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

Diff of /code/trunk/ChangeLog

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

revision 55 by nigel, Sat Feb 24 21:39:46 2007 UTC revision 1011 by ph10, Sat Aug 25 11:36:15 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.32
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 in the interpreters so that it now matches a
54        Unicode extended grapheme cluster.
55    
56    
57    
58    Version 8.31 06-July-2012
59    -------------------------
60    
61    1.  Fixing a wrong JIT test case and some compiler warnings.
62    
63    2.  Removed a bashism from the RunTest script.
64    
65    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
66        to unsigned type, result still unsigned" that was given by an MS compiler
67        on encountering the code "-sizeof(xxx)".
68    
69    4.  Partial matching support is added to the JIT compiler.
70    
71    5.  Fixed several bugs concerned with partial matching of items that consist
72        of more than one character:
73    
74        (a) /^(..)\1/ did not partially match "aba" because checking references was
75            done on an "all or nothing" basis. This also applied to repeated
76            references.
77    
78        (b) \R did not give a hard partial match if \r was found at the end of the
79            subject.
80    
81        (c) \X did not give a hard partial match after matching one or more
82            characters at the end of the subject.
83    
84        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
85            a partial match for the string "\r".
86    
87        (e) When newline was set to CRLF, the metacharacter "." did not recognize
88            a partial match for a CR character at the end of the subject string.
89    
90    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
91        running pcretest, the text "(JIT)" added to the output whenever JIT is
92        actually used to run the match.
93    
94    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
95        or /S+[+] with a digit between 1 and 7.
96    
97    8.  OP_NOT now supports any UTF character not just single-byte ones.
98    
99    9.  (*MARK) control verb is now supported by the JIT compiler.
100    
101    10. The command "./RunTest list" lists the available tests without actually
102        running any of them. (Because I keep forgetting what they all are.)
103    
104    11. Add PCRE_INFO_MAXLOOKBEHIND.
105    
106    12. Applied a (slightly modified) user-supplied patch that improves performance
107        when the heap is used for recursion (compiled with --disable-stack-for-
108        recursion). Instead of malloc and free for each heap frame each time a
109        logical recursion happens, frames are retained on a chain and re-used where
110        possible. This sometimes gives as much as 30% improvement.
111    
112    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
113        call.
114    
115    14. As documented, (*COMMIT) is now confined to within a positive assertion.
116    
117    15. It is now possible to link pcretest with libedit as an alternative to
118        libreadline.
119    
120    16. (*COMMIT) control verb is now supported by the JIT compiler.
121    
122    17. The Unicode data tables have been updated to Unicode 6.1.0.
123    
124    18. Added --file-list option to pcregrep.
125    
126    19. Added binary file support to pcregrep, including the -a, --binary-files,
127        -I, and --text options.
128    
129    20. The madvise function is renamed for posix_madvise for QNX compatibility
130        reasons. Fixed by Giuseppe D'Angelo.
131    
132    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
133    
134    22. Changed the option for creating symbolic links for 16-bit man pages from
135        -s to -sf so that re-installing does not cause issues.
136    
137    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
138    
139    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
140        matches in certain environments (the workspace was not being correctly
141        retained). Also added to pcre_dfa_exec() a simple plausibility check on
142        some of the workspace data at the beginning of a restart.
143    
144    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
145        was not doing so when it should - probably a typo introduced by SVN 528
146        (change 8.10/14).
147    
148    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
149        \w+ when the character tables indicated that \x{c4} was a word character.
150        There were several related cases, all because the tests for doing a table
151        lookup were testing for characters less than 127 instead of 255.
152    
153    27. If a pattern contains capturing parentheses that are not used in a match,
154        their slots in the ovector are set to -1. For those that are higher than
155        any matched groups, this happens at the end of processing. In the case when
156        there were back references that the ovector was too small to contain
157        (causing temporary malloc'd memory to be used during matching), and the
158        highest capturing number was not used, memory off the end of the ovector
159        was incorrectly being set to -1. (It was using the size of the temporary
160        memory instead of the true size.)
161    
162    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
163        ovector size, it uses memory at the end of the block that it has got.
164    
165    29. Check for an overlong MARK name and give an error at compile time. The
166        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
167    
168    30. JIT compiler update.
169    
170    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
171        Rill for the patch.
172    
173    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
174    
175    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
176    
177    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
178        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
179        was enabled.
180    
181    35. Improve JIT code generation for greedy plus quantifier.
182    
183    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
184        1 to "aa" instead of to an empty string. The bug affected repeated groups
185        that could potentially match an empty string.
186    
187    37. Optimizing single character iterators in JIT.
188    
189    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
190        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
191        codepoints that are too big for the mode are faulted, and in a UTF mode,
192        disallowed codepoints are also faulted.
193    
194    39. If PCRE was compiled with UTF support, in three places in the DFA
195        matcher there was code that should only have been obeyed in UTF mode, but
196        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
197        processing when bytes with values greater than 127 were present. In 16-bit
198        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
199        both cases the values are those that cannot be the first data item in a UTF
200        character. The three items that might have provoked this were recursions,
201        possessively repeated groups, and atomic groups.
202    
203    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
204        and pcregrep, because some OS require shared objects to be explicitly
205        passed to ld, causing the link step to fail if they are not.
206    
207    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
208        mode, patterns that started with \h* or \R* might be incorrectly matched.
209    
210    
211    Version 8.30 04-February-2012
212    -----------------------------
213    
214    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
215        name is defined in ctype.h.
216    
217    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
218        only in quite long subpatterns.
219    
220    3.  Removed the function pcre_info(), which has been obsolete and deprecated
221        since it was replaced by pcre_fullinfo() in February 2000.
222    
223    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
224        match a (*MARK), and the match failed at the start of the subject, a
225        reference to memory before the start of the subject could occur. This bug
226        was introduced by fix 17 of release 8.21.
227    
228    5.  A reference to an unset group with zero minimum repetition was giving
229        totally wrong answers (in non-JavaScript-compatibility mode). For example,
230        /(another)?(\1?)test/ matched against "hello world test". This bug was
231        introduced in release 8.13.
232    
233    6.  Add support for 16-bit character strings (a large amount of work involving
234        many changes and refactorings).
235    
236    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
237        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
238        from a file.
239    
240    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
241        rounding is not applied in this particular case).
242    
243    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
244        if they appear, or are escaped, in patterns.
245    
246    10. Get rid of a number of -Wunused-but-set-variable warnings.
247    
248    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
249        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
250        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
251        also returns the mark "x". This bug applied to capturing parentheses,
252        non-capturing parentheses, and atomic parentheses. It also applied to some
253        assertions.
254    
255    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
256        information out of configure.ac instead of relying on pcre.h.generic, which
257        is not stored in the repository.
258    
259    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
260        -lreadline.
261    
262    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
263    
264    15. Applied Graycode's patch to put the top-level frame on the stack rather
265        than the heap when not using the stack for recursion. This gives a
266        performance improvement in many cases when recursion is not deep.
267    
268    16. Experimental code added to "pcretest -C" to output the stack frame size.
269    
270    
271    Version 8.21 12-Dec-2011
272    ------------------------
273    
274    1.  Updating the JIT compiler.
275    
276    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
277        are added as well.
278    
279    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
280        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
281        calling _pcre_jit_exec. Some extra comments are added.
282    
283    4.  (*MARK) settings inside atomic groups that do not contain any capturing
284        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
285        was introduced by change 18 for 8.20.
286    
287    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
288        ECMA-262 standard.
289    
290    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
291        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
292        This bug was probably introduced by change 9 of 8.13.
293    
294    7.  While fixing 6 above, I noticed that a number of other items were being
295        incorrectly rejected as "not fixed length". This arose partly because newer
296        opcodes had not been added to the fixed-length checking code. I have (a)
297        corrected the bug and added tests for these items, and (b) arranged for an
298        error to occur if an unknown opcode is encountered while checking for fixed
299        length instead of just assuming "not fixed length". The items that were
300        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
301        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
302        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
303    
304    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
305        being incorrectly compiled and would have given unpredicatble results.
306    
307    9.  A possessively repeated subpattern with minimum repeat count greater than
308        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
309        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
310        the first (A) could occur when it should not.
311    
312    10. Add a cast and remove a redundant test from the code.
313    
314    11. JIT should use pcre_malloc/pcre_free for allocation.
315    
316    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
317        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
318        is anything other than /usr, -L is still shown).
319    
320    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
321    
322    14. Perl does not support \N without a following name in a [] class; PCRE now
323        also gives an error.
324    
325    15. If a forward reference was repeated with an upper limit of around 2000,
326        it caused the error "internal error: overran compiling workspace". The
327        maximum number of forward references (including repeats) was limited by the
328        internal workspace, and dependent on the LINK_SIZE. The code has been
329        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
330        the default depends on LINK_SIZE. There is a new upper limit (for safety)
331        of around 200,000 forward references. While doing this, I also speeded up
332        the filling in of repeated forward references.
333    
334    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
335        incorrectly expecting the subject to contain another "a" after the start.
336    
337    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
338        in the match, the SKIP should be ignored. This was not happening; instead
339        the SKIP was being treated as NOMATCH. For patterns such as
340        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
341        tested.
342    
343    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
344        now much more compatible with Perl, in particular in cases where the result
345        is a non-match for a non-anchored pattern. For example, if
346        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
347        "m", where previously it did not return a name. A side effect of this
348        change is that for partial matches, the last encountered mark name is
349        returned, as for non matches. A number of tests that were previously not
350        Perl-compatible have been moved into the Perl-compatible test files. The
351        refactoring has had the pleasing side effect of removing one argument from
352        the match() function, thus reducing its stack requirements.
353    
354    19. If the /S+ option was used in pcretest to study a pattern using JIT,
355        subsequent uses of /S (without +) incorrectly behaved like /S+.
356    
357    21. Retrieve executable code size support for the JIT compiler and fixing
358        some warnings.
359    
360    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
361        not work when the shorter character appeared right at the end of the
362        subject string.
363    
364    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
365        systems.
366    
367    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
368        output it when the /M option is used in pcretest.
369    
370    25. The CheckMan script was not being included in the distribution. Also, added
371        an explicit "perl" to run Perl scripts from the PrepareRelease script
372        because this is reportedly needed in Windows.
373    
374    26. If study data was being save in a file and studying had not found a set of
375        "starts with" bytes for the pattern, the data written to the file (though
376        never used) was taken from uninitialized memory and so caused valgrind to
377        complain.
378    
379    27. Updated RunTest.bat as provided by Sheri Pierce.
380    
381    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
382    
383    29. Computation of memory usage for the table of capturing group names was
384        giving an unnecessarily large value.
385    
386    
387    Version 8.20 21-Oct-2011
388    ------------------------
389    
390    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
391        a POSIX class. After further experiments with Perl, which convinced me that
392        Perl has bugs and confusions, a closing square bracket is no longer allowed
393        in a POSIX name. This bug also affected patterns with classes that started
394        with full stops.
395    
396    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
397        captured substring, but while checking the failing first alternative,
398        substring 1 is temporarily captured. If the output vector supplied to
399        pcre_exec() was not big enough for this capture, the yield of the function
400        was still zero ("insufficient space for captured substrings"). This cannot
401        be totally fixed without adding another stack variable, which seems a lot
402        of expense for a edge case. However, I have improved the situation in cases
403        such as /(a)(b)x|abc/ matched against "abc", where the return code
404        indicates that fewer than the maximum number of slots in the ovector have
405        been set.
406    
407    3.  Related to (2) above: when there are more back references in a pattern than
408        slots in the output vector, pcre_exec() uses temporary memory during
409        matching, and copies in the captures as far as possible afterwards. It was
410        using the entire output vector, but this conflicts with the specification
411        that only 2/3 is used for passing back captured substrings. Now it uses
412        only the first 2/3, for compatibility. This is, of course, another edge
413        case.
414    
415    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
416        main code base, and can be used by building with --enable-jit. When this is
417        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
418        runtime --no-jit option is given.
419    
420    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
421        ovector, the return from the function was zero, implying that there were
422        other matches that did not fit. The correct "exactly full" value is now
423        returned.
424    
425    6.  If a subpattern that was called recursively or as a subroutine contained
426        (*PRUNE) or any other control that caused it to give a non-standard return,
427        invalid errors such as "Error -26 (nested recursion at the same subject
428        position)" or even infinite loops could occur.
429    
430    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
431        computing the minimum length on reaching *ACCEPT, and so ended up with the
432        wrong value of 1 rather than 0. Further investigation indicates that
433        computing a minimum subject length in the presence of *ACCEPT is difficult
434        (think back references, subroutine calls), and so I have changed the code
435        so that no minimum is registered for a pattern that contains *ACCEPT.
436    
437    8.  If (*THEN) was present in the first (true) branch of a conditional group,
438        it was not handled as intended. [But see 16 below.]
439    
440    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
441        Sheri Pierce.
442    
443    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
444        the first byte in a match must be "a".
445    
446    11. Change 17 for 8.13 increased the recursion depth for patterns like
447        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
448        pattern contains any instances of (*THEN). If it does not, the old
449        optimizations are restored. It would be nice to do this on a per-group
450        basis, but at the moment that is not feasible.
451    
452    12. In some environments, the output of pcretest -C is CRLF terminated. This
453        broke RunTest's code that checks for the link size. A single white space
454        character after the value is now allowed for.
455    
456    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
457        For "fr", it uses the Windows-specific input and output files.
458    
459    14. If (*THEN) appeared in a group that was called recursively or as a
460        subroutine, it did not work as intended. [But see next item.]
461    
462    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
463        pattern fragments (but not containing any | characters). If A and B are
464        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
465        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
466        D. In other words, Perl considers parentheses that do not contain any |
467        characters to be part of a surrounding alternative, whereas PCRE was
468        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
469        differently. PCRE now behaves in the same way as Perl, except in the case
470        of subroutine/recursion calls such as (?1) which have in any case always
471        been different (but PCRE had them first :-).
472    
473    16. Related to 15 above: Perl does not treat the | in a conditional group as
474        creating alternatives. Such a group is treated in the same way as an
475        ordinary group without any | characters when processing (*THEN). PCRE has
476        been changed to match Perl's behaviour.
477    
478    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
479        RunGrepTest script failed.
480    
481    18. Change 22 for version 13 caused atomic groups to use more stack. This is
482        inevitable for groups that contain captures, but it can lead to a lot of
483        stack use in large patterns. The old behaviour has been restored for atomic
484        groups that do not contain any capturing parentheses.
485    
486    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
487        suppress the check for a minimum subject length at run time. (If it was
488        given to pcre_exec() or pcre_dfa_exec() it did work.)
489    
490    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
491        fail to work when decoding hex characters in data strings in EBCDIC
492        environments.
493    
494    21. It appears that in at least one Mac OS environment, the isxdigit() function
495        is implemented as a macro that evaluates to its argument more than once,
496        contravening the C 90 Standard (I haven't checked a later standard). There
497        was an instance in pcretest which caused it to go wrong when processing
498        \x{...} escapes in subject strings. The has been rewritten to avoid using
499        things like p++ in the argument of isxdigit().
500    
501    
502    Version 8.13 16-Aug-2011
503    ------------------------
504    
505    1.  The Unicode data tables have been updated to Unicode 6.0.0.
506    
507    2.  Two minor typos in pcre_internal.h have been fixed.
508    
509    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
510        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
511        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
512    
513    4.  There were a number of related bugs in the code for matching backrefences
514        caselessly in UTF-8 mode when codes for the characters concerned were
515        different numbers of bytes. For example, U+023A and U+2C65 are an upper
516        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
517        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
518        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
519        2-byte code at the end of the subject (it thought there wasn't enough data
520        left).
521    
522    5.  Comprehensive information about what went wrong is now returned by
523        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
524        as the output vector has at least 2 elements. The offset of the start of
525        the failing character and a reason code are placed in the vector.
526    
527    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
528        now returned is for the first byte of the failing character, instead of the
529        last byte inspected. This is an incompatible change, but I hope it is small
530        enough not to be a problem. It makes the returned offset consistent with
531        pcre_exec() and pcre_dfa_exec().
532    
533    7.  pcretest now gives a text phrase as well as the error number when
534        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
535        failure, the offset and reason code are output.
536    
537    8.  When \R was used with a maximizing quantifier it failed to skip backwards
538        over a \r\n pair if the subsequent match failed. Instead, it just skipped
539        back over a single character (\n). This seems wrong (because it treated the
540        two characters as a single entity when going forwards), conflicts with the
541        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
542        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
543        has been changed.
544    
545    9.  Some internal refactoring has changed the processing so that the handling
546        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
547        time (the PCRE_DOTALL option was changed this way some time ago: version
548        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
549        which was always a bit of a fudge. It also means that there is one less
550        argument for the match() function, which reduces its stack requirements
551        slightly. This change also fixes an incompatibility with Perl: the pattern
552        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
553    
554    10. More internal refactoring has drastically reduced the number of recursive
555        calls to match() for possessively repeated groups such as (abc)++ when
556        using pcre_exec().
557    
558    11. While implementing 10, a number of bugs in the handling of groups were
559        discovered and fixed:
560    
561        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
562        (a|)*(?1) gave a compile-time internal error.
563        ((a|)+)+  did not notice that the outer group could match an empty string.
564        (^a|^)+   was not marked as anchored.
565        (.*a|.*)+ was not marked as matching at start or after a newline.
566    
567    12. Yet more internal refactoring has removed another argument from the match()
568        function. Special calls to this function are now indicated by setting a
569        value in a variable in the "match data" data block.
570    
571    13. Be more explicit in pcre_study() instead of relying on "default" for
572        opcodes that mean there is no starting character; this means that when new
573        ones are added and accidentally left out of pcre_study(), testing should
574        pick them up.
575    
576    14. The -s option of pcretest has been documented for ages as being an old
577        synonym of -m (show memory usage). I have changed it to mean "force study
578        for every regex", that is, assume /S for every regex. This is similar to -i
579        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
580        using it. It makes it easier to run collections of tests with and without
581        study enabled, and thereby test pcre_study() more easily. All the standard
582        tests are now run with and without -s (but some patterns can be marked as
583        "never study" - see 20 below).
584    
585    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
586        restoration of the capturing data to the outer values was not happening
587        correctly.
588    
589    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
590        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
591        pattern had matched an empty string, and so incorrectly returned a no
592        match.
593    
594    17. There was optimizing code for the last branch of non-capturing parentheses,
595        and also for the obeyed branch of a conditional subexpression, which used
596        tail recursion to cut down on stack usage. Unfortunately, now that there is
597        the possibility of (*THEN) occurring in these branches, tail recursion is
598        no longer possible because the return has to be checked for (*THEN). These
599        two optimizations have therefore been removed. [But see 8.20/11 above.]
600    
601    18. If a pattern containing \R was studied, it was assumed that \R always
602        matched two bytes, thus causing the minimum subject length to be
603        incorrectly computed because \R can also match just one byte.
604    
605    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
606        was incorrectly computed.
607    
608    20. If /S is present twice on a test pattern in pcretest input, it now
609        *disables* studying, thereby overriding the use of -s on the command line
610        (see 14 above). This is necessary for one or two tests to keep the output
611        identical in both cases.
612    
613    21. When (*ACCEPT) was used in an assertion that matched an empty string and
614        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
615    
616    22. When an atomic group that contained a capturing parenthesis was
617        successfully matched, but the branch in which it appeared failed, the
618        capturing was not being forgotten if a higher numbered group was later
619        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
620        group 1 to "a", when in fact it should be unset. This applied to multi-
621        branched capturing and non-capturing groups, repeated or not, and also to
622        positive assertions (capturing in negative assertions does not happen
623        in PCRE) and also to nested atomic groups.
624    
625    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
626        subject after a captured substring, to make it easier to tell which of a
627        number of identical substrings has been captured.
628    
629    24. The way atomic groups are processed by pcre_exec() has been changed so that
630        if they are repeated, backtracking one repetition now resets captured
631        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
632        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
633        "aaa". Previously, it would have been "a". As part of this code
634        refactoring, the way recursive calls are handled has also been changed.
635    
636    25. If an assertion condition captured any substrings, they were not passed
637        back unless some other capturing happened later. For example, if
638        (?(?=(a))a) was matched against "a", no capturing was returned.
639    
640    26. When studying a pattern that contained subroutine calls or assertions,
641        the code for finding the minimum length of a possible match was handling
642        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
643        group 1 called group 2 while simultaneously a separate group 2 called group
644        1). A stack overflow occurred in this case. I have fixed this by limiting
645        the recursion depth to 10.
646    
647    27. Updated RunTest.bat in the distribution to the version supplied by Tom
648        Fortmann. This supports explicit test numbers on the command line, and has
649        argument validation and error reporting.
650    
651    28. An instance of \X with an unlimited repeat could fail if at any point the
652        first character it looked at was a mark character.
653    
654    29. Some minor code refactoring concerning Unicode properties and scripts
655        should reduce the stack requirement of match() slightly.
656    
657    30. Added the '=' option to pcretest to check the setting of unused capturing
658        slots at the end of the pattern, which are documented as being -1, but are
659        not included in the return count.
660    
661    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
662        compiled something random. Now it gives a compile-time error (as does
663        Perl).
664    
665    32. A *MARK encountered during the processing of a positive assertion is now
666        recorded and passed back (compatible with Perl).
667    
668    33. If --only-matching or --colour was set on a pcregrep call whose pattern
669        had alternative anchored branches, the search for a second match in a line
670        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
671        matched the line "0102" twice. The same bug affected patterns that started
672        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
673        twice.
674    
675    34. Previously, PCRE did not allow quantification of assertions. However, Perl
676        does, and because of capturing effects, quantifying parenthesized
677        assertions may at times be useful. Quantifiers are now allowed for
678        parenthesized assertions.
679    
680    35. A minor code tidy in pcre_compile() when checking options for \R usage.
681    
682    36. \g was being checked for fancy things in a character class, when it should
683        just be a literal "g".
684    
685    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
686        appearance of a nested POSIX class supersedes an apparent external class.
687        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
688        unescaped square brackets may also appear as part of class names. For
689        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
690        more like Perl. (But see 8.20/1 above.)
691    
692    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
693        was because it thought it was \N{name}, which is not supported).
694    
695    39. Add minix to OS list not supporting the -S option in pcretest.
696    
697    40. PCRE tries to detect cases of infinite recursion at compile time, but it
698        cannot analyze patterns in sufficient detail to catch mutual recursions
699        such as ((?1))((?2)). There is now a runtime test that gives an error if a
700        subgroup is called recursively as a subpattern for a second time at the
701        same position in the subject string. In previous releases this might have
702        been caught by the recursion limit, or it might have run out of stack.
703    
704    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
705        happen only once. PCRE was, however incorrectly giving a compile time error
706        "recursive call could loop indefinitely" because it cannot analyze the
707        pattern in sufficient detail. The compile time test no longer happens when
708        PCRE is compiling a conditional subpattern, but actual runaway loops are
709        now caught at runtime (see 40 above).
710    
711    42. It seems that Perl allows any characters other than a closing parenthesis
712        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
713        has been changed to be the same.
714    
715    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
716        as not to get warnings when autogen.sh is called. Also changed
717        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
718    
719    44. To help people who use pcregrep to scan files containing exceedingly long
720        lines, the following changes have been made:
721    
722        (a) The default value of the buffer size parameter has been increased from
723            8K to 20K. (The actual buffer used is three times this size.)
724    
725        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
726            PCRE is built.
727    
728        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
729            to be set at run time.
730    
731        (d) Numerical values in pcregrep options can be followed by K or M, for
732            example --buffer-size=50K.
733    
734        (e) If a line being scanned overflows pcregrep's buffer, an error is now
735            given and the return code is set to 2.
736    
737    45. Add a pointer to the latest mark to the callout data block.
738    
739    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
740        partial match of an empty string instead of no match. This was specific to
741        the use of ".".
742    
743    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
744        complete match instead of a partial match. This bug was dependent on both
745        the PCRE_UTF8 and PCRE_DOTALL options being set.
746    
747    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
748        starting byte set, because \b was not being ignored.
749    
750    
751    Version 8.12 15-Jan-2011
752    ------------------------
753    
754    1.  Fixed some typos in the markup of the man pages, and wrote a script that
755        checks for such things as part of the documentation building process.
756    
757    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
758        --match-limit and --recursion-limit options (added for 8.11). In
759        particular, this made one of the standard tests fail. (The integer value
760        went into the wrong half of a long int.)
761    
762    3.  If the --colour option was given to pcregrep with -v (invert match), it
763        did strange things, either producing crazy output, or crashing. It should,
764        of course, ignore a request for colour when reporting lines that do not
765        match.
766    
767    4.  Another pcregrep bug caused similar problems if --colour was specified with
768        -M (multiline) and the pattern match finished with a line ending.
769    
770    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
771        matched in multiline mode, the following line was shown as part of the
772        match. This seems wrong, so I have changed it.
773    
774    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
775        the check for further matches in the same line (so they could be coloured)
776        to overrun the end of the current line. If another match was found, it was
777        incorrectly shown (and then shown again when found in the next line).
778    
779    7.  If pcregrep was compiled under Windows, there was a reference to the
780        function pcregrep_exit() before it was defined. I am assuming this was
781        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
782        reported by a user. I've moved the definition above the reference.
783    
784    
785    Version 8.11 10-Dec-2010
786    ------------------------
787    
788    1.  (*THEN) was not working properly if there were untried alternatives prior
789        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
790        backtracked to try for "b" instead of moving to the next alternative branch
791        at the same level (in this case, to look for "c"). The Perl documentation
792        is clear that when (*THEN) is backtracked onto, it goes to the "next
793        alternative in the innermost enclosing group".
794    
795    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
796        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
797        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
798        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
799        (*THEN).
800    
801    3.  If \s appeared in a character class, it removed the VT character from
802        the class, even if it had been included by some previous item, for example
803        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
804        of \s, but is part of the POSIX "space" class.)
805    
806    4.  A partial match never returns an empty string (because you can always
807        match an empty string at the end of the subject); however the checking for
808        an empty string was starting at the "start of match" point. This has been
809        changed to the "earliest inspected character" point, because the returned
810        data for a partial match starts at this character. This means that, for
811        example, /(?<=abc)def/ gives a partial match for the subject "abc"
812        (previously it gave "no match").
813    
814    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
815        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
816        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
817        has an implication that the given string is incomplete (because a partial
818        match is preferred over a full match). For this reason, these items now
819        give a partial match in this situation. [Aside: previously, the one case
820        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
821        match rather than a full match, which was wrong by the old rules, but is
822        now correct.]
823    
824    6.  There was a bug in the handling of #-introduced comments, recognized when
825        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
826        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
827        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
828        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
829        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
830        places in pcre_compile().
831    
832    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
833        comments when looking ahead for named forward references to subpatterns,
834        the only newline sequence it recognized was NL. It now handles newlines
835        according to the set newline convention.
836    
837    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
838        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
839        cater for a lack of strerror(). These oversights have been fixed.
840    
841    9.  Added --match-limit and --recursion-limit to pcregrep.
842    
843    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
844    
845    11. When the -o option was used, pcregrep was setting a return code of 1, even
846        when matches were found, and --line-buffered was not being honoured.
847    
848    12. Added an optional parentheses number to the -o and --only-matching options
849        of pcregrep.
850    
851    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
852        can match an empty string. The code to do it in pcretest and pcredemo
853        needed fixing:
854    
855        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
856            only one byte after an empty string match just before CRLF (this case
857            just got forgotten; "any" and "anycrlf" were OK).
858    
859        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
860            mode when an empty string match preceded an ASCII character followed by
861            a non-ASCII character. (The code for advancing by one character rather
862            than one byte was nonsense.)
863    
864        (c) The pcredemo.c sample program did not have any code at all to handle
865            the cases when CRLF is a valid newline sequence.
866    
867    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
868        as a starting offset was within the subject string. There is now a new
869        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
870        negative or greater than the length of the string. In order to test this,
871        pcretest is extended to allow the setting of negative starting offsets.
872    
873    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
874        starting offset points to the beginning of a UTF-8 character was
875        unnecessarily clumsy. I tidied it up.
876    
877    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
878        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
879    
880    17. Nobody had reported that the --include_dir option, which was added in
881        release 7.7 should have been called --include-dir (hyphen, not underscore)
882        for compatibility with GNU grep. I have changed it to --include-dir, but
883        left --include_dir as an undocumented synonym, and the same for
884        --exclude-dir, though that is not available in GNU grep, at least as of
885        release 2.5.4.
886    
887    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
888        characters from a string of bytes) have been redefined so as not to use
889        loops, in order to improve performance in some environments. At the same
890        time, I abstracted some of the common code into auxiliary macros to save
891        repetition (this should not affect the compiled code).
892    
893    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
894        compile-time error is now given if \c is not followed by an ASCII
895        character, that is, a byte less than 128. (In EBCDIC mode, the code is
896        different, and any byte value is allowed.)
897    
898    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
899        START_OPTIMIZE option, which is now allowed at compile time - but just
900        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
901        to pcregrep and other applications that have no direct access to PCRE
902        options. The new /Y option in pcretest sets this option when calling
903        pcre_compile().
904    
905    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
906        back references. Groups containing recursive back references were forced to
907        be atomic by that change, but in the case of named groups, the amount of
908        memory required was incorrectly computed, leading to "Failed: internal
909        error: code overflow". This has been fixed.
910    
911    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
912        pcretest.c, to avoid build problems in some Borland environments.
913    
914    
915    Version 8.10 25-Jun-2010
916    ------------------------
917    
918    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
919        THEN.
920    
921    2.  (*ACCEPT) was not working when inside an atomic group.
922    
923    3.  Inside a character class, \B is treated as a literal by default, but
924        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
925        causes the error). The code is unchanged, but I tidied the documentation.
926    
927    4.  Inside a character class, PCRE always treated \R and \X as literals,
928        whereas Perl faults them if its -w option is set. I have changed PCRE so
929        that it faults them when PCRE_EXTRA is set.
930    
931    5.  Added support for \N, which always matches any character other than
932        newline. (It is the same as "." when PCRE_DOTALL is not set.)
933    
934    6.  When compiling pcregrep with newer versions of gcc which may have
935        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
936        declared with attribute warn_unused_result" were given. Just casting the
937        result to (void) does not stop the warnings; a more elaborate fudge is
938        needed. I've used a macro to implement this.
939    
940    7.  Minor change to pcretest.c to avoid a compiler warning.
941    
942    8.  Added four artifical Unicode properties to help with an option to make
943        \s etc use properties (see next item). The new properties are: Xan
944        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
945    
946    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
947        use Unicode properties. (*UCP) at the start of a pattern can be used to set
948        this option. Modified pcretest to add /W to test this facility. Added
949        REG_UCP to make it available via the POSIX interface.
950    
951    10. Added --line-buffered to pcregrep.
952    
953    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
954        studied, and the match started with a letter with a code point greater than
955        127 whose first byte was different to the first byte of the other case of
956        the letter, the other case of this starting letter was not recognized
957        (#976).
958    
959    12. If a pattern that was studied started with a repeated Unicode property
960        test, for example, \p{Nd}+, there was the theoretical possibility of
961        setting up an incorrect bitmap of starting bytes, but fortunately it could
962        not have actually happened in practice until change 8 above was made (it
963        added property types that matched character-matching opcodes).
964    
965    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
966        possible starting bytes for non-anchored patterns.
967    
968    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
969        \R, and also a number of cases that involve Unicode properties, both
970        explicit and implicit when PCRE_UCP is set.
971    
972    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
973        input, it could crash or give wrong results if characters with values
974        greater than 0xc0 were present in the subject string. (Detail: it assumed
975        UTF-8 input when processing these items.)
976    
977    16. Added a lot of (int) casts to avoid compiler warnings in systems where
978        size_t is 64-bit (#991).
979    
980    17. Added a check for running out of memory when PCRE is compiled with
981        --disable-stack-for-recursion (#990).
982    
983    18. If the last data line in a file for pcretest does not have a newline on
984        the end, a newline was missing in the output.
985    
986    19. The default pcre_chartables.c file recognizes only ASCII characters (values
987        less than 128) in its various bitmaps. However, there is a facility for
988        generating tables according to the current locale when PCRE is compiled. It
989        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
990        characters, are recognized by isspace() and therefore were getting set in
991        these tables, and indeed these tables seem to approximate to ISO 8859. This
992        caused a problem in UTF-8 mode when pcre_study() was used to create a list
993        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
994        which of course cannot start UTF-8 characters. I have changed the code so
995        that only real ASCII characters (less than 128) and the correct starting
996        bytes for UTF-8 encodings are set for characters greater than 127 when in
997        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
998        altogether.)
999    
1000    20. Added the /T option to pcretest so as to be able to run tests with non-
1001        standard character tables, thus making it possible to include the tests
1002        used for 19 above in the standard set of tests.
1003    
1004    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1005        reference to a subpattern the other side of a comment that contains an
1006        opening parenthesis caused either an internal compiling error, or a
1007        reference to the wrong subpattern.
1008    
1009    
1010    Version 8.02 19-Mar-2010
1011    ------------------------
1012    
1013    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1014    
1015    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1016        configured.
1017    
1018    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1019        original author of that file, following a query about its status.
1020    
1021    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1022        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1023    
1024    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1025        quantifier applied to a forward-referencing subroutine call, could compile
1026        incorrect code or give the error "internal error: previously-checked
1027        referenced subpattern not found".
1028    
1029    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1030        variables to point to external functions. For these systems, therefore,
1031        pcre_malloc etc. are now initialized to local functions that call the
1032        relevant global functions.
1033    
1034    7.  There were two entries missing in the vectors called coptable and poptable
1035        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1036        I've fixed the data, and added a kludgy way of testing at compile time that
1037        the lengths are correct (equal to the number of opcodes).
1038    
1039    8.  Following on from 7, I added a similar kludge to check the length of the
1040        eint vector in pcreposix.c.
1041    
1042    9.  Error texts for pcre_compile() are held as one long string to avoid too
1043        much relocation at load time. To find a text, the string is searched,
1044        counting zeros. There was no check for running off the end of the string,
1045        which could happen if a new error number was added without updating the
1046        string.
1047    
1048    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1049    
1050    11. \K was not working if it appeared in an atomic group or in a group that
1051        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1052        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1053        the assertion is positive, but not if it is negative.
1054    
1055    12. Change 11 fortuitously reduced the size of the stack frame used in the
1056        "match()" function of pcre_exec.c by one pointer. Forthcoming
1057        implementation of support for (*MARK) will need an extra pointer on the
1058        stack; I have reserved it now, so that the stack frame size does not
1059        decrease.
1060    
1061    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1062        item in branch that calls a recursion is a subroutine call - as in the
1063        second branch in the above example - was incorrectly given the compile-
1064        time error "recursive call could loop indefinitely" because pcre_compile()
1065        was not correctly checking the subroutine for matching a non-empty string.
1066    
1067    14. The checks for overrunning compiling workspace could trigger after an
1068        overrun had occurred. This is a "should never occur" error, but it can be
1069        triggered by pathological patterns such as hundreds of nested parentheses.
1070        The checks now trigger 100 bytes before the end of the workspace.
1071    
1072    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1073    
1074    
1075    Version 8.01 19-Jan-2010
1076    ------------------------
1077    
1078    1.  If a pattern contained a conditional subpattern with only one branch (in
1079        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1080        computed the wrong minimum data length (which is of course zero for such
1081        subpatterns). This could cause incorrect "no match" results.
1082    
1083    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1084        the pattern is reset in the first branch, pcre_compile() failed with
1085        "internal error: code overflow at offset...". This happened only when
1086        the reset was to the original external option setting. (An optimization
1087        abstracts leading options settings into an external setting, which was the
1088        cause of this.)
1089    
1090    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1091        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1092        assertion pattern did not match (meaning that the assertion was true), it
1093        was incorrectly treated as false if the SKIP had been reached during the
1094        matching. This also applied to assertions used as conditions.
1095    
1096    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1097        assertion subpattern, including such a pattern used as a condition,
1098        unpredictable results occurred, instead of the error return
1099        PCRE_ERROR_DFA_UITEM.
1100    
1101    5.  The C++ GlobalReplace function was not working like Perl for the special
1102        situation when an empty string is matched. It now does the fancy magic
1103        stuff that is necessary.
1104    
1105    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1106        removed. (These were left over from very, very early versions of PCRE.)
1107    
1108    7.  Some cosmetic changes to the code to make life easier when compiling it
1109        as part of something else:
1110    
1111        (a) Change DEBUG to PCRE_DEBUG.
1112    
1113        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1114            called "current" as "current_branch", to prevent a collision with the
1115            Linux macro when compiled as a kernel module.
1116    
1117        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1118            prevent a collision with the Linux macro when compiled as a kernel
1119            module.
1120    
1121    8.  In pcre_compile() there are some checks for integer overflows that used to
1122        cast potentially large values to (double). This has been changed to that
1123        when building, a check for int64_t is made, and if it is found, it is used
1124        instead, thus avoiding the use of floating point arithmetic. (There is no
1125        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1126        double.
1127    
1128    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1129        2005 (difference between two addresses compared to an unsigned value).
1130    
1131    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1132        custom one, because of the following reported problem in Windows:
1133    
1134          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1135              under Win32.
1136          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1137              therefore missing the function definition.
1138          - The compiler thus generates a "C" signature for the test function.
1139          - The linker fails to find the "C" function.
1140          - PCRE fails to configure if asked to do so against libbz2.
1141    
1142    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1143        messages were output:
1144    
1145          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1146          rerunning libtoolize, to keep the correct libtool macros in-tree.
1147          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1148    
1149        I have done both of these things.
1150    
1151    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1152        most of the time, it *can* run out if it is given a pattern that contains a
1153        runaway infinite recursion. I updated the discussion in the pcrestack man
1154        page.
1155    
1156    13. Now that we have gone to the x.xx style of version numbers, the minor
1157        version may start with zero. Using 08 or 09 is a bad idea because users
1158        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1159        interpreted as invalid octal numbers. I've updated the previous comment in
1160        configure.ac, and also added a check that gives an error if 08 or 09 are
1161        used.
1162    
1163    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1164        causing partial matching to fail when the end of the subject matched \W
1165        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1166    
1167    15. There were some discrepancies between the declarations in pcre_internal.h
1168        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1169        their definitions. The declarations used "const uschar *" and the
1170        definitions used USPTR. Even though USPTR is normally defined as "const
1171        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1172        reported that: "This difference in casting confuses some C++ compilers, for
1173        example, SunCC recognizes above declarations as different functions and
1174        generates broken code for hbpcre." I have changed the declarations to use
1175        USPTR.
1176    
1177    16. GNU libtool is named differently on some systems. The autogen.sh script now
1178        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1179        (FreeBSD).
1180    
1181    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1182        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1183        comment: "Figure out how to create a longlong from a string: strtoll and
1184        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1185        instance, but it only takes 2 args instead of 3!"
1186    
1187    18. A subtle bug concerned with back references has been fixed by a change of
1188        specification, with a corresponding code fix. A pattern such as
1189        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1190        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1191        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1192        same bug. Such groups have to be quantified to be useful, or contained
1193        inside another quantified group. (If there's no repetition, the reference
1194        can never match.) The problem arises because, having left the group and
1195        moved on to the rest of the pattern, a later failure that backtracks into
1196        the group uses the captured value from the final iteration of the group
1197        rather than the correct earlier one. I have fixed this in PCRE by forcing
1198        any group that contains a reference to itself to be an atomic group; that
1199        is, there cannot be any backtracking into it once it has completed. This is
1200        similar to recursive and subroutine calls.
1201    
1202    
1203    Version 8.00 19-Oct-09
1204    ----------------------
1205    
1206    1.  The table for translating pcre_compile() error codes into POSIX error codes
1207        was out-of-date, and there was no check on the pcre_compile() error code
1208        being within the table. This could lead to an OK return being given in
1209        error.
1210    
1211    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1212        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1213        in a Windows environment.
1214    
1215    3.  The pcregrep --count option prints the count for each file even when it is
1216        zero, as does GNU grep. However, pcregrep was also printing all files when
1217        --files-with-matches was added. Now, when both options are given, it prints
1218        counts only for those files that have at least one match. (GNU grep just
1219        prints the file name in this circumstance, but including the count seems
1220        more useful - otherwise, why use --count?) Also ensured that the
1221        combination -clh just lists non-zero counts, with no names.
1222    
1223    4.  The long form of the pcregrep -F option was incorrectly implemented as
1224        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1225        but it seems right to fix it, and I didn't think it was worth preserving
1226        the old behaviour.
1227    
1228    5.  The command line items --regex=pattern and --regexp=pattern were not
1229        recognized by pcregrep, which required --regex pattern or --regexp pattern
1230        (with a space rather than an '='). The man page documented the '=' forms,
1231        which are compatible with GNU grep; these now work.
1232    
1233    6.  No libpcreposix.pc file was created for pkg-config; there was just
1234        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1235    
1236    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1237        when UCP support is not needed, by modifying the Python script that
1238        generates it from Unicode data files. This should not matter if the module
1239        is correctly used as a library, but I received one complaint about 50K of
1240        unwanted data. My guess is that the person linked everything into his
1241        program rather than using a library. Anyway, it does no harm.
1242    
1243    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1244        was a minimum greater than 1 for a wide character in a possessive
1245        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1246        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1247        character. Chaos in the form of incorrect output or a compiling loop could
1248        result.
1249    
1250    9.  The restrictions on what a pattern can contain when partial matching is
1251        requested for pcre_exec() have been removed. All patterns can now be
1252        partially matched by this function. In addition, if there are at least two
1253        slots in the offset vector, the offset of the earliest inspected character
1254        for the match and the offset of the end of the subject are set in them when
1255        PCRE_ERROR_PARTIAL is returned.
1256    
1257    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1258        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1259        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1260        and may be more useful for multi-segment matching.
1261    
1262    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1263        used to be given if ever the end of the subject was reached; now it is
1264        given only if matching could not proceed because another character was
1265        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1266        string "Z", which now yields "no match" instead of "partial match". In the
1267        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1268        final character ended with (*FAIL).
1269    
1270    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1271        if the pattern had a "must contain" character that was already found in the
1272        earlier partial match, unless partial matching was again requested. For
1273        example, with the pattern /dog.(body)?/, the "must contain" character is
1274        "g". If the first part-match was for the string "dog", restarting with
1275        "sbody" failed. This bug has been fixed.
1276    
1277    13. The string returned by pcre_dfa_exec() after a partial match has been
1278        changed so that it starts at the first inspected character rather than the
1279        first character of the match. This makes a difference only if the pattern
1280        starts with a lookbehind assertion or \b or \B (\K is not supported by
1281        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1282        matching functions compatible, and I think it's the right thing to do.
1283    
1284    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1285        so that the demonstration program is easily available in environments where
1286        PCRE has not been installed from source.
1287    
1288    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1289        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1290        library.
1291    
1292    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1293        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1294        is not the first non-POSIX option to be added. Clearly some people find
1295        these options useful.
1296    
1297    17. If a caller to the POSIX matching function regexec() passes a non-zero
1298        value for nmatch with a NULL value for pmatch, the value of
1299        nmatch is forced to zero.
1300    
1301    18. RunGrepTest did not have a test for the availability of the -u option of
1302        the diff command, as RunTest does. It now checks in the same way as
1303        RunTest, and also checks for the -b option.
1304    
1305    19. If an odd number of negated classes containing just a single character
1306        interposed, within parentheses, between a forward reference to a named
1307        subpattern and the definition of the subpattern, compilation crashed with
1308        an internal error, complaining that it could not find the referenced
1309        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1310        [The bug was that it was starting one character too far in when skipping
1311        over the character class, thus treating the ] as data rather than
1312        terminating the class. This meant it could skip too much.]
1313    
1314    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1315        /g option in pcretest when the pattern contains \K, which makes it possible
1316        to have an empty string match not at the start, even when the pattern is
1317        anchored. Updated pcretest and pcredemo to use this option.
1318    
1319    21. If the maximum number of capturing subpatterns in a recursion was greater
1320        than the maximum at the outer level, the higher number was returned, but
1321        with unset values at the outer level. The correct (outer level) value is
1322        now given.
1323    
1324    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1325        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1326        make it do so. The string so far is captured, making this feature
1327        compatible with Perl.
1328    
1329    23. The tests have been re-organized, adding tests 11 and 12, to make it
1330        possible to check the Perl 5.10 features against Perl 5.10.
1331    
1332    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1333        pattern matches a fixed length string. PCRE did not allow this; now it
1334        does. Neither allows recursion.
1335    
1336    25. I finally figured out how to implement a request to provide the minimum
1337        length of subject string that was needed in order to match a given pattern.
1338        (It was back references and recursion that I had previously got hung up
1339        on.) This code has now been added to pcre_study(); it finds a lower bound
1340        to the length of subject needed. It is not necessarily the greatest lower
1341        bound, but using it to avoid searching strings that are too short does give
1342        some useful speed-ups. The value is available to calling programs via
1343        pcre_fullinfo().
1344    
1345    26. While implementing 25, I discovered to my embarrassment that pcretest had
1346        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1347        study optimizations had never been tested with that matching function.
1348        Oops. What is worse, even when it was passed study data, there was a bug in
1349        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1350        were also very few tests of studied patterns with pcre_dfa_exec().
1351    
1352    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1353        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1354        on the other side of the coin, they are no longer allowed to have different
1355        names, because these cannot be distinguished in PCRE, and this has caused
1356        confusion. (This is a difference from Perl.)
1357    
1358    28. When duplicate subpattern names are present (necessarily with different
1359        numbers, as required by 27 above), and a test is made by name in a
1360        conditional pattern, either for a subpattern having been matched, or for
1361        recursion in such a pattern, all the associated numbered subpatterns are
1362        tested, and the overall condition is true if the condition is true for any
1363        one of them. This is the way Perl works, and is also more like the way
1364        testing by number works.
1365    
1366    
1367    Version 7.9 11-Apr-09
1368    ---------------------
1369    
1370    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1371        (pcretest), all targets were linked against these libraries. This included
1372        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1373        libraries. This caused unwanted dependencies to be created. This problem
1374        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1375        pcretest is linked with readline.
1376    
1377    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1378        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1379        moved outside it again, because FALSE and TRUE are already defined in AIX,
1380        but BOOL is not.
1381    
1382    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1383        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1384    
1385    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1386        hyphens) following file names and line numbers when outputting matching
1387        lines. This is not true; no spaces are inserted. I have also clarified the
1388        wording for the --colour (or --color) option.
1389    
1390    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1391        was not coloured; this is different to GNU grep, so I have changed it to be
1392        the same.
1393    
1394    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1395        each matching line was coloured. Now it goes on to look for further matches
1396        of any of the test patterns, which is the same behaviour as GNU grep.
1397    
1398    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1399        doesn't make sense to accept an empty string match in pcregrep, so I have
1400        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1401        seems to be how GNU grep behaves.
1402    
1403    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1404        start or after a newline", because the conditional assertion was not being
1405        correctly handled. The rule now is that both the assertion and what follows
1406        in the first alternative must satisfy the test.
1407    
1408    9.  If auto-callout was enabled in a pattern with a conditional group whose
1409        condition was an assertion, PCRE could crash during matching, both with
1410        pcre_exec() and pcre_dfa_exec().
1411    
1412    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1413        used for matching.
1414    
1415    11. Unicode property support in character classes was not working for
1416        characters (bytes) greater than 127 when not in UTF-8 mode.
1417    
1418    12. Added the -M command line option to pcretest.
1419    
1420    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1421    
1422    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1423    
1424    16. Added comments and documentation about mis-use of no_arg in the C++
1425        wrapper.
1426    
1427    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1428        from Martin Jerabek that uses macro names for all relevant character and
1429        string constants.
1430    
1431    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1432        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1433        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1434        these, but not everybody uses configure.
1435    
1436    19. A conditional group that had only one branch was not being correctly
1437        recognized as an item that could match an empty string. This meant that an
1438        enclosing group might also not be so recognized, causing infinite looping
1439        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1440        with the subject "ab", where knowledge that the repeated group can match
1441        nothing is needed in order to break the loop.
1442    
1443    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1444        exec(), but without supplying a callout function, matching went wrong.
1445    
1446    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1447        leak if the size of the offset vector was greater than 30. When the vector
1448        is smaller, the saved offsets during recursion go onto a local stack
1449        vector, but for larger vectors malloc() is used. It was failing to free
1450        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1451        error, in fact).
1452    
1453    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1454        heapframe that is used only when UTF-8 support is enabled. This caused no
1455        problem, but was untidy.
1456    
1457    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1458        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1459        included within another project.
1460    
1461    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1462        slightly modified by me:
1463    
1464          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1465              not building pcregrep.
1466    
1467          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1468              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1469    
1470    25. Forward references, both numeric and by name, in patterns that made use of
1471        duplicate group numbers, could behave incorrectly or give incorrect errors,
1472        because when scanning forward to find the reference group, PCRE was not
1473        taking into account the duplicate group numbers. A pattern such as
1474        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1475    
1476    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1477        the feature of a custom pointer more persuasive (as requested by a user).
1478    
1479    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1480        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1481        pre-defined.
1482    
1483    28. Added support for (*UTF8) at the start of a pattern.
1484    
1485    29. Arrange for flags added by the "release type" setting in CMake to be shown
1486        in the configuration summary.
1487    
1488    
1489    Version 7.8 05-Sep-08
1490    ---------------------
1491    
1492    1.  Replaced UCP searching code with optimized version as implemented for Ad
1493        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1494        stage table and inline lookup instead of a function, giving speed ups of 2
1495        to 5 times on some simple patterns that I tested. Permission was given to
1496        distribute the MultiStage2.py script that generates the tables (it's not in
1497        the tarball, but is in the Subversion repository).
1498    
1499    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1500        scripts.
1501    
1502    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1503        a group with a zero qualifier. The result of the study could be incorrect,
1504        or the function might crash, depending on the pattern.
1505    
1506    4.  Caseless matching was not working for non-ASCII characters in back
1507        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1508        It now works when Unicode Property Support is available.
1509    
1510    5.  In pcretest, an escape such as \x{de} in the data was always generating
1511        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1512        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1513        truncation.
1514    
1515    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1516    
1517    7.  Added two (int) casts to pcregrep when printing the difference of two
1518        pointers, in case they are 64-bit values.
1519    
1520    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1521        test 2 if it fails.
1522    
1523    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1524        and a #define of that name to empty if it is not externally set. This is to
1525        allow users of MSVC to set it if necessary.
1526    
1527    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1528        the convenience functions in the pcre_get.c source file.
1529    
1530    11. An option change at the start of a pattern that had top-level alternatives
1531        could cause overwriting and/or a crash. This command provoked a crash in
1532        some environments:
1533    
1534          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1535    
1536        This potential security problem was recorded as CVE-2008-2371.
1537    
1538    12. For a pattern where the match had to start at the beginning or immediately
1539        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1540        pcre_dfa_exec() could read past the end of the passed subject if there was
1541        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1542        pcretest so that it places the subject at the end of its malloc-ed buffer.
1543    
1544    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1545        exec() might read past the end of the data buffer in UTF-8 mode.
1546    
1547    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1548        the data contained the byte 0x85 as part of a UTF-8 character within its
1549        first line. This applied both to normal and DFA matching.
1550    
1551    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1552        /^[^d]*?$/8 failed to match "abc".
1553    
1554    16. Added a missing copyright notice to pcrecpp_internal.h.
1555    
1556    17. Make it more clear in the documentation that values returned from
1557        pcre_exec() in ovector are byte offsets, not character counts.
1558    
1559    18. Tidied a few places to stop certain compilers from issuing warnings.
1560    
1561    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1562        supplied by Stefan Weber. I made a further small update for 7.8 because
1563        there is a change of source arrangements: the pcre_searchfuncs.c module is
1564        replaced by pcre_ucd.c.
1565    
1566    
1567    Version 7.7 07-May-08
1568    ---------------------
1569    
1570    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1571        a string to a long long, pretend we don't even have a long long." This is
1572        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1573    
1574    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1575        pre-7.6 versions, which defined a global no_arg variable instead of putting
1576        it in the RE class. (See also #8 below.)
1577    
1578    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1579        Lopes.
1580    
1581    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1582    
1583        (1) The include/exclude patterns were being applied to the whole pathnames
1584            of files, instead of just to the final components.
1585    
1586        (2) If there was more than one level of directory, the subdirectories were
1587            skipped unless they satisfied the include/exclude conditions. This is
1588            inconsistent with GNU grep (and could even be seen as contrary to the
1589            pcregrep specification - which I improved to make it absolutely clear).
1590            The action now is always to scan all levels of directory, and just
1591            apply the include/exclude patterns to regular files.
1592    
1593    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1594        --exclude_dir in the tests to avoid scanning .svn directories.
1595    
1596    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1597        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1598        doesn't support NULs in patterns.
1599    
1600    7.  Added some missing "const"s to declarations of static tables in
1601        pcre_compile.c and pcre_dfa_exec.c.
1602    
1603    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1604        caused by fix #2  above. (Subsequently also a second patch to fix the
1605        first patch. And a third patch - this was a messy problem.)
1606    
1607    9.  Applied Craig's patch to remove the use of push_back().
1608    
1609    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1610        matching function regexec().
1611    
1612    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1613        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1614        references. PCRE supports relative numbers with this syntax (I don't think
1615        Oniguruma does).
1616    
1617    12. Previously, a group with a zero repeat such as (...){0} was completely
1618        omitted from the compiled regex. However, this means that if the group
1619        was called as a subroutine from elsewhere in the pattern, things went wrong
1620        (an internal error was given). Such groups are now left in the compiled
1621        pattern, with a new opcode that causes them to be skipped at execution
1622        time.
1623    
1624    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1625        to the way PCRE behaves:
1626    
1627        (a) A lone ] character is dis-allowed (Perl treats it as data).
1628    
1629        (b) A back reference to an unmatched subpattern matches an empty string
1630            (Perl fails the current match path).
1631    
1632        (c) A data ] in a character class must be notated as \] because if the
1633            first data character in a class is ], it defines an empty class. (In
1634            Perl it is not possible to have an empty class.) The empty class []
1635            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1636            The negative empty class [^] matches any one character, independently
1637            of the DOTALL setting.
1638    
1639    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1640        non-existent subpattern following a character class starting with ']' and
1641        containing () gave an internal compiling error instead of "reference to
1642        non-existent subpattern". Fortunately, when the pattern did exist, the
1643        compiled code was correct. (When scanning forwards to check for the
1644        existencd of the subpattern, it was treating the data ']' as terminating
1645        the class, so got the count wrong. When actually compiling, the reference
1646        was subsequently set up correctly.)
1647    
1648    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1649        it was being rejected as not supported by pcre_dfa_exec(), even though
1650        other assertions are supported. I have made pcre_dfa_exec() support
1651        (*FAIL).
1652    
1653    16. The implementation of 13c above involved the invention of a new opcode,
1654        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1655        cannot be changed at match time, I realized I could make a small
1656        improvement to matching performance by compiling OP_ALLANY instead of
1657        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1658        on the OP_ANY path.
1659    
1660    17. Compiling pcretest on Windows with readline support failed without the
1661        following two fixes: (1) Make the unistd.h include conditional on
1662        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1663    
1664    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1665        ncurses library to be included for pcretest when ReadLine support is
1666        requested, but also to allow for it to be overridden. This patch came from
1667        Daniel Bergström.
1668    
1669    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1670        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1671        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1672        spotting this.
1673    
1674    
1675    Version 7.6 28-Jan-08
1676    ---------------------
1677    
1678    1.  A character class containing a very large number of characters with
1679        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1680        overflow.
1681    
1682    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1683        HAVE_LONG_LONG is not defined.
1684    
1685    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1686        bring them up to date and include new features. This patch includes:
1687    
1688        - Fixed PH's badly added libz and libbz2 support.
1689        - Fixed a problem with static linking.
1690        - Added pcredemo. [But later removed - see 7 below.]
1691        - Fixed dftables problem and added an option.
1692        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1693            HAVE_LONG_LONG.
1694        - Added readline support for pcretest.
1695        - Added an listing of the option settings after cmake has run.
1696    
1697    4.  A user submitted a patch to Makefile that makes it easy to create
1698        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1699        Makefile.am that cause it to include this special target, without
1700        affecting anything else. Note that the same mingw target plus all
1701        the other distribution libraries and programs are now supported
1702        when configuring with CMake (see 6 below) instead of with
1703        Configure/Make.
1704    
1705    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1706        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1707        exported in the Windows port". It has not yet been confirmed that the patch
1708        solves the problem, but it does no harm.
1709    
1710    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1711        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1712        with CMake, and also correct the comment about stack recursion.
1713    
1714    7.  Remove the automatic building of pcredemo from the ./configure system and
1715        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1716        of a program that users should build themselves after PCRE is installed, so
1717        building it automatically is not really right. What is more, it gave
1718        trouble in some build environments.
1719    
1720    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1721    
1722    
1723    Version 7.5 10-Jan-08
1724    ---------------------
1725    
1726    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1727        values in parens when parsing an RE using the C++ wrapper."
1728    
1729    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1730        Characters greater than 255 were excluded from the class instead of being
1731        included.
1732    
1733    3.  The same bug as (2) above applied to negated POSIX classes such as
1734        [:^space:].
1735    
1736    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1737        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1738        I have changed it.
1739    
1740    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1741        first named subpattern) and a construct such as (?&a) would reference the
1742        first named subpattern whose name started with "a" (in other words, the
1743        length check was missing). Both these problems are fixed. "Subpattern name
1744        expected" is now given for (?&) (a zero-length name), and this patch also
1745        makes it give the same error for \k'' (previously it complained that that
1746        was a reference to a non-existent subpattern).
1747    
1748    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1749        this is right because (?- can be followed by option settings as well as by
1750        digits. I have, however, made the messages clearer.
1751    
1752    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1753        than the number used in the conditional) now cause a compile-time error.
1754        This is actually not compatible with Perl, which accepts such patterns, but
1755        treats the conditional as always being FALSE (as PCRE used to), but it
1756        seems to me that giving a diagnostic is better.
1757    
1758    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1759        and messages.
1760    
1761    9.  Fix two occurrences of "backslash" in comments that should have been
1762        "backspace".
1763    
1764    10. Remove two redundant lines of code that can never be obeyed (their function
1765        was moved elsewhere).
1766    
1767    11. The program that makes PCRE's Unicode character property table had a bug
1768        which caused it to generate incorrect table entries for sequences of
1769        characters that have the same character type, but are in different scripts.
1770        It amalgamated them into a single range, with the script of the first of
1771        them. In other words, some characters were in the wrong script. There were
1772        thirteen such cases, affecting characters in the following ranges:
1773    
1774          U+002b0 - U+002c1
1775          U+0060c - U+0060d
1776          U+0061e - U+00612
1777          U+0064b - U+0065e
1778          U+0074d - U+0076d
1779          U+01800 - U+01805
1780          U+01d00 - U+01d77
1781          U+01d9b - U+01dbf
1782          U+0200b - U+0200f
1783          U+030fc - U+030fe
1784          U+03260 - U+0327f
1785          U+0fb46 - U+0fbb1
1786          U+10450 - U+1049d
1787    
1788    12. The -o option (show only the matching part of a line) for pcregrep was not
1789        compatible with GNU grep in that, if there was more than one match in a
1790        line, it showed only the first of them. It now behaves in the same way as
1791        GNU grep.
1792    
1793    13. If the -o and -v options were combined for pcregrep, it printed a blank
1794        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1795        does the same. The return code can be used to tell if there were any
1796        non-matching lines.
1797    
1798    14. Added --file-offsets and --line-offsets to pcregrep.
1799    
1800    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1801        infinitely looping recursion. The bug was that positive lookaheads were not
1802        being skipped when checking for a possible empty match (negative lookaheads
1803        and both kinds of lookbehind were skipped).
1804    
1805    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1806        inclusion of <windows.h> to before rather than after the definition of
1807        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1808    
1809    17. Specifying a possessive quantifier with a specific limit for a Unicode
1810        character property caused pcre_compile() to compile bad code, which led at
1811        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1812        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1813        caused the error; without that there was no problem.
1814    
1815    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1816    
1817    19. Added --enable-pcretest-libreadline.
1818    
1819    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1820        RE::GlobalReplace(). As a result, the number of replacements returned was
1821        double what it should be. I removed one of the increments, but Craig sent a
1822        later patch that removed the other one (the right fix) and added unit tests
1823        that check the return values (which was not done before).
1824    
1825    21. Several CMake things:
1826    
1827        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1828            the names libpcre and libpcreposix, not just pcre and pcreposix.
1829    
1830        (2) The above change means that pcretest and pcregrep are now correctly
1831            linked with the newly-built libraries, not previously installed ones.
1832    
1833        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1834    
1835    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1836        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1837        UTF-8 newline character). The key issue is that the pattern starts .*;
1838        this means that the match must be either at the beginning, or after a
1839        newline. The bug was in the code for advancing after a failed match and
1840        checking that the new position followed a newline. It was not taking
1841        account of UTF-8 characters correctly.
1842    
1843    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1844        character classes. PCRE was not treating the sequence [:...:] as a
1845        character class unless the ... were all letters. Perl, however, seems to
1846        allow any characters between [: and :], though of course it rejects as
1847        unknown any "names" that contain non-letters, because all the known class
1848        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1849        for example, whereas PCRE did not - it did not recognize a POSIX character
1850        class. This seemed a bit dangerous, so the code has been changed to be
1851        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1852        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1853        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1854        Perl does, and where it didn't before.
1855    
1856    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1857        Windows environments %n is disabled by default.
1858    
1859    
1860    Version 7.4 21-Sep-07
1861    ---------------------
1862    
1863    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1864        means that a class such as [\s] counted as "explicit reference to CR or
1865        LF". That isn't really right - the whole point of the change was to try to
1866        help when there was an actual mention of one of the two characters. So now
1867        the change happens only if \r or \n (or a literal CR or LF) character is
1868        encountered.
1869    
1870    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1871        of both had grown to the point where there were only 3 bits left.
1872        Fortunately, there was spare space in the data structure, and so I have
1873        moved the internal flags into a new 16-bit field to free up more option
1874        bits.
1875    
1876    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1877        but did not set the internal JCHANGED flag - either of these is enough to
1878        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1879        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1880        start sets both bits.
1881    
1882    4.  Added options (at build time, compile time, exec time) to change \R from
1883        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1884    
1885    5.  doc/pcresyntax.html was missing from the distribution.
1886    
1887    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1888        compatibility, even though it is no longer used.
1889    
1890    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1891        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1892        windows.h file is present (where different names are used). [This was
1893        reversed later after testing - see 16 below.]
1894    
1895    8.  Changed all #include <config.h> to #include "config.h". There were also
1896        some further <pcre.h> cases that I changed to "pcre.h".
1897    
1898    9.  When pcregrep was used with the --colour option, it missed the line ending
1899        sequence off the lines that it output.
1900    
1901    10. It was pointed out to me that arrays of string pointers cause lots of
1902        relocations when a shared library is dynamically loaded. A technique of
1903        using a single long string with a table of offsets can drastically reduce
1904        these. I have refactored PCRE in four places to do this. The result is
1905        dramatic:
1906    
1907          Originally:                          290
1908          After changing UCP table:            187
1909          After changing error message table:   43
1910          After changing table of "verbs"       36
1911          After changing table of Posix names   22
1912    
1913        Thanks to the folks working on Gregex for glib for this insight.
1914    
1915    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1916        unicode-properties was also set.
1917    
1918    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1919    
1920    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1921        checked only for CRLF.
1922    
1923    14. Added casts to pcretest.c to avoid compiler warnings.
1924    
1925    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1926    
1927    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1928        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1929        entirely. This removes changes made in 7 above.
1930    
1931    17. The CMake files have been updated, and there is now more information about
1932        building with CMake in the NON-UNIX-USE document.
1933    
1934    
1935    Version 7.3 28-Aug-07
1936    ---------------------
1937    
1938     1. In the rejigging of the build system that eventually resulted in 7.1, the
1939        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1940        brackets there is not right, since it causes compilers to look for an
1941        installed pcre.h, not the version that is in the source that is being
1942        compiled (which of course may be different). I have changed it back to:
1943    
1944          #include "pcre.h"
1945    
1946        I have a vague recollection that the change was concerned with compiling in
1947        different directories, but in the new build system, that is taken care of
1948        by the VPATH setting the Makefile.
1949    
1950     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1951        when the subject happened to end in the byte 0x85 (e.g. if the last
1952        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1953        characters but of course it shouldn't be taken as a newline when it is part
1954        of another character. The bug was that, for an unlimited repeat of . in
1955        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1956        characters when looking for a newline.
1957    
1958     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1959    
1960     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1961        in debug output.
1962    
1963     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1964        long printing in the pcrecpp unittest when running under MinGW.
1965    
1966     6. ESC_K was left out of the EBCDIC table.
1967    
1968     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1969        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1970        limit also applies to "virtual nesting" when a pattern is recursive, and in
1971        this case 1000 isn't so big. I have been able to remove this limit at the
1972        expense of backing off one optimization in certain circumstances. Normally,
1973        when pcre_exec() would call its internal match() function recursively and
1974        immediately return the result unconditionally, it uses a "tail recursion"
1975        feature to save stack. However, when a subpattern that can match an empty
1976        string has an unlimited repetition quantifier, it no longer makes this
1977        optimization. That gives it a stack frame in which to save the data for
1978        checking that an empty string has been matched. Previously this was taken
1979        from the 1000-entry workspace that had been reserved. So now there is no
1980        explicit limit, but more stack is used.
1981    
1982     8. Applied Daniel's patches to solve problems with the import/export magic
1983        syntax that is required for Windows, and which was going wrong for the
1984        pcreposix and pcrecpp parts of the library. These were overlooked when this
1985        problem was solved for the main library.
1986    
1987     9. There were some crude static tests to avoid integer overflow when computing
1988        the size of patterns that contain repeated groups with explicit upper
1989        limits. As the maximum quantifier is 65535, the maximum group length was
1990        set at 30,000 so that the product of these two numbers did not overflow a
1991        32-bit integer. However, it turns out that people want to use groups that
1992        are longer than 30,000 bytes (though not repeat them that many times).
1993        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1994        made it possible to implement the integer overflow checks in a much more
1995        dynamic way, which I have now done. The artificial limitation on group
1996        length has been removed - we now have only the limit on the total length of
1997        the compiled pattern, which depends on the LINK_SIZE setting.
1998    
1999    10. Fixed a bug in the documentation for get/copy named substring when
2000        duplicate names are permitted. If none of the named substrings are set, the
2001        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2002        empty string.
2003    
2004    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2005        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2006        because the ] is interpreted as the first data character and the
2007        terminating ] is not found. PCRE has been made compatible with Perl in this
2008        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2009        cause memory overwriting.
2010    
2011    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2012        string has been matched (to stop an infinite loop). It was not recognizing
2013        a conditional subpattern that could match an empty string if that
2014        subpattern was within another subpattern. For example, it looped when
2015        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2016        condition was not nested. This bug has been fixed.
2017    
2018    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2019        past the start of the subject in the presence of bytes with the top bit
2020        set, for example "\x8aBCD".
2021    
2022    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2023        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2024    
2025    14. Optimized (?!) to (*FAIL).
2026    
2027    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2028        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2029        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2030        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2031        does: it's just the validity check that is more restrictive.
2032    
2033    16. Inserted checks for integer overflows during escape sequence (backslash)
2034        processing, and also fixed erroneous offset values for syntax errors during
2035        backslash processing.
2036    
2037    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2038        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2039    
2040    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2041        caused an overrun.
2042    
2043    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2044        something other than just ASCII characters) inside a group that had an
2045        unlimited repeat caused a loop at compile time (while checking to see
2046        whether the group could match an empty string).
2047    
2048    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2049        [\P{Any}] did so. (Error in the code for printing property names.)
2050    
2051    21. An orphan \E inside a character class could cause a crash.
2052    
2053    22. A repeated capturing bracket such as (A)? could cause a wild memory
2054        reference during compilation.
2055    
2056    23. There are several functions in pcre_compile() that scan along a compiled
2057        expression for various reasons (e.g. to see if it's fixed length for look
2058        behind). There were bugs in these functions when a repeated \p or \P was
2059        present in the pattern. These operators have additional parameters compared
2060        with \d, etc, and these were not being taken into account when moving along
2061        the compiled data. Specifically:
2062    
2063        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2064            length.
2065    
2066        (b) An item such as \pL+ within a repeated group could cause crashes or
2067            loops.
2068    
2069        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2070            "reference to non-existent subpattern" error.
2071    
2072        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2073    
2074    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2075        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2076    
2077    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2078    
2079    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2080        character were causing crashes (broken optimization).
2081    
2082    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2083        \p or \P) caused a compile-time loop.
2084    
2085    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2086        break. For example, the unstudied pattern [\r\n]A does not match the string
2087        "\r\nA" because change 7.0/46 below moves the current point on by two
2088        characters after failing to match at the start. However, the pattern \nA
2089        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2090        the same is true. There doesn't seem any very clean way out of this, but
2091        what I have chosen to do makes the common cases work: PCRE now takes note
2092        of whether there can be an explicit match for \r or \n anywhere in the
2093        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2094        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2095        pattern has explicit CR or LF references.
2096    
2097    29. Added (*CR) etc for changing newline setting at start of pattern.
2098    
2099    
2100    Version 7.2 19-Jun-07
2101    ---------------------
2102    
2103     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2104        which is apparently normally available under Windows.
2105    
2106     2. Re-jig the pcregrep tests with different newline settings in an attempt
2107        to make them independent of the local environment's newline setting.
2108    
2109     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2110    
2111     4. Some of the "internals" tests were previously cut out when the link size
2112        was not 2, because the output contained actual offsets. The recent new
2113        "Z" feature of pcretest means that these can be cut out, making the tests
2114        usable with all link sizes.
2115    
2116     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2117        stack recursion. This gives a massive performance boost under BSD, but just
2118        a small improvement under Linux. However, it saves one field in the frame
2119        in all cases.
2120    
2121     6. Added more features from the forthcoming Perl 5.10:
2122    
2123        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2124            recursion call. It refers to the nth most recently opened parentheses.
2125    
2126        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2127            to be opened parentheses.
2128    
2129        (c) Conditions that refer to capturing parentheses can be specified
2130            relatively, for example, (?(-2)... or (?(+3)...
2131    
2132        (d) \K resets the start of the current match so that everything before
2133            is not part of it.
2134    
2135        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2136    
2137        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2138            reference syntax.
2139    
2140        (g) (?| introduces a group in which the numbering of parentheses in each
2141            alternative starts with the same number.
2142    
2143        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2144    
2145     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2146        PCRE_INFO_JCHANGED.
2147    
2148     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2149        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2150        for detecting groups that can match an empty string.
2151    
2152     9. A pattern with a very large number of alternatives (more than several
2153        hundred) was running out of internal workspace during the pre-compile
2154        phase, where pcre_compile() figures out how much memory will be needed. A
2155        bit of new cunning has reduced the workspace needed for groups with
2156        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2157        workspace instead of running out of the 4096 that are available.
2158    
2159    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2160    
2161    11. Applied patch from Google to remove an optimization that didn't quite work.
2162        The report of the bug said:
2163    
2164          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2165          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2166          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2167    
2168    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2169        it matched the wrong number of bytes.
2170    
2171    
2172    Version 7.1 24-Apr-07
2173    ---------------------
2174    
2175     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2176        that is more "standard", making use of automake and other Autotools. There
2177        is some re-arrangement of the files and adjustment of comments consequent
2178        on this.
2179    
2180     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2181        for recursive directory scanning broke on some systems because the files
2182        are not scanned in any specific order and on different systems the order
2183        was different. A call to "sort" has been inserted into RunGrepTest for the
2184        approprate test as a short-term fix. In the longer term there may be an
2185        alternative.
2186    
2187     3. I had an email from Eric Raymond about problems translating some of PCRE's
2188        man pages to HTML (despite the fact that I distribute HTML pages, some
2189        people do their own conversions for various reasons). The problems
2190        concerned the use of low-level troff macros .br and .in. I have therefore
2191        removed all such uses from the man pages (some were redundant, some could
2192        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2193        HTML has been updated to handle .nf/.fi and to complain if it encounters
2194        .br or .in.
2195    
2196     4. Updated comments in configure.ac that get placed in config.h.in and also
2197        arranged for config.h to be included in the distribution, with the name
2198        config.h.generic, for the benefit of those who have to compile without
2199        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2200    
2201     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2202        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2203        makevp.bat for the current PCRE, using the additional files
2204        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2205    
2206     6. A Windows user reported a minor discrepancy with test 2, which turned out
2207        to be caused by a trailing space on an input line that had got lost in his
2208        copy. The trailing space was an accident, so I've just removed it.
2209    
2210     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2211        that is needed.
2212    
2213     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2214        as "const" (a) because they are and (b) because it helps the PHP
2215        maintainers who have recently made a script to detect big data structures
2216        in the php code that should be moved to the .rodata section. I remembered
2217        to update Builducptable as well, so it won't revert if ucptable.h is ever
2218        re-created.
2219    
2220     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2221        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2222        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2223        support is not required. This saves 1.5-2K of code, which is important in
2224        some applications.
2225    
2226        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2227        so as not to refer to the tables, even though these functions will never be
2228        called when UTF-8 support is disabled. Otherwise there are problems with a
2229        shared library.
2230    
2231    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2232    
2233        (a) It was defining its arguments as char * instead of void *.
2234    
2235        (b) It was assuming that all moves were upwards in memory; this was true
2236            a long time ago when I wrote it, but is no longer the case.
2237    
2238        The emulated memove() is provided for those environments that have neither
2239        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2240        is clearly not the case, as these two bugs were recently reported.
2241    
2242    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2243        and Detrail to create the HTML documentation, the .txt form of the man
2244        pages, and it removes trailing spaces from listed files. It also creates
2245        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2246        case, it wraps all the #defines with #ifndefs. This script should be run
2247        before "make dist".
2248    
2249    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2250        with Unicode property support.
2251    
2252        (a) For a maximizing quantifier, if the two different cases of the
2253            character were of different lengths in their UTF-8 codings (there are
2254            some cases like this - I found 11), and the matching function had to
2255            back up over a mixture of the two cases, it incorrectly assumed they
2256            were both the same length.
2257    
2258        (b) When PCRE was configured to use the heap rather than the stack for
2259            recursion during matching, it was not correctly preserving the data for
2260            the other case of a UTF-8 character when checking ahead for a match
2261            while processing a minimizing repeat. If the check also involved
2262            matching a wide character, but failed, corruption could cause an
2263            erroneous result when trying to check for a repeat of the original
2264            character.
2265    
2266    13. Some tidying changes to the testing mechanism:
2267    
2268        (a) The RunTest script now detects the internal link size and whether there
2269            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2270            values substituted by "configure". (The RunGrepTest script already did
2271            this for UTF-8.) The configure.ac script no longer substitutes the
2272            relevant variables.
2273    
2274        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2275            with length and offset values. This means that the output is different
2276            for different internal link sizes. Test 2 is skipped for link sizes
2277            other than 2 because of this, bypassing the problem. Unfortunately,
2278            there was also a test in test 3 (the locale tests) that used /B and
2279            failed for link sizes other than 2. Rather than cut the whole test out,
2280            I have added a new /Z option to pcretest that replaces the length and
2281            offset values with spaces. This is now used to make test 3 independent
2282            of link size. (Test 2 will be tidied up later.)
2283    
2284    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2285        segmentation fault instead of returning the appropriate error message.
2286    
2287    15. In multiline mode when the newline sequence was set to "any", the pattern
2288        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2289        This doesn't seem right; it now treats the CRLF combination as the line
2290        ending, and so does not match in that case. It's only a pattern such as ^$
2291        that would hit this one: something like ^ABC$ would have failed after \r
2292        and then tried again after \r\n.
2293    
2294    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2295        in an attempt to make files that differ only in their line terminators
2296        compare equal. This works on Linux.
2297    
2298    17. Under certain error circumstances pcregrep might try to free random memory
2299        as it exited. This is now fixed, thanks to valgrind.
2300    
2301    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2302        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2303        was because its rules for how to advance for /g after matching an empty
2304        string at the end of a line did not allow for this case. They now check for
2305        it specially.
2306    
2307    20. pcretest is supposed to handle patterns and data of any length, by
2308        extending its buffers when necessary. It was getting this wrong when the
2309        buffer for a data line had to be extended.
2310    
2311    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2312        CRLF as a newline sequence.
2313    
2314    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2315        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2316        I have nevertheless tidied it up.
2317    
2318    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2319    
2320    24. Added a man page for pcre-config.
2321    
2322    
2323    Version 7.0 19-Dec-06
2324    ---------------------
2325    
2326     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2327        moving to gcc 4.1.1.
2328    
2329     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2330        sys/time.h, which is documented as needed for this function. It doesn't
2331        seem to matter on Linux, but it showed up on some releases of OS X.
2332    
2333     3. It seems that there are systems where bytes whose values are greater than
2334        127 match isprint() in the "C" locale. The "C" locale should be the
2335        default when a C program starts up. In most systems, only ASCII printing
2336        characters match isprint(). This difference caused the output from pcretest
2337        to vary, making some of the tests fail. I have changed pcretest so that:
2338    
2339        (a) When it is outputting text in the compiled version of a pattern, bytes
2340            other than 32-126 are always shown as hex escapes.
2341    
2342        (b) When it is outputting text that is a matched part of a subject string,
2343            it does the same, unless a different locale has been set for the match
2344            (using the /L modifier). In this case, it uses isprint() to decide.
2345    
2346     4. Fixed a major bug that caused incorrect computation of the amount of memory
2347        required for a compiled pattern when options that changed within the
2348        pattern affected the logic of the preliminary scan that determines the
2349        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2350        that the computed length was too small. The symptoms of this bug were
2351        either the PCRE error "internal error: code overflow" from pcre_compile(),
2352        or a glibc crash with a message such as "pcretest: free(): invalid next
2353        size (fast)". Examples of patterns that provoked this bug (shown in
2354        pcretest format) are:
2355    
2356          /(?-x: )/x
2357          /(?x)(?-x: \s*#\s*)/
2358          /((?i)[\x{c0}])/8
2359          /(?i:[\x{c0}])/8
2360    
2361        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2362        is now done differently.
2363    
2364     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2365        wrapper classes; (b) implement a new function in the C++ scanner that is
2366        more efficient than the old way of doing things because it avoids levels of
2367        recursion in the regex matching; (c) add a paragraph to the documentation
2368        for the FullMatch() function.
2369    
2370     6. The escape sequence \n was being treated as whatever was defined as
2371        "newline". Not only was this contrary to the documentation, which states
2372        that \n is character 10 (hex 0A), but it also went horribly wrong when
2373        "newline" was defined as CRLF. This has been fixed.
2374    
2375     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2376        was being set to -1 for the "end of line" case (supposedly a value that no
2377        character can have). Though this value is never used (the check for end of
2378        line is "zero bytes in current character"), it caused compiler complaints.
2379        I've changed it to 0xffffffff.
2380    
2381     8. In pcre_version.c, the version string was being built by a sequence of
2382        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2383        string (as it is for production releases) called a macro with an empty
2384        argument. The C standard says the result of this is undefined. The gcc
2385        compiler treats it as an empty string (which was what was wanted) but it is
2386        reported that Visual C gives an error. The source has been hacked around to
2387        avoid this problem.
2388    
2389     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2390        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2391        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2392        of them did).
2393    
2394    10. Originally, pcretest opened its input and output without "b"; then I was
2395        told that "b" was needed in some environments, so it was added for release
2396        5.0 to both the input and output. (It makes no difference on Unix-like
2397        systems.) Later I was told that it is wrong for the input on Windows. I've
2398        now abstracted the modes into two macros, to make it easier to fiddle with
2399        them, and removed "b" from the input mode under Windows.
2400    
2401    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2402    
2403    12. Added -help and --help to pcretest as an official way of being reminded
2404        of the options.
2405    
2406    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2407        and pcrecpp.cc because they annoy compilers at high warning levels.
2408    
2409    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2410    
2411    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2412        scripts are not C programs :-) and which was not noticed because it works
2413        on Linux.
2414    
2415    16. pcretest is supposed to handle any length of pattern and data line (as one
2416        line or as a continued sequence of lines) by extending its input buffer if
2417        necessary. This feature was broken for very long pattern lines, leading to
2418        a string of junk being passed to pcre_compile() if the pattern was longer
2419        than about 50K.
2420    
2421    17. I have done a major re-factoring of the way pcre_compile() computes the
2422        amount of memory needed for a compiled pattern. Previously, there was code
2423        that made a preliminary scan of the pattern in order to do this. That was
2424        OK when PCRE was new, but as the facilities have expanded, it has become
2425        harder and harder to keep it in step with the real compile phase, and there
2426        have been a number of bugs (see for example, 4 above). I have now found a
2427        cunning way of running the real compile function in a "fake" mode that
2428        enables it to compute how much memory it would need, while actually only
2429        ever using a few hundred bytes of working memory and without too many
2430        tests of the mode. This should make future maintenance and development
2431        easier. A side effect of this work is that the limit of 200 on the nesting
2432        depth of parentheses has been removed (though this was never a serious
2433        limitation, I suspect). However, there is a downside: pcre_compile() now
2434        runs more slowly than before (30% or more, depending on the pattern). I
2435        hope this isn't a big issue. There is no effect on runtime performance.
2436    
2437    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2438        newline (only possible for the last line of a file) and it was a
2439        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2440    
2441    19. Added additional timing features to pcretest. (1) The -tm option now times
2442        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2443        separate command line item, by a number that specifies the number of
2444        repeats to use when timing. The default is 50000; this gives better
2445        precision, but takes uncomfortably long for very large patterns.
2446    
2447    20. Extended pcre_study() to be more clever in cases where a branch of a
2448        subpattern has no definite first character. For example, (a*|b*)[cd] would
2449        previously give no result from pcre_study(). Now it recognizes that the
2450        first character must be a, b, c, or d.
2451    
2452    21. There was an incorrect error "recursive call could loop indefinitely" if
2453        a subpattern (or the entire pattern) that was being tested for matching an
2454        empty string contained only one non-empty item after a nested subpattern.
2455        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2456        incorrectly, because the \d was being skipped in the check.
2457    
2458    22. The pcretest program now has a new pattern option /B and a command line
2459        option -b, which is equivalent to adding /B to every pattern. This causes
2460        it to show the compiled bytecode, without the additional information that
2461        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2462        is the same as /B/I).
2463    
2464    23. A new optimization is now able automatically to treat some sequences such
2465        as a*b as a*+b. More specifically, if something simple (such as a character
2466        or a simple class like \d) has an unlimited quantifier, and is followed by
2467        something that cannot possibly match the quantified thing, the quantifier
2468        is automatically "possessified".
2469    
2470    24. A recursive reference to a subpattern whose number was greater than 39
2471        went wrong under certain circumstances in UTF-8 mode. This bug could also
2472        have affected the operation of pcre_study().
2473    
2474    25. Realized that a little bit of performance could be had by replacing
2475        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2476    
2477    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2478    
2479    27. Possessive quantifiers such as a++ were previously implemented by turning
2480        them into atomic groups such as ($>a+). Now they have their own opcodes,
2481        which improves performance. This includes the automatically created ones
2482        from 23 above.
2483    
2484    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2485        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2486        the first matched character to be a colon. This applied both to named and
2487        numbered groups.
2488    
2489    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2490    
2491    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2492        building PCRE on Windows easier, so I have included it in the distribution.
2493    
2494    31. There is now a check in pcretest against a ridiculously large number being
2495        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2496        loop, the loop is abandoned.
2497    
2498    32. Forward references to subpatterns in conditions such as (?(2)...) where
2499        subpattern 2 is defined later cause pcre_compile() to search forwards in
2500        the pattern for the relevant set of parentheses. This search went wrong
2501        when there were unescaped parentheses in a character class, parentheses
2502        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2503    
2504    33. "Subroutine" calls and backreferences were previously restricted to
2505        referencing subpatterns earlier in the regex. This restriction has now
2506        been removed.
2507    
2508    34. Added a number of extra features that are going to be in Perl 5.10. On the
2509        whole, these are just syntactic alternatives for features that PCRE had
2510        previously implemented using the Python syntax or my own invention. The
2511        other formats are all retained for compatibility.
2512    
2513        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2514            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2515            also .NET compatible.
2516    
2517        (b) A recursion or subroutine call to a named group can now be defined as
2518            (?&name) as well as (?P>name).
2519    
2520        (c) A backreference to a named group can now be defined as \k<name> or
2521            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2522            5.10, are also .NET compatible.
2523    
2524        (d) A conditional reference to a named group can now use the syntax
2525            (?(<name>) or (?('name') as well as (?(name).
2526    
2527        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2528            groups (named and numbered) that are never evaluated inline, but can be
2529            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2530            is always false. There may be only one alternative in such a group.
2531    
2532        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2533            as the simple (?(R). The condition is true only if the most recent
2534            recursion is that of the given number or name. It does not search out
2535            through the entire recursion stack.
2536    
2537        (g) The escape \gN or \g{N} has been added, where N is a positive or
2538            negative number, specifying an absolute or relative reference.
2539    
2540    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2541        some "unreachable code" warnings.
2542    
2543    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2544        things, this adds five new scripts.
2545    
2546    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2547        There were also incompatibilities regarding the handling of \Q..\E inside
2548        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2549        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2550    
2551    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2552        matches an empty string, and forcibly breaks the loop. There were bugs in
2553        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2554        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2555        separate and independent bugs (that affected different cases) have been
2556        fixed.
2557    
2558    39. Refactored the code to abolish the use of different opcodes for small
2559        capturing bracket numbers. This is a tidy that I avoided doing when I
2560        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2561        The new approach is not only tidier, it makes it possible to reduce the
2562        memory needed to fix the previous bug (38).
2563    
2564    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2565        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2566        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2567        mode.
2568    
2569    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2570        report.
2571    
2572    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2573        copy construction and assignment in the C++ wrapper.
2574    
2575    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2576        couple of bugs that could have given wrong results in the "--newline=crlf"
2577        case.
2578    
2579    44. Added a number of casts and did some reorganization of signed/unsigned int
2580        variables following suggestions from Dair Grant. Also renamed the variable
2581        "this" as "item" because it is a C++ keyword.
2582    
2583    45. Arranged for dftables to add
2584    
2585          #include "pcre_internal.h"
2586    
2587        to pcre_chartables.c because without it, gcc 4.x may remove the array
2588        definition from the final binary if PCRE is built into a static library and
2589        dead code stripping is activated.
2590    
2591    46. For an unanchored pattern, if a match attempt fails at the start of a
2592        newline sequence, and the newline setting is CRLF or ANY, and the next two
2593        characters are CRLF, advance by two characters instead of one.
2594    
2595    
2596    Version 6.7 04-Jul-06
2597    ---------------------
2598    
2599     1. In order to handle tests when input lines are enormously long, pcretest has
2600        been re-factored so that it automatically extends its buffers when
2601        necessary. The code is crude, but this _is_ just a test program. The
2602        default size has been increased from 32K to 50K.
2603    
2604     2. The code in pcre_study() was using the value of the re argument before
2605        testing it for NULL. (Of course, in any sensible call of the function, it
2606        won't be NULL.)
2607    
2608     3. The memmove() emulation function in pcre_internal.h, which is used on
2609        systems that lack both memmove() and bcopy() - that is, hardly ever -
2610        was missing a "static" storage class specifier.
2611    
2612     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2613        containing an extended class (one that cannot be represented by a bitmap
2614        because it contains high-valued characters or Unicode property items, e.g.
2615        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2616        pattern, but PCRE should not loop if you do not (it no longer does).
2617        [Detail: two cases were found: (a) a repeated subpattern containing an
2618        extended class; (b) a recursive reference to a subpattern that followed a
2619        previous extended class. It wasn't skipping over the extended class
2620        correctly when UTF-8 mode was not set.]
2621    
2622     5. A negated single-character class was not being recognized as fixed-length
2623        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2624        compile error "lookbehind assertion is not fixed length".
2625    
2626     6. The RunPerlTest auxiliary script was showing an unexpected difference
2627        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2628        write a Perl script that can interpret lines of an input file either as
2629        byte characters or as UTF-8, which is what "perltest" was being required to
2630        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2631        can't do is switch easily at run time between having the "use utf8;" pragma
2632        or not. In the end, I fudged it by using the RunPerlTest script to insert
2633        "use utf8;" explicitly for the UTF-8 tests.
2634    
2635     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2636        the end of the subject string, contrary to the documentation and to what
2637        Perl does. This was true of both matching functions. Now it matches only at
2638        the start of the subject and immediately after *internal* newlines.
2639    
2640     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2641        a pointer to an int instead of a pointer to an unsigned long int. This
2642        caused problems on 64-bit systems.
2643    
2644     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2645        instance of the 'standard' template library not being so standard".
2646    
2647    10. There was no check on the number of named subpatterns nor the maximum
2648        length of a subpattern name. The product of these values is used to compute
2649        the size of the memory block for a compiled pattern. By supplying a very
2650        long subpattern name and a large number of named subpatterns, the size
2651        computation could be caused to overflow. This is now prevented by limiting
2652        the length of names to 32 characters, and the number of named subpatterns
2653        to 10,000.
2654    
2655    11. Subpatterns that are repeated with specific counts have to be replicated in
2656        the compiled pattern. The size of memory for this was computed from the
2657        length of the subpattern and the repeat count. The latter is limited to
2658        65535, but there was no limit on the former, meaning that integer overflow
2659        could in principle occur. The compiled length of a repeated subpattern is
2660        now limited to 30,000 bytes in order to prevent this.
2661    
2662    12. Added the optional facility to have named substrings with the same name.
2663    
2664    13. Added the ability to use a named substring as a condition, using the
2665        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2666        are numbers (not recommended). Forward references are permitted.
2667    
2668    14. Added forward references in named backreferences (if you see what I mean).
2669    
2670    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2671        pattern could run off the end of the subject. For example, the pattern
2672        "(?s)(.{1,5})"8 did this with the subject "ab".
2673    
2674    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2675        PCRE_CASELESS was set when matching characters that were quantified with ?
2676        or *.
2677    
2678    17. A character class other than a single negated character that had a minimum
2679        but no maximum quantifier - for example [ab]{6,} - was not handled
2680        correctly by pce_dfa_exec(). It would match only one character.
2681    
2682    18. A valid (though odd) pattern that looked like a POSIX character
2683        class but used an invalid character after [ (for example [[,abc,]]) caused
2684        pcre_compile() to give the error "Failed: internal error: code overflow" or
2685        in some cases to crash with a glibc free() error. This could even happen if
2686        the pattern terminated after [[ but there just happened to be a sequence of
2687        letters, a binary zero, and a closing ] in the memory that followed.
2688    
2689    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2690        over the years. Originally (before any Unicode support), just the bottom 8
2691        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2692        output from "man perlunicode" includes this:
2693    
2694          The regular expression compiler produces polymorphic opcodes.  That
2695          is, the pattern adapts to the data and automatically switches to
2696          the Unicode character scheme when presented with Unicode data--or
2697          instead uses a traditional byte scheme when presented with byte
2698          data.
2699    
2700        Sadly, a wide octal escape does not cause a switch, and in a string with
2701        no other multibyte characters, these octal escapes are treated as before.
2702        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2703        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2704        Unicode string.
2705    
2706        I have not perpetrated such confusion in PCRE. Up till now, it took just
2707        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2708        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2709        translate to the appropriate multibyte character.
2710    
2711    29. Applied some refactoring to reduce the number of warnings from Microsoft
2712        and Borland compilers. This has included removing the fudge introduced
2713        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2714        a warning about an unused variable.
2715    
2716    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2717        characters since release 4.0, because Perl (from release 5.004) does not.
2718        [Or at least, is documented not to: some releases seem to be in conflict
2719        with the documentation.] However, when a pattern was studied with
2720        pcre_study() and all its branches started with \s, PCRE still included VT
2721        as a possible starting character. Of course, this did no harm; it just
2722        caused an unnecessary match attempt.
2723    
2724    22. Removed a now-redundant internal flag bit that recorded the fact that case
2725        dependency changed within the pattern. This was once needed for "required
2726        byte" processing, but is no longer used. This recovers a now-scarce options
2727        bit. Also moved the least significant internal flag bit to the most-
2728        significant bit of the word, which was not previously used (hangover from
2729        the days when it was an int rather than a uint) to free up another bit for
2730        the future.
2731    
2732    23. Added support for CRLF line endings as well as CR and LF. As well as the
2733        default being selectable at build time, it can now be changed at runtime
2734        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2735        specify that it is scanning data with non-default line endings.
2736    
2737    24. Changed the definition of CXXLINK to make it agree with the definition of
2738        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2739    
2740    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2741        recursions. This makes a big different to stack usage for some patterns.
2742    
2743    26. If a subpattern containing a named recursion or subroutine reference such
2744        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2745        the space required for the compiled pattern went wrong and gave too small a
2746        value. Depending on the environment, this could lead to "Failed: internal
2747        error: code overflow at offset 49" or "glibc detected double free or
2748        corruption" errors.
2749    
2750    27. Applied patches from Google (a) to support the new newline modes and (b) to
2751        advance over multibyte UTF-8 characters in GlobalReplace.
2752    
2753    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2754        difference for some implementation of PCRE in some Windows version.
2755    
2756    29. Added some extra testing facilities to pcretest:
2757    
2758        \q<number>   in a data line sets the "match limit" value
2759        \Q<number>   in a data line sets the "match recursion limt" value
2760        -S <number>  sets the stack size, where <number> is in megabytes
2761    
2762        The -S option isn't available for Windows.
2763    
2764    
2765    Version 6.6 06-Feb-06
2766    ---------------------
2767    
2768     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2769        in pcreposix.h. I have copied the definition from pcre.h.
2770    
2771     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2772        because pcre.h is no longer a built file.
2773    
2774     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2775        not normally included in the compiled code.
2776    
2777    
2778    Version 6.5 01-Feb-06
2779    ---------------------
2780    
2781     1. When using the partial match feature with pcre_dfa_exec(), it was not
2782        anchoring the second and subsequent partial matches at the new starting
2783        point. This could lead to incorrect results. For example, with the pattern
2784        /1234/, partially matching against "123" and then "a4" gave a match.
2785    
2786     2. Changes to pcregrep:
2787    
2788        (a) All non-match returns from pcre_exec() were being treated as failures
2789            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2790            error message is output. Some extra information is given for the
2791            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2792            probably the only errors that are likely to be caused by users (by
2793            specifying a regex that has nested indefinite repeats, for instance).
2794            If there are more than 20 of these errors, pcregrep is abandoned.
2795    
2796        (b) A binary zero was treated as data while matching, but terminated the
2797            output line if it was written out. This has been fixed: binary zeroes
2798            are now no different to any other data bytes.
2799    
2800        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2801            used to set a locale for matching. The --locale=xxxx long option has
2802            been added (no short equivalent) to specify a locale explicitly on the
2803            pcregrep command, overriding the environment variables.
2804    
2805        (d) When -B was used with -n, some line numbers in the output were one less
2806            than they should have been.
2807    
2808        (e) Added the -o (--only-matching) option.
2809    
2810        (f) If -A or -C was used with -c (count only), some lines of context were
2811            accidentally printed for the final match.
2812    
2813        (g) Added the -H (--with-filename) option.
2814    
2815        (h) The combination of options -rh failed to suppress file names for files
2816            that were found from directory arguments.
2817    
2818        (i) Added the -D (--devices) and -d (--directories) options.
2819    
2820        (j) Added the -F (--fixed-strings) option.
2821    
2822        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2823    
2824        (l) Added the --colo(u)r option.
2825    
2826        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2827            is not present by default.
2828    
2829     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2830        items such as (?R) or (?1), when the recursion could match a number of
2831        alternatives. If it matched one of the alternatives, but subsequently,
2832        outside the recursion, there was a failure, the code tried to back up into
2833        the recursion. However, because of the way PCRE is implemented, this is not
2834        possible, and the result was an incorrect result from the match.
2835    
2836        In order to prevent this happening, the specification of recursion has
2837        been changed so that all such subpatterns are automatically treated as
2838        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2839    
2840     4. I had overlooked the fact that, in some locales, there are characters for
2841        which isalpha() is true but neither isupper() nor islower() are true. In
2842        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2843        and ordfeminine) are like this. This affected the treatment of \w and \W
2844        when they appeared in character classes, but not when they appeared outside
2845        a character class. The bit map for "word" characters is now created
2846        separately from the results of isalnum() instead of just taking it from the
2847        upper, lower, and digit maps. (Plus the underscore character, of course.)
2848    
2849     5. The above bug also affected the handling of POSIX character classes such as
2850        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2851        permanent tables. Instead, the bit maps for such a class were previously
2852        created as the appropriate unions of the upper, lower, and digit bitmaps.
2853        Now they are created by subtraction from the [[:word:]] class, which has
2854        its own bitmap.
2855    
2856     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2857        It is created by subtracting the vertical space characters (\x09, \x0a,
2858        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2859        subtraction was done in the overall bitmap for a character class, meaning
2860        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2861        be recognized. This bug has been fixed.
2862    
2863     7. Patches from the folks at Google:
2864    
2865          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2866          real life, but is still worth protecting against".
2867    
2868          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2869          regular expressions".
2870    
2871          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2872          have it.
2873    
2874          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2875          "configure" and the latter not, in order to fix a problem somebody had
2876          with compiling the Arg class on HP-UX.
2877    
2878          (e) Improve the error-handling of the C++ wrapper a little bit.
2879    
2880          (f) New tests for checking recursion limiting.
2881    
2882     8. The pcre_memmove() function, which is used only if the environment does not
2883        have a standard memmove() function (and is therefore rarely compiled),
2884        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2885        returning a result (though PCRE never actually uses the result).
2886    
2887     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2888        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2889        returned instead of calling malloc() with an overflowing number that would
2890        most likely cause subsequent chaos.
2891    
2892    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2893    
2894    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2895        with this option is matched, the nmatch and pmatch options of regexec() are
2896        ignored.
2897    
2898    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2899        provided in case anyone wants to the the POSIX interface with UTF-8
2900        strings.
2901    
2902    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2903        C++ linking (needed for some HP-UX environments).
2904    
2905    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2906        (unused parameter) and in the pcre_printint() function (omitted "default"
2907        switch label when the default is to do nothing).
2908    
2909    15. Added some code to make it possible, when PCRE is compiled as a C++
2910        library, to replace subject pointers for pcre_exec() with a smart pointer
2911        class, thus making it possible to process discontinuous strings.
2912    
2913    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2914        much the same function. They were added by different people who were trying
2915        to make PCRE easy to compile on non-Unix systems. It has been suggested
2916        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2917        for compiling on Windows systems. I have therefore replaced it with
2918        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2919        defaults to "extern" for C or "extern C" for C++, which works fine on
2920        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2921        SCOPE with something explicit in config.h. In addition:
2922    
2923        (a) pcreposix.h still had just "extern" instead of either of these macros;
2924            I have replaced it with PCRE_DATA_SCOPE.
2925    
2926        (b) Functions such as _pcre_xclass(), which are internal to the library,
2927            but external in the C sense, all had PCRE_EXPORT in their definitions.
2928            This is apparently wrong for the Windows case, so I have removed it.
2929            (It makes no difference on Unix-like systems.)
2930    
2931    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2932        of recursive calls to match(). This is different to MATCH_LIMIT because
2933        that limits the total number of calls to match(), not all of which increase
2934        the depth of recursion. Limiting the recursion depth limits the amount of
2935        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2936        when PCRE is compiled, and changed at run time. A patch from Google adds
2937        this functionality to the C++ interface.
2938    
2939    18. Changes to the handling of Unicode character properties:
2940    
2941        (a) Updated the table to Unicode 4.1.0.
2942    
2943        (b) Recognize characters that are not in the table as "Cn" (undefined).
2944    
2945        (c) I revised the way the table is implemented to a much improved format
2946            which includes recognition of ranges. It now supports the ranges that
2947            are defined in UnicodeData.txt, and it also amalgamates other
2948            characters into ranges. This has reduced the number of entries in the
2949            table from around 16,000 to around 3,000, thus reducing its size
2950            considerably. I realized I did not need to use a tree structure after
2951            all - a binary chop search is just as efficient. Having reduced the
2952            number of entries, I extended their size from 6 bytes to 8 bytes to
2953            allow for more data.
2954    
2955        (d) Added support for Unicode script names via properties such as \p{Han}.
2956    
2957    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2958        matching that character.
2959    
2960    20. When matching a repeated Unicode property with a minimum greater than zero,
2961        (for example \pL{2,}), PCRE could look past the end of the subject if it
2962        reached it while seeking the minimum number of characters. This could
2963        happen only if some of the characters were more than one byte long, because
2964        there is a check for at least the minimum number of bytes.
2965    
2966    21. Refactored the implementation of \p and \P so as to be more general, to
2967        allow for more different types of property in future. This has changed the
2968        compiled form incompatibly. Anybody with saved compiled patterns that use
2969        \p or \P will have to recompile them.
2970    
2971    22. Added "Any" and "L&" to the supported property types.
2972    
2973    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2974        but give a compile time error if the value is greater than 0xff.
2975    
2976    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2977        accidentally not being installed or uninstalled.
2978    
2979    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2980        made were to insert the current release number. This seemed silly, because
2981        it made things harder for people building PCRE on systems that don't run
2982        "configure". I have turned pcre.h into a distributed file, no longer built
2983        by "configure", with the version identification directly included. There is
2984        no longer a pcre.h.in file.
2985    
2986        However, this change necessitated a change to the pcre-config script as
2987        well. It is built from pcre-config.in, and one of the substitutions was the
2988        release number. I have updated configure.ac so that ./configure now finds
2989        the release number by grepping pcre.h.
2990    
2991    26. Added the ability to run the tests under valgrind.
2992    
2993    
2994    Version 6.4 05-Sep-05
2995    ---------------------
2996    
2997     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2998        "--" to be printed when multiple files were scanned, even when none of the
2999        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3000        consider it to be a bug, and have restored the previous behaviour.
3001    
3002     2. A couple of code tidies to get rid of compiler warnings.
3003    
3004     3. The pcretest program used to cheat by referring to symbols in the library
3005        whose names begin with _pcre_. These are internal symbols that are not
3006        really supposed to be visible externally, and in some environments it is
3007        possible to suppress them. The cheating is now confined to including
3008        certain files from the library's source, which is a bit cleaner.
3009    
3010     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3011        file's purpose clearer.
3012    
3013     5. Reorganized pcre_ucp_findchar().
3014    
3015    
3016    Version 6.3 15-Aug-05
3017    ---------------------
3018    
3019     1. The file libpcre.pc.in did not have general read permission in the tarball.
3020    
3021     2. There were some problems when building without C++ support:
3022    
3023        (a) If C++ support was not built, "make install" and "make test" still
3024            tried to test it.
3025    
3026        (b) There were problems when the value of CXX was explicitly set. Some
3027            changes have been made to try to fix these, and ...
3028    
3029        (c) --disable-cpp can now be used to explicitly disable C++ support.
3030    
3031        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3032            backslash in a target when C++ was disabled. This confuses some
3033            versions of "make", apparently. Using an intermediate variable solves
3034            this. (Same for CPP_LOBJ.)
3035    
3036     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3037        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3038        necessary on certain architectures.
3039    
3040     4. Added a setting of -export-symbols-regex to the link command to remove
3041        those symbols that are exported in the C sense, but actually are local
3042        within the library, and not documented. Their names all begin with
3043        "_pcre_". This is not a perfect job, because (a) we have to except some
3044        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3045        available (and never for static libraries). I have made a note to try to
3046        find a way round (a) in the future.
3047    
3048    
3049    Version 6.2 01-Aug-05
3050    ---------------------
3051    
3052     1. There was no test for integer overflow of quantifier values. A construction
3053        such as {1111111111111111} would give undefined results. What is worse, if
3054        a minimum quantifier for a parenthesized subpattern overflowed and became
3055        negative, the calculation of the memory size went wrong. This could have
3056        led to memory overwriting.
3057    
3058     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3059    
3060     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3061        operating environments where this matters.
3062    
3063     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3064        PCRE options from within the C++ wrapper.
3065    
3066     5. Named capturing subpatterns were not being correctly counted when a pattern
3067        was compiled. This caused two problems: (a) If there were more than 100
3068        such subpatterns, the calculation of the memory needed for the whole
3069        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3070        back references of the form \12, where the number was greater than 9, were
3071        not recognized as back references, even though there were sufficient
3072        previous subpatterns.
3073    
3074     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3075        versions of gcc, e.g. 2.95.4.
3076    
3077    
3078    Version 6.1 21-Jun-05
3079    ---------------------
3080    
3081     1. There was one reference to the variable "posix" in pcretest.c that was not
3082        surrounded by "#if !defined NOPOSIX".
3083    
3084     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3085        the cross-check on the old pcre_info() function, for the benefit of the
3086        cut-down version of PCRE that is currently imported into Exim.
3087    
3088     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3089        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3090        patterns that start (?i) but I don't think that matters. The use of (?i) is
3091        just an example; this all applies to the other options as well.
3092    
3093     4. Since libtool seems to echo the compile commands it is issuing, the output
3094        from "make" can be reduced a bit by putting "@" in front of each libtool
3095        compile command.
3096    
3097     5. Patch from the folks at Google for configure.in to be a bit more thorough
3098        in checking for a suitable C++ installation before trying to compile the
3099        C++ stuff. This should fix a reported problem when a compiler was present,
3100        but no suitable headers.
3101    
3102     6. The man pages all had just "PCRE" as their title. I have changed them to
3103        be the relevant file name. I have also arranged that these names are
3104        retained in the file doc/pcre.txt, which is a concatenation in text format
3105        of all the man pages except the little individual ones for each function.
3106    
3107     7. The NON-UNIX-USE file had not been updated for the different set of source
3108        files that come with release 6. I also added a few comments about the C++
3109        wrapper.
3110    
3111    
3112    Version 6.0 07-Jun-05
3113    ---------------------
3114    
3115     1. Some minor internal re-organization to help with my DFA experiments.
3116    
3117     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3118        didn't matter for the library itself when fully configured, but did matter
3119        when compiling without UCP support, or within Exim, where the ucp files are
3120        not imported.
3121    
3122     3. Refactoring of the library code to split up the various functions into
3123        different source modules. The addition of the new DFA matching code (see
3124        below) to a single monolithic source would have made it really too
3125        unwieldy, quite apart from causing all the code to be include in a
3126        statically linked application, when only some functions are used. This is
3127        relevant even without the DFA addition now that patterns can be compiled in
3128        one application and matched in another.
3129    
3130        The downside of splitting up is that there have to be some external
3131        functions and data tables that are used internally in different modules of
3132        the library but which are not part of the API. These have all had their
3133        names changed to start with "_pcre_" so that they are unlikely to clash
3134        with other external names.
3135    
3136     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3137        a different (DFA) algorithm. Although it is slower than the original
3138        function, it does have some advantages for certain types of matching
3139        problem.
3140    
3141     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3142        including restarting after a partial match.
3143    
3144     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3145        defined when compiling for Windows was sent to me. I have put it into the
3146        code, though I have no means of testing or verifying it.
3147    
3148     7. Added the pcre_refcount() auxiliary function.
3149    
3150     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3151        match before or at the first newline in the subject string. In pcretest,
3152        the /f option on a pattern can be used to set this.
3153    
3154     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3155        would behave wrongly. This has been present in PCRE since release 4.0.
3156    
3157    10. A number of changes to the pcregrep command:
3158    
3159        (a) Refactored how -x works; insert ^(...)$ instead of setting
3160            PCRE_ANCHORED and checking the length, in preparation for adding
3161            something similar for -w.
3162    
3163        (b) Added the -w (match as a word) option.
3164    
3165        (c) Refactored the way lines are read and buffered so as to have more
3166            than one at a time available.
3167    
3168        (d) Implemented a pcregrep test script.
3169    
3170        (e) Added the -M (multiline match) option. This allows patterns to match
3171            over several lines of the subject. The buffering ensures that at least
3172            8K, or the rest of the document (whichever is the shorter) is available
3173            for matching (and similarly the previous 8K for lookbehind assertions).
3174    
3175        (f) Changed the --help output so that it now says
3176    
3177              -w, --word-regex(p)
3178    
3179            instead of two lines, one with "regex" and the other with "regexp"
3180            because that confused at least one person since the short forms are the
3181            same. (This required a bit of code, as the output is generated
3182            automatically from a table. It wasn't just a text change.)
3183    
3184        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3185            option but starts with a hyphen. Could be a pattern or a path name
3186            starting with a hyphen, for instance.
3187    
3188        (h) "-" can be given as a file name to represent stdin.
3189    
3190        (i) When file names are being printed, "(standard input)" is used for
3191            the standard input, for compatibility with GNU grep. Previously
3192            "<stdin>" was used.
3193    
3194        (j) The option --label=xxx can be used to supply a name to be used for
3195            stdin when file names are being printed. There is no short form.
3196    
3197        (k) Re-factored the options decoding logic because we are going to add
3198            two more options that take data. Such options can now be given in four
3199            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3200    
3201        (l) Added the -A, -B, and -C options for requesting that lines of context
3202            around matches be printed.
3203    
3204        (m) Added the -L option to print the names of files that do not contain
3205            any matching lines, that is, the complement of -l.
3206    
3207        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3208            continue to scan other files.
3209    
3210        (o) The -s option was incorrectly implemented. For compatibility with other
3211            greps, it now suppresses the error message for a non-existent or non-
3212            accessible file (but not the return code). There is a new option called
3213            -q that suppresses the output of matching lines, which was what -s was
3214            previously doing.
3215    
3216        (p) Added --include and --exclude options to specify files for inclusion
3217            and exclusion when recursing.
3218    
3219    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3220        Hopefully, it now does.
3221    
3222    12. Missing cast in pcre_study().
3223    
3224    13. Added an "uninstall" target to the makefile.
3225    
3226    14. Replaced "extern" in the function prototypes in Makefile.in with
3227        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3228        world, but is set differently for Windows.
3229    
3230    15. Added a second compiling function called pcre_compile2(). The only
3231        difference is that it has an extra argument, which is a pointer to an
3232        integer error code. When there is a compile-time failure, this is set
3233        non-zero, in addition to the error test pointer being set to point to an
3234        error message. The new argument may be NULL if no error number is required
3235        (but then you may as well call pcre_compile(), which is now just a
3236        wrapper). This facility is provided because some applications need a
3237        numeric error indication, but it has also enabled me to tidy up the way
3238        compile-time errors are handled in the POSIX wrapper.
3239    
3240    16. Added VPATH=.libs to the makefile; this should help when building with one
3241        prefix path and installing with another. (Or so I'm told by someone who
3242        knows more about this stuff than I do.)
3243    
3244    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3245        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3246        match everything, including newlines. This is not POSIX-compatible, but
3247        somebody wanted the feature. From pcretest it can be activated by using
3248        both the P and the s flags.
3249    
3250    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3251    
3252    19. libpcre.pc was being incorrectly installed as executable.
3253    
3254    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3255        it now also looks for '\r' so that it will work unmodified on Windows.
3256    
3257    21. Added Google's contributed C++ wrapper to the distribution.
3258    
3259    22. Added some untidy missing memory free() calls in pcretest, to keep
3260        Electric Fence happy when testing.
3261    
3262    
3263    
3264    Version 5.0 13-Sep-04
3265    ---------------------
3266    
3267     1. Internal change: literal characters are no longer packed up into items
3268        containing multiple characters in a single byte-string. Each character
3269        is now matched using a separate opcode. However, there may be more than one
3270        byte in the character in UTF-8 mode.
3271    
3272     2. The pcre_callout_block structure has two new fields: pattern_position and
3273        next_item_length. These contain the offset in the pattern to the next match
3274        item, and its length, respectively.
3275    
3276     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3277        insertion of callouts before each pattern item. Added the /C option to
3278        pcretest to make use of this.
3279    
3280     4. On the advice of a Windows user, the lines
3281    
3282          #if defined(_WIN32) || defined(WIN32)
3283          _setmode( _fileno( stdout ), 0x8000 );
3284          #endif  /* defined(_WIN32) || defined(WIN32) */
3285    
3286        have been added to the source of pcretest. This apparently does useful
3287        magic in relation to line terminators.
3288    
3289     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3290        for the benefit of those environments where the "b" makes a difference.
3291    
3292     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3293        to know about it. I have put a hack into configure.in that adds in code
3294        to set GCC=yes if CC=icc. This seems to end up at a point in the
3295        generated configure script that is early enough to affect the setting of
3296        compiler options, which is what is needed, but I have no means of testing
3297        whether it really works. (The user who reported this had patched the
3298        generated configure script, which of course I cannot do.)
3299    
3300        LATER: After change 22 below (new libtool files), the configure script
3301        seems to know about icc (and also ecc). Therefore, I have commented out
3302        this hack in configure.in.
3303    
3304     7. Added support for pkg-config (2 patches were sent in).
3305    
3306     8. Negated POSIX character classes that used a combination of internal tables
3307        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3308        [[:^ascii]]. Typically, they would match almost any characters. The other
3309        POSIX classes were not broken in this way.
3310    
3311     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3312        to find the match, as PCRE was deluded into thinking that the match had to
3313        start at the start point or following a newline. The same bug applied to
3314        patterns with negative forward assertions or any backward assertions
3315        preceding ".*" at the start, unless the pattern required a fixed first
3316        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3317    
3318    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3319        starting at the last subject character, bytes beyond the end of the subject
3320        string were read.
3321    
3322    11. Renamed the variable "class" as "classbits" to make life easier for C++
3323        users. (Previously there was a macro definition, but it apparently wasn't
3324        enough.)
3325    
3326    12. Added the new field "tables" to the extra data so that tables can be passed
3327        in at exec time, or the internal tables can be re-selected. This allows
3328        a compiled regex to be saved and re-used at a later time by a different
3329        program that might have everything at different addresses.
3330    
3331    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3332        -R library as well as a -L library.
3333    
3334    14. The debugging options of pcretest (-d on the command line or D on a
3335        pattern) showed incorrect output for anything following an extended class
3336        that contained multibyte characters and which was followed by a quantifier.
3337    
3338    15. Added optional support for general category Unicode character properties
3339        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3340        support. It adds about 90K to the size of the library. The meanings of the
3341        inbuilt class escapes such as \d and \s have NOT been changed.
3342    
3343    16. Updated pcredemo.c to include calls to free() to release the memory for the
3344        compiled pattern.
3345    
3346    17. The generated file chartables.c was being created in the source directory
3347        instead of in the building directory. This caused the build to fail if the
3348        source directory was different from the building directory, and was
3349        read-only.
3350    
3351    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3352        file. No doubt somebody will tell me if they don't make sense... Also added
3353        Dan Mooney's comments about building on OpenVMS.
3354    
3355    19. Added support for partial matching via the PCRE_PARTIAL option for
3356        pcre_exec() and the \P data escape in pcretest.
3357    
3358    20. Extended pcretest with 3 new pattern features:
3359    
3360        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3361              write the compiled pattern to the file whose name is "rest-of-line".
3362              This is a straight binary dump of the data, with the saved pointer to
3363              the character tables forced to be NULL. The study data, if any, is
3364              written too. After writing, pcretest reads a new pattern.
3365    
3366        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3367              compiled pattern from the given file. There must not be any
3368              occurrences of "<" in the file name (pretty unlikely); if there are,
3369              pcretest will instead treat the initial "<" as a pattern delimiter.
3370              After reading in the pattern, pcretest goes on to read data lines as
3371              usual.
3372    
3373        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3374              and 16-bit fields in a compiled pattern, to simulate a pattern that
3375              was compiled on a host of opposite endianness.
3376    
3377    21. The pcre-exec() function can now cope with patterns that were compiled on
3378        hosts of opposite endianness, with this restriction:
3379    
3380          As for any compiled expression that is saved and used later, the tables
3381          pointer field cannot be preserved; the extra_data field in the arguments
3382          to pcre_exec() should be used to pass in a tables address if a value
3383          other than the default internal tables were used at compile time.
3384    
3385    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3386        now diagnosed as an error. Previously, most of the time, a negative number
3387        would have been treated as zero, but if in addition "ovector" was passed as
3388        NULL, a crash could occur.
3389    
3390    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3391        new versions from the libtool 1.5 distribution (the last one is a copy of
3392        a file called libtool.m4). This seems to have fixed the need to patch
3393        "configure" to support Darwin 1.3 (which I used to do). However, I still
3394        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3395        workstation).
3396    
3397    24. Changed the PCRE licence to be the more standard "BSD" licence.
3398    
3399    
3400    Version 4.5 01-Dec-03
3401    ---------------------
3402    
3403     1. There has been some re-arrangement of the code for the match() function so
3404        that it can be compiled in a version that does not call itself recursively.
3405        Instead, it keeps those local variables that need separate instances for
3406        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3407        needs to "recurse". Keeping track of where control must go is done by means
3408        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3409        hide most of the details from the main code, and operates only if
3410        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3411        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3412        operating.
3413    
3414        To make it easier for callers to provide specially tailored get/free
3415        functions for this usage, two new functions, pcre_stack_malloc, and
3416        pcre_stack_free, are used. They are always called in strict stacking order,
3417        and the size of block requested is always the same.
3418    
3419        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3420        PCRE has been compiled to use the stack or the heap for recursion. The
3421        -C option of pcretest uses this to show which version is compiled.
3422    
3423        A new data escape \S, is added to pcretest; it causes the amounts of store
3424        obtained and freed by both kinds of malloc/free at match time to be added
3425        to the output.
3426    
3427     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3428        what's available on my current Linux desktop machine.
3429    
3430     3. When matching a UTF-8 string, the test for a valid string at the start has
3431        been extended. If start_offset is not zero, PCRE now checks that it points
3432        to a byte that is the start of a UTF-8 character. If not, it returns
3433        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3434        this is necessary because there may be backward assertions in the pattern.
3435        When matching the same subject several times, it may save resources to use
3436        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3437    
3438     4. The code for checking the validity of UTF-8 strings has been tightened so
3439        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3440        containing "overlong sequences".
3441    
3442     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3443        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3444        should have been "&", but it just so happened that all the cases this let
3445        through by mistake were picked up later in the function.
3446    
3447     6. I had used a variable called "isblank" - this is a C99 function, causing
3448        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3449    
3450     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3451        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3452        so the version is shown; (c) stop "make test" echoing "./RunTest".
3453    
3454     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3455    
3456     9. The prototype for memmove() for systems that don't have it was using
3457        size_t, but the inclusion of the header that defines size_t was later. I've
3458        moved the #includes for the C headers earlier to avoid this.
3459    
3460    10. Added some adjustments to the code to make it easier to compiler on certain
3461        special systems:
3462    
3463          (a) Some "const" qualifiers were missing.
3464          (b) Added the macro EXPORT before all exported functions; by default this
3465              is defined to be empty.
3466          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3467              that it reads its output file name as an argument instead of writing
3468              to the standard output and assuming this can be redirected.
3469    
3470    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3471        class containing characters with values greater than 255, PCRE compilation
3472        went into a loop.
3473    
3474    12. A recursive reference to a subpattern that was within another subpattern
3475        that had a minimum quantifier of zero caused PCRE to crash. For example,
3476        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3477        recursion. If the recursively-called subpattern itself had a zero repeat,
3478        that was OK.
3479    
3480    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3481        buffer into which it was copied (for escape processing) was still set at
3482        1024, so long lines caused crashes.
3483    
3484    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3485        "internal error: code overflow...". This applied to any character class
3486        that was followed by a possessive quantifier.
3487    
3488    15. Modified the Makefile to add libpcre.la as a prerequisite for
3489        libpcreposix.la because I was told this is needed for a parallel build to
3490        work.
3491    
3492    16. If a pattern that contained .* following optional items at the start was
3493        studied, the wrong optimizing data was generated, leading to matching
3494        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3495        matching string must start with a or b or c. The correct conclusion for
3496        this pattern is that a match can start with any character.
3497    
3498    
3499    Version 4.4 13-Aug-03
3500    ---------------------
3501    
3502     1. In UTF-8 mode, a character class containing characters with values between
3503        127 and 255 was not handled correctly if the compiled pattern was studied.
3504        In fixing this, I have also improved the studying algorithm for such
3505        classes (slightly).
3506    
3507     2. Three internal functions had redundant arguments passed to them. Removal
3508        might give a very teeny performance improvement.
3509    
3510     3. Documentation bug: the value of the capture_top field in a callout is *one
3511        more than* the number of the hightest numbered captured substring.
3512    
3513     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3514        in incorrectly linking with a previously installed version. They now link
3515        explicitly with libpcre.la.
3516    
3517     5. configure.in no longer needs to recognize Cygwin specially.
3518    
3519     6. A problem in pcre.in for Windows platforms is fixed.
3520    
3521     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3522        pcretest, it used to include the size of the study block as part of its
3523        output. Unfortunately, the structure contains a field that has a different
3524        size on different hardware architectures. This meant that the tests that
3525        showed this size failed. As the block is currently always of a fixed size,
3526        this information isn't actually particularly useful in pcretest output, so
3527        I have just removed it.
3528    
3529     8. Three pre-processor statements accidentally did not start in column 1.
3530        Sadly, there are *still* compilers around that complain, even though
3531        standard C has not required this for well over a decade. Sigh.
3532    
3533     9. In pcretest, the code for checking callouts passed small integers in the
3534        callout_data field, which is a void * field. However, some picky compilers
3535        complained about the casts involved for this on 64-bit systems. Now
3536        pcretest passes the address of the small integer instead, which should get
3537        rid of the warnings.
3538    
3539    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3540        both compile and run time, and gives an error if an invalid UTF-8 sequence
3541        is found. There is a option for disabling this check in cases where the
3542        string is known to be correct and/or the maximum performance is wanted.
3543    
3544    11. In response to a bug report, I changed one line in Makefile.in from
3545    
3546            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3547        to
3548            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3549    
3550        to look similar to other lines, but I have no way of telling whether this
3551        is the right thing to do, as I do not use Windows. No doubt I'll get told
3552        if it's wrong...
3553    
3554    
3555    Version 4.3 21-May-03
3556    ---------------------
3557    
3558    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3559       Makefile.
3560    
3561    2. Some refactoring to improve the quality of the code:
3562    
3563       (i)   The utf8_table... variables are now declared "const".
3564    
3565       (ii)  The code for \cx, which used the "case flipping" table to upper case
3566             lower case letters, now just substracts 32. This is ASCII-specific,
3567             but the whole concept of \cx is ASCII-specific, so it seems
3568             reasonable.
3569    
3570       (iii) PCRE was using its character types table to recognize decimal and
3571             hexadecimal digits in the pattern. This is silly, because it handles
3572             only 0-9, a-f, and A-F, but the character types table is locale-
3573             specific, which means strange things might happen. A private
3574             table is now used for this - though it costs 256 bytes, a table is
3575             much faster than multiple explicit tests. Of course, the standard
3576             character types table is still used for matching digits in subject
3577             strings against \d.
3578    
3579       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3580             ending in _t are). So I've renamed it as ESC_tee.
3581    
3582    3. The first argument for regexec() in the POSIX wrapper should have been
3583       defined as "const".
3584    
3585    4. Changed pcretest to use malloc() for its buffers so that they can be
3586       Electric Fenced for debugging.
3587    
3588    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3589       to read one or more bytes before the start of the subject string. Often this
3590       had no effect on PCRE's behaviour, but in some circumstances it could
3591       provoke a segmentation fault.
3592    
3593    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3594       to try to read one or more bytes before the start of the subject string.
3595    
3596    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3597       UTF-8 support could misbehave in various ways if the subject string
3598       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3599       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3600       back over UTF-8 characters.)
3601    
3602    
3603    Version 4.2 14-Apr-03
3604    ---------------------
3605    
3606    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3607    
3608    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3609         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3610         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3611         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3612         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3613           and BUILD_EXEEXT
3614         Note: automatic setting of the BUILD variables is not yet working
3615         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3616           compile-time but not at link-time
3617         [LINK]: use for linking executables only
3618         make different versions for Windows and non-Windows
3619         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3620           libraries
3621         [LINK_FOR_BUILD]: new variable
3622         [OBJEXT]: use throughout
3623         [EXEEXT]: use throughout
3624         <winshared>: new target
3625         <wininstall>: new target
3626         <dftables.o>: use native compiler
3627         <dftables>: use native linker
3628         <install>: handle Windows platform correctly
3629         <clean>: ditto
3630         <check>: ditto
3631         copy DLL to top builddir before testing
3632    
3633       As part of these changes, -no-undefined was removed again. This was reported
3634       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3635       in any case.
3636    
3637    3. Some tidies to get rid of compiler warnings:
3638    
3639       . In the match_data structure, match_limit was an unsigned long int, whereas
3640         match_call_count was an int. I've made them both unsigned long ints.
3641    
3642       . In pcretest the fact that a const uschar * doesn't automatically cast to
3643         a void * provoked a warning.
3644    
3645       . Turning on some more compiler warnings threw up some "shadow" variables
3646         and a few more missing casts.
3647    
3648    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3649       option, a class that contained a single character with a value between 128
3650       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3651    
3652    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3653       option, a class that contained several characters, but with at least one
3654       whose value was between 128 and 255 caused PCRE to crash.
3655    
3656    
3657    Version 4.1 12-Mar-03
3658    ---------------------
3659    
3660    1. Compiling with gcc -pedantic found a couple of places where casts were
3661    needed, and a string in dftables.c that was longer than standard compilers are
3662    required to support.
3663    
3664    2. Compiling with Sun's compiler found a few more places where the code could
3665    be tidied up in order to avoid warnings.
3666    
3667    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3668    first of these names is deprecated in the latest Autoconf in favour of the name
3669    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3670    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3671    analogy I have changed it to CFLAGS_FOR_BUILD.
3672    
3673    4. Added -no-undefined to the linking command in the Makefile, because this is
3674    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3675    linking step for the pcreposix library.
3676    
3677    5. PCRE was failing to diagnose the case of two named groups with the same
3678    name.
3679    
3680    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3681    literal character that is needed in the subject for a match, and scans along to
3682    ensure that it is present before embarking on the full matching process. This
3683    saves time in cases of nested unlimited repeats that are never going to match.
3684    Problem: the scan can take a lot of time if the subject is very long (e.g.
3685    megabytes), thus penalizing straightforward matches. It is now done only if the
3686    amount of subject to be scanned is less than 1000 bytes.
3687    
3688    7. A lesser problem with the same optimization is that it was recording the
3689    first character of an anchored pattern as "needed", thus provoking a search
3690    right along the subject, even when the first match of the pattern was going to
3691    fail. The "needed" character is now not set for anchored patterns, unless it
3692    follows something in the pattern that is of non-fixed length. Thus, it still
3693    fulfils its original purpose of finding quick non-matches in cases of nested
3694    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3695    
3696    
3697    Version 4.0 17-Feb-03
3698    ---------------------
3699    
3700    1. If a comment in an extended regex that started immediately after a meta-item
3701    extended to the end of string, PCRE compiled incorrect data. This could lead to
3702    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3703    
3704    2. Moved to autoconf 2.53 and libtool 1.4.2.
3705    
3706    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3707    the special perltest8 script is no longer needed - all the tests can be run
3708    from a single perltest script.
3709    
3710    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3711    by \s. It has now been removed in PCRE. This means it isn't recognized as
3712    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3713    class [:space:] *does* include VT, thereby creating a mess.
3714    
3715    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3716    space and tab.
3717    
3718    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3719    its new features into the main test script, reducing the number of scripts.
3720    
3721    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3722    were backward compatible, and made the (?i) apply to the whole pattern, as if
3723    /i were given. Now it behaves more logically, and applies the option setting
3724    only to what follows. PCRE has been changed to follow suit. However, if it
3725    finds options settings right at the start of the pattern, it extracts them into
3726    the global options, as before. Thus, they show up in the info data.
3727    
3728    8. Added support for the \Q...\E escape sequence. Characters in between are
3729    treated as literals. This is slightly different from Perl in that $ and @ are
3730    also handled as literals inside the quotes. In Perl, they will cause variable
3731    interpolation. Note the following examples:
3732    
3733        Pattern            PCRE matches      Perl matches
3734    
3735        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3736        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3737        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3738    
3739    For compatibility with Perl, \Q...\E sequences are recognized inside character
3740    classes as well as outside them.
3741    
3742    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3743    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3744    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3745    signed/unsigned warnings.
3746    
3747    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3748    option for pcretest, so I've replaced it by a simple function that does just
3749    that job.
3750    
3751    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3752    "pcregrep -".
3753    
3754    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3755    Java package. This provides some syntactic sugar for simple cases of what my
3756    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3757    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3758    item, you can use this simplified notation. Note that only makes sense with
3759    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3760    greediness, whatever the setting of the PCRE_UNGREEDY option.
3761    
3762    13. A change of greediness default within a pattern was not taking effect at
3763    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3764    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3765    was abstracted outside.
3766    
3767    14. PCRE now supports the \G assertion. It is true when the current matching
3768    position is at the start point of the match. This differs from \A when the
3769    starting offset is non-zero. Used with the /g option of pcretest (or similar
3770    code), it works in the same way as it does for Perl's /g option. If all
3771    alternatives of a regex begin with \G, the expression is anchored to the start
3772    match position, and the "anchored" flag is set in the compiled expression.
3773    
3774    15. Some bugs concerning the handling of certain option changes within patterns
3775    have been fixed. These applied to options other than (?ims). For example,
3776    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3777    been the other way round. Some of this was related to change 7 above.
3778    
3779    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3780    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3781    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3782    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3783    
3784    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3785    mode. Unlike ".", it always matches newline, whatever the setting of
3786    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3787    assertions. Perl allows it, but it doesn't (in general) work because it can't<