty unlikely); if there are,
3477              pcretest will instead treat the initial "<" as a pattern delimiter.
3478              After reading in the pattern, pcretest goes on to read data lines as
3479              usual.
3480    
3481        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3482              and 16-bit fields in a compiled pattern, to simulate a pattern that
3483              was compiled on a host of opposite endianness.
3484    
3485    21. The pcre-exec() function can now cope with patterns that were compiled on
3486        hosts of opposite endianness, with this restriction:
3487    
3488          As for any compiled expression that is saved and used later, the tables
3489          pointer field cannot be preserved; the extra_data field in the arguments
3490          to pcre_exec() should be used to pass in a tables address if a value
3491          other than the default internal tables were used at compile time.
3492    
3493    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3494        now diagnosed as an error. Previously, most of the time, a negative number
3495        would have been treated as zero, but if in addition "ovector" was passed as
3496        NULL, a crash could occur.
3497    
3498    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3499        new versions from the libtool 1.5 distribution (the last one is a copy of
3500        a file called libtool.m4). This seems to have fixed the need to patch
3501        "configure" to support Darwin 1.3 (which I used to do). However, I still
3502        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3503        workstation).
3504    
3505    24. Changed the PCRE licence to be the more standard "BSD" licence.
3506    
3507    
3508    Version 4.5 01-Dec-03
3509    ---------------------
3510    
3511     1. There has been some re-arrangement of the code for the match() function so
3512        that it can be compiled in a version that does not call itself recursively.
3513        Instead, it keeps those local variables that need separate instances for
3514        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3515        needs to "recurse". Keeping track of where control must go is done by means
3516        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3517        hide most of the details from the main code, and operates only if
3518        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3519        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3520        operating.
3521    
3522        To make it easier for callers to provide specially tailored get/free
3523        functions for this usage, two new functions, pcre_stack_malloc, and
3524        pcre_stack_free, are used. They are always called in strict stacking order,
3525        and the size of block requested is always the same.
3526    
3527        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3528        PCRE has been compiled to use the stack or the heap for recursion. The
3529        -C option of pcretest uses this to show which version is compiled.
3530