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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.81  
changed lines
  Added in v.994

  ViewVC Help
Powered by ViewVC 1.1.5