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

Diff of /code/trunk/ChangeLog

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

revision 23 by nigel, Sat Feb 24 21:38:41 2007 UTC revision 1003 by ph10, Wed Aug 15 16:45:36 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.32
5    ------------
6    
7    1.  Improved JIT compiler optimizations for first character search and single
8        character iterators.
9    
10    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11        Patch by Daniel Richard G.
12    
13    3.  Single character iterator optimizations in the JIT compiler.
14    
15    4.  Improved JIT compiler optimizations for character ranges.
16    
17    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18        Reported by Giuseppe D'Angelo.
19    
20    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21        of a line, was being set incorrectly in cases where .* appeared inside
22        atomic brackets at the start of a pattern, or where there was a subsequent
23        *PRUNE or *SKIP.
24    
25    7.  Improved instruction cache flush for POWER/PowerPC.
26        Patch by Daniel Richard G.
27    
28    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
29        grep:
30    
31        (a) There is now no limit to the number of patterns to be matched.
32    
33        (b) An error is given if a pattern is too long.
34    
35        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36            are now supported.
37    
38        (d) --exclude-from and --include-from (multiple use) have been added.
39    
40        (e) Exclusions and inclusions now apply to all files and directories, not
41            just to those obtained from scanning a directory recursively.
42    
43        (f) Multiple uses of -f and --file-list are now supported.
44    
45        (g) In a Windows environment, the default for -d has been changed from
46            "read" (the GNU grep default) to "skip", because otherwise the presence
47            of a directory in the file list provokes an error.
48    
49        (h) The documentation has been revised and clarified in places.
50    
51    
52    Version 8.31 06-July-2012
53    -------------------------
54    
55    1.  Fixing a wrong JIT test case and some compiler warnings.
56    
57    2.  Removed a bashism from the RunTest script.
58    
59    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
60        to unsigned type, result still unsigned" that was given by an MS compiler
61        on encountering the code "-sizeof(xxx)".
62    
63    4.  Partial matching support is added to the JIT compiler.
64    
65    5.  Fixed several bugs concerned with partial matching of items that consist
66        of more than one character:
67    
68        (a) /^(..)\1/ did not partially match "aba" because checking references was
69            done on an "all or nothing" basis. This also applied to repeated
70            references.
71    
72        (b) \R did not give a hard partial match if \r was found at the end of the
73            subject.
74    
75        (c) \X did not give a hard partial match after matching one or more
76            characters at the end of the subject.
77    
78        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
79            a partial match for the string "\r".
80    
81        (e) When newline was set to CRLF, the metacharacter "." did not recognize
82            a partial match for a CR character at the end of the subject string.
83    
84    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
85        running pcretest, the text "(JIT)" added to the output whenever JIT is
86        actually used to run the match.
87    
88    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
89        or /S+[+] with a digit between 1 and 7.
90    
91    8.  OP_NOT now supports any UTF character not just single-byte ones.
92    
93    9.  (*MARK) control verb is now supported by the JIT compiler.
94    
95    10. The command "./RunTest list" lists the available tests without actually
96        running any of them. (Because I keep forgetting what they all are.)
97    
98    11. Add PCRE_INFO_MAXLOOKBEHIND.
99    
100    12. Applied a (slightly modified) user-supplied patch that improves performance
101        when the heap is used for recursion (compiled with --disable-stack-for-
102        recursion). Instead of malloc and free for each heap frame each time a
103        logical recursion happens, frames are retained on a chain and re-used where
104        possible. This sometimes gives as much as 30% improvement.
105    
106    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
107        call.
108    
109    14. As documented, (*COMMIT) is now confined to within a positive assertion.
110    
111    15. It is now possible to link pcretest with libedit as an alternative to
112        libreadline.
113    
114    16. (*COMMIT) control verb is now supported by the JIT compiler.
115    
116    17. The Unicode data tables have been updated to Unicode 6.1.0.
117    
118    18. Added --file-list option to pcregrep.
119    
120    19. Added binary file support to pcregrep, including the -a, --binary-files,
121        -I, and --text options.
122    
123    20. The madvise function is renamed for posix_madvise for QNX compatibility
124        reasons. Fixed by Giuseppe D'Angelo.
125    
126    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
127    
128    22. Changed the option for creating symbolic links for 16-bit man pages from
129        -s to -sf so that re-installing does not cause issues.
130    
131    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
132    
133    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
134        matches in certain environments (the workspace was not being correctly
135        retained). Also added to pcre_dfa_exec() a simple plausibility check on
136        some of the workspace data at the beginning of a restart.
137    
138    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
139        was not doing so when it should - probably a typo introduced by SVN 528
140        (change 8.10/14).
141    
142    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
143        \w+ when the character tables indicated that \x{c4} was a word character.
144        There were several related cases, all because the tests for doing a table
145        lookup were testing for characters less than 127 instead of 255.
146    
147    27. If a pattern contains capturing parentheses that are not used in a match,
148        their slots in the ovector are set to -1. For those that are higher than
149        any matched groups, this happens at the end of processing. In the case when
150        there were back references that the ovector was too small to contain
151        (causing temporary malloc'd memory to be used during matching), and the
152        highest capturing number was not used, memory off the end of the ovector
153        was incorrectly being set to -1. (It was using the size of the temporary
154        memory instead of the true size.)
155    
156    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
157        ovector size, it uses memory at the end of the block that it has got.
158    
159    29. Check for an overlong MARK name and give an error at compile time. The
160        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
161    
162    30. JIT compiler update.
163    
164    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
165        Rill for the patch.
166    
167    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
168    
169    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
170    
171    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
172        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
173        was enabled.
174    
175    35. Improve JIT code generation for greedy plus quantifier.
176    
177    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
178        1 to "aa" instead of to an empty string. The bug affected repeated groups
179        that could potentially match an empty string.
180    
181    37. Optimizing single character iterators in JIT.
182    
183    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
184        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
185        codepoints that are too big for the mode are faulted, and in a UTF mode,
186        disallowed codepoints are also faulted.
187    
188    39. If PCRE was compiled with UTF support, in three places in the DFA
189        matcher there was code that should only have been obeyed in UTF mode, but
190        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
191        processing when bytes with values greater than 127 were present. In 16-bit
192        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
193        both cases the values are those that cannot be the first data item in a UTF
194        character. The three items that might have provoked this were recursions,
195        possessively repeated groups, and atomic groups.
196    
197    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
198        and pcregrep, because some OS require shared objects to be explicitly
199        passed to ld, causing the link step to fail if they are not.
200    
201    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
202        mode, patterns that started with \h* or \R* might be incorrectly matched.
203    
204    
205    Version 8.30 04-February-2012
206    -----------------------------
207    
208    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
209        name is defined in ctype.h.
210    
211    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
212        only in quite long subpatterns.
213    
214    3.  Removed the function pcre_info(), which has been obsolete and deprecated
215        since it was replaced by pcre_fullinfo() in February 2000.
216    
217    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
218        match a (*MARK), and the match failed at the start of the subject, a
219        reference to memory before the start of the subject could occur. This bug
220        was introduced by fix 17 of release 8.21.
221    
222    5.  A reference to an unset group with zero minimum repetition was giving
223        totally wrong answers (in non-JavaScript-compatibility mode). For example,
224        /(another)?(\1?)test/ matched against "hello world test". This bug was
225        introduced in release 8.13.
226    
227    6.  Add support for 16-bit character strings (a large amount of work involving
228        many changes and refactorings).
229    
230    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
231        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
232        from a file.
233    
234    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
235        rounding is not applied in this particular case).
236    
237    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
238        if they appear, or are escaped, in patterns.
239    
240    10. Get rid of a number of -Wunused-but-set-variable warnings.
241    
242    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
243        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
244        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
245        also returns the mark "x". This bug applied to capturing parentheses,
246        non-capturing parentheses, and atomic parentheses. It also applied to some
247        assertions.
248    
249    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
250        information out of configure.ac instead of relying on pcre.h.generic, which
251        is not stored in the repository.
252    
253    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
254        -lreadline.
255    
256    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
257    
258    15. Applied Graycode's patch to put the top-level frame on the stack rather
259        than the heap when not using the stack for recursion. This gives a
260        performance improvement in many cases when recursion is not deep.
261    
262    16. Experimental code added to "pcretest -C" to output the stack frame size.
263    
264    
265    Version 8.21 12-Dec-2011
266    ------------------------
267    
268    1.  Updating the JIT compiler.
269    
270    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
271        are added as well.
272    
273    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
274        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
275        calling _pcre_jit_exec. Some extra comments are added.
276    
277    4.  (*MARK) settings inside atomic groups that do not contain any capturing
278        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
279        was introduced by change 18 for 8.20.
280    
281    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
282        ECMA-262 standard.
283    
284    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
285        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
286        This bug was probably introduced by change 9 of 8.13.
287    
288    7.  While fixing 6 above, I noticed that a number of other items were being
289        incorrectly rejected as "not fixed length". This arose partly because newer
290        opcodes had not been added to the fixed-length checking code. I have (a)
291        corrected the bug and added tests for these items, and (b) arranged for an
292        error to occur if an unknown opcode is encountered while checking for fixed
293        length instead of just assuming "not fixed length". The items that were
294        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
295        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
296        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
297    
298    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
299        being incorrectly compiled and would have given unpredicatble results.
300    
301    9.  A possessively repeated subpattern with minimum repeat count greater than
302        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
303        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
304        the first (A) could occur when it should not.
305    
306    10. Add a cast and remove a redundant test from the code.
307    
308    11. JIT should use pcre_malloc/pcre_free for allocation.
309    
310    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
311        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
312        is anything other than /usr, -L is still shown).
313    
314    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
315    
316    14. Perl does not support \N without a following name in a [] class; PCRE now
317        also gives an error.
318    
319    15. If a forward reference was repeated with an upper limit of around 2000,
320        it caused the error "internal error: overran compiling workspace". The
321        maximum number of forward references (including repeats) was limited by the
322        internal workspace, and dependent on the LINK_SIZE. The code has been
323        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
324        the default depends on LINK_SIZE. There is a new upper limit (for safety)
325        of around 200,000 forward references. While doing this, I also speeded up
326        the filling in of repeated forward references.
327    
328    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
329        incorrectly expecting the subject to contain another "a" after the start.
330    
331    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
332        in the match, the SKIP should be ignored. This was not happening; instead
333        the SKIP was being treated as NOMATCH. For patterns such as
334        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
335        tested.
336    
337    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
338        now much more compatible with Perl, in particular in cases where the result
339        is a non-match for a non-anchored pattern. For example, if
340        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
341        "m", where previously it did not return a name. A side effect of this
342        change is that for partial matches, the last encountered mark name is
343        returned, as for non matches. A number of tests that were previously not
344        Perl-compatible have been moved into the Perl-compatible test files. The
345        refactoring has had the pleasing side effect of removing one argument from
346        the match() function, thus reducing its stack requirements.
347    
348    19. If the /S+ option was used in pcretest to study a pattern using JIT,
349        subsequent uses of /S (without +) incorrectly behaved like /S+.
350    
351    21. Retrieve executable code size support for the JIT compiler and fixing
352        some warnings.
353    
354    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
355        not work when the shorter character appeared right at the end of the
356        subject string.
357    
358    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
359        systems.
360    
361    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
362        output it when the /M option is used in pcretest.
363    
364    25. The CheckMan script was not being included in the distribution. Also, added
365        an explicit "perl" to run Perl scripts from the PrepareRelease script
366        because this is reportedly needed in Windows.
367    
368    26. If study data was being save in a file and studying had not found a set of
369        "starts with" bytes for the pattern, the data written to the file (though
370        never used) was taken from uninitialized memory and so caused valgrind to
371        complain.
372    
373    27. Updated RunTest.bat as provided by Sheri Pierce.
374    
375    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
376    
377    29. Computation of memory usage for the table of capturing group names was
378        giving an unnecessarily large value.
379    
380    
381    Version 8.20 21-Oct-2011
382    ------------------------
383    
384    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
385        a POSIX class. After further experiments with Perl, which convinced me that
386        Perl has bugs and confusions, a closing square bracket is no longer allowed
387        in a POSIX name. This bug also affected patterns with classes that started
388        with full stops.
389    
390    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
391        captured substring, but while checking the failing first alternative,
392        substring 1 is temporarily captured. If the output vector supplied to
393        pcre_exec() was not big enough for this capture, the yield of the function
394        was still zero ("insufficient space for captured substrings"). This cannot
395        be totally fixed without adding another stack variable, which seems a lot
396        of expense for a edge case. However, I have improved the situation in cases
397        such as /(a)(b)x|abc/ matched against "abc", where the return code
398        indicates that fewer than the maximum number of slots in the ovector have
399        been set.
400    
401    3.  Related to (2) above: when there are more back references in a pattern than
402        slots in the output vector, pcre_exec() uses temporary memory during
403        matching, and copies in the captures as far as possible afterwards. It was
404        using the entire output vector, but this conflicts with the specification
405        that only 2/3 is used for passing back captured substrings. Now it uses
406        only the first 2/3, for compatibility. This is, of course, another edge
407        case.
408    
409    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
410        main code base, and can be used by building with --enable-jit. When this is
411        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
412        runtime --no-jit option is given.
413    
414    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
415        ovector, the return from the function was zero, implying that there were
416        other matches that did not fit. The correct "exactly full" value is now
417        returned.
418    
419    6.  If a subpattern that was called recursively or as a subroutine contained
420        (*PRUNE) or any other control that caused it to give a non-standard return,
421        invalid errors such as "Error -26 (nested recursion at the same subject
422        position)" or even infinite loops could occur.
423    
424    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
425        computing the minimum length on reaching *ACCEPT, and so ended up with the
426        wrong value of 1 rather than 0. Further investigation indicates that
427        computing a minimum subject length in the presence of *ACCEPT is difficult
428        (think back references, subroutine calls), and so I have changed the code
429        so that no minimum is registered for a pattern that contains *ACCEPT.
430    
431    8.  If (*THEN) was present in the first (true) branch of a conditional group,
432        it was not handled as intended. [But see 16 below.]
433    
434    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
435        Sheri Pierce.
436    
437    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
438        the first byte in a match must be "a".
439    
440    11. Change 17 for 8.13 increased the recursion depth for patterns like
441        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
442        pattern contains any instances of (*THEN). If it does not, the old
443        optimizations are restored. It would be nice to do this on a per-group
444        basis, but at the moment that is not feasible.
445    
446    12. In some environments, the output of pcretest -C is CRLF terminated. This
447        broke RunTest's code that checks for the link size. A single white space
448        character after the value is now allowed for.
449    
450    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
451        For "fr", it uses the Windows-specific input and output files.
452    
453    14. If (*THEN) appeared in a group that was called recursively or as a
454        subroutine, it did not work as intended. [But see next item.]
455    
456    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
457        pattern fragments (but not containing any | characters). If A and B are
458        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
459        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
460        D. In other words, Perl considers parentheses that do not contain any |
461        characters to be part of a surrounding alternative, whereas PCRE was
462        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
463        differently. PCRE now behaves in the same way as Perl, except in the case
464        of subroutine/recursion calls such as (?1) which have in any case always
465        been different (but PCRE had them first :-).
466    
467    16. Related to 15 above: Perl does not treat the | in a conditional group as
468        creating alternatives. Such a group is treated in the same way as an
469        ordinary group without any | characters when processing (*THEN). PCRE has
470        been changed to match Perl's behaviour.
471    
472    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
473        RunGrepTest script failed.
474    
475    18. Change 22 for version 13 caused atomic groups to use more stack. This is
476        inevitable for groups that contain captures, but it can lead to a lot of
477        stack use in large patterns. The old behaviour has been restored for atomic
478        groups that do not contain any capturing parentheses.
479    
480    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
481        suppress the check for a minimum subject length at run time. (If it was
482        given to pcre_exec() or pcre_dfa_exec() it did work.)
483    
484    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
485        fail to work when decoding hex characters in data strings in EBCDIC
486        environments.
487    
488    21. It appears that in at least one Mac OS environment, the isxdigit() function
489        is implemented as a macro that evaluates to its argument more than once,
490        contravening the C 90 Standard (I haven't checked a later standard). There
491        was an instance in pcretest which caused it to go wrong when processing
492        \x{...} escapes in subject strings. The has been rewritten to avoid using
493        things like p++ in the argument of isxdigit().
494    
495    
496    Version 8.13 16-Aug-2011
497    ------------------------
498    
499    1.  The Unicode data tables have been updated to Unicode 6.0.0.
500    
501    2.  Two minor typos in pcre_internal.h have been fixed.
502    
503    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
504        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
505        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
506    
507    4.  There were a number of related bugs in the code for matching backrefences
508        caselessly in UTF-8 mode when codes for the characters concerned were
509        different numbers of bytes. For example, U+023A and U+2C65 are an upper
510        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
511        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
512        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
513        2-byte code at the end of the subject (it thought there wasn't enough data
514        left).
515    
516    5.  Comprehensive information about what went wrong is now returned by
517        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
518        as the output vector has at least 2 elements. The offset of the start of
519        the failing character and a reason code are placed in the vector.
520    
521    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
522        now returned is for the first byte of the failing character, instead of the
523        last byte inspected. This is an incompatible change, but I hope it is small
524        enough not to be a problem. It makes the returned offset consistent with
525        pcre_exec() and pcre_dfa_exec().
526    
527    7.  pcretest now gives a text phrase as well as the error number when
528        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
529        failure, the offset and reason code are output.
530    
531    8.  When \R was used with a maximizing quantifier it failed to skip backwards
532        over a \r\n pair if the subsequent match failed. Instead, it just skipped
533        back over a single character (\n). This seems wrong (because it treated the
534        two characters as a single entity when going forwards), conflicts with the
535        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
536        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
537        has been changed.
538    
539    9.  Some internal refactoring has changed the processing so that the handling
540        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
541        time (the PCRE_DOTALL option was changed this way some time ago: version
542        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
543        which was always a bit of a fudge. It also means that there is one less
544        argument for the match() function, which reduces its stack requirements
545        slightly. This change also fixes an incompatibility with Perl: the pattern
546        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
547    
548    10. More internal refactoring has drastically reduced the number of recursive
549        calls to match() for possessively repeated groups such as (abc)++ when
550        using pcre_exec().
551    
552    11. While implementing 10, a number of bugs in the handling of groups were
553        discovered and fixed:
554    
555        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
556        (a|)*(?1) gave a compile-time internal error.
557        ((a|)+)+  did not notice that the outer group could match an empty string.
558        (^a|^)+   was not marked as anchored.
559        (.*a|.*)+ was not marked as matching at start or after a newline.
560    
561    12. Yet more internal refactoring has removed another argument from the match()
562        function. Special calls to this function are now indicated by setting a
563        value in a variable in the "match data" data block.
564    
565    13. Be more explicit in pcre_study() instead of relying on "default" for
566        opcodes that mean there is no starting character; this means that when new
567        ones are added and accidentally left out of pcre_study(), testing should
568        pick them up.
569    
570    14. The -s option of pcretest has been documented for ages as being an old
571        synonym of -m (show memory usage). I have changed it to mean "force study
572        for every regex", that is, assume /S for every regex. This is similar to -i
573        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
574        using it. It makes it easier to run collections of tests with and without
575        study enabled, and thereby test pcre_study() more easily. All the standard
576        tests are now run with and without -s (but some patterns can be marked as
577        "never study" - see 20 below).
578    
579    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
580        restoration of the capturing data to the outer values was not happening
581        correctly.
582    
583    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
584        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
585        pattern had matched an empty string, and so incorrectly returned a no
586        match.
587    
588    17. There was optimizing code for the last branch of non-capturing parentheses,
589        and also for the obeyed branch of a conditional subexpression, which used
590        tail recursion to cut down on stack usage. Unfortunately, now that there is
591        the possibility of (*THEN) occurring in these branches, tail recursion is
592        no longer possible because the return has to be checked for (*THEN). These
593        two optimizations have therefore been removed. [But see 8.20/11 above.]
594    
595    18. If a pattern containing \R was studied, it was assumed that \R always
596        matched two bytes, thus causing the minimum subject length to be
597        incorrectly computed because \R can also match just one byte.
598    
599    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
600        was incorrectly computed.
601    
602    20. If /S is present twice on a test pattern in pcretest input, it now
603        *disables* studying, thereby overriding the use of -s on the command line
604        (see 14 above). This is necessary for one or two tests to keep the output
605        identical in both cases.
606    
607    21. When (*ACCEPT) was used in an assertion that matched an empty string and
608        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
609    
610    22. When an atomic group that contained a capturing parenthesis was
611        successfully matched, but the branch in which it appeared failed, the
612        capturing was not being forgotten if a higher numbered group was later
613        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
614        group 1 to "a", when in fact it should be unset. This applied to multi-
615        branched capturing and non-capturing groups, repeated or not, and also to
616        positive assertions (capturing in negative assertions does not happen
617        in PCRE) and also to nested atomic groups.
618    
619    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
620        subject after a captured substring, to make it easier to tell which of a
621        number of identical substrings has been captured.
622    
623    24. The way atomic groups are processed by pcre_exec() has been changed so that
624        if they are repeated, backtracking one repetition now resets captured
625        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
626        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
627        "aaa". Previously, it would have been "a". As part of this code
628        refactoring, the way recursive calls are handled has also been changed.
629    
630    25. If an assertion condition captured any substrings, they were not passed
631        back unless some other capturing happened later. For example, if
632        (?(?=(a))a) was matched against "a", no capturing was returned.
633    
634    26. When studying a pattern that contained subroutine calls or assertions,
635        the code for finding the minimum length of a possible match was handling
636        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
637        group 1 called group 2 while simultaneously a separate group 2 called group
638        1). A stack overflow occurred in this case. I have fixed this by limiting
639        the recursion depth to 10.
640    
641    27. Updated RunTest.bat in the distribution to the version supplied by Tom
642        Fortmann. This supports explicit test numbers on the command line, and has
643        argument validation and error reporting.
644    
645    28. An instance of \X with an unlimited repeat could fail if at any point the
646        first character it looked at was a mark character.
647    
648    29. Some minor code refactoring concerning Unicode properties and scripts
649        should reduce the stack requirement of match() slightly.
650    
651    30. Added the '=' option to pcretest to check the setting of unused capturing
652        slots at the end of the pattern, which are documented as being -1, but are
653        not included in the return count.
654    
655    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
656        compiled something random. Now it gives a compile-time error (as does
657        Perl).
658    
659    32. A *MARK encountered during the processing of a positive assertion is now
660        recorded and passed back (compatible with Perl).
661    
662    33. If --only-matching or --colour was set on a pcregrep call whose pattern
663        had alternative anchored branches, the search for a second match in a line
664        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
665        matched the line "0102" twice. The same bug affected patterns that started
666        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
667        twice.
668    
669    34. Previously, PCRE did not allow quantification of assertions. However, Perl
670        does, and because of capturing effects, quantifying parenthesized
671        assertions may at times be useful. Quantifiers are now allowed for
672        parenthesized assertions.
673    
674    35. A minor code tidy in pcre_compile() when checking options for \R usage.
675    
676    36. \g was being checked for fancy things in a character class, when it should
677        just be a literal "g".
678    
679    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
680        appearance of a nested POSIX class supersedes an apparent external class.
681        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
682        unescaped square brackets may also appear as part of class names. For
683        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
684        more like Perl. (But see 8.20/1 above.)
685    
686    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
687        was because it thought it was \N{name}, which is not supported).
688    
689    39. Add minix to OS list not supporting the -S option in pcretest.
690    
691    40. PCRE tries to detect cases of infinite recursion at compile time, but it
692        cannot analyze patterns in sufficient detail to catch mutual recursions
693        such as ((?1))((?2)). There is now a runtime test that gives an error if a
694        subgroup is called recursively as a subpattern for a second time at the
695        same position in the subject string. In previous releases this might have
696        been caught by the recursion limit, or it might have run out of stack.
697    
698    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
699        happen only once. PCRE was, however incorrectly giving a compile time error
700        "recursive call could loop indefinitely" because it cannot analyze the
701        pattern in sufficient detail. The compile time test no longer happens when
702        PCRE is compiling a conditional subpattern, but actual runaway loops are
703        now caught at runtime (see 40 above).
704    
705    42. It seems that Perl allows any characters other than a closing parenthesis
706        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
707        has been changed to be the same.
708    
709    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
710        as not to get warnings when autogen.sh is called. Also changed
711        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
712    
713    44. To help people who use pcregrep to scan files containing exceedingly long
714        lines, the following changes have been made:
715    
716        (a) The default value of the buffer size parameter has been increased from
717            8K to 20K. (The actual buffer used is three times this size.)
718    
719        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
720            PCRE is built.
721    
722        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
723            to be set at run time.
724    
725        (d) Numerical values in pcregrep options can be followed by K or M, for
726            example --buffer-size=50K.
727    
728        (e) If a line being scanned overflows pcregrep's buffer, an error is now
729            given and the return code is set to 2.
730    
731    45. Add a pointer to the latest mark to the callout data block.
732    
733    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
734        partial match of an empty string instead of no match. This was specific to
735        the use of ".".
736    
737    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
738        complete match instead of a partial match. This bug was dependent on both
739        the PCRE_UTF8 and PCRE_DOTALL options being set.
740    
741    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
742        starting byte set, because \b was not being ignored.
743    
744    
745    Version 8.12 15-Jan-2011
746    ------------------------
747    
748    1.  Fixed some typos in the markup of the man pages, and wrote a script that
749        checks for such things as part of the documentation building process.
750    
751    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
752        --match-limit and --recursion-limit options (added for 8.11). In
753        particular, this made one of the standard tests fail. (The integer value
754        went into the wrong half of a long int.)
755    
756    3.  If the --colour option was given to pcregrep with -v (invert match), it
757        did strange things, either producing crazy output, or crashing. It should,
758        of course, ignore a request for colour when reporting lines that do not
759        match.
760    
761    4.  Another pcregrep bug caused similar problems if --colour was specified with
762        -M (multiline) and the pattern match finished with a line ending.
763    
764    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
765        matched in multiline mode, the following line was shown as part of the
766        match. This seems wrong, so I have changed it.
767    
768    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
769        the check for further matches in the same line (so they could be coloured)
770        to overrun the end of the current line. If another match was found, it was
771        incorrectly shown (and then shown again when found in the next line).
772    
773    7.  If pcregrep was compiled under Windows, there was a reference to the
774        function pcregrep_exit() before it was defined. I am assuming this was
775        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
776        reported by a user. I've moved the definition above the reference.
777    
778    
779    Version 8.11 10-Dec-2010
780    ------------------------
781    
782    1.  (*THEN) was not working properly if there were untried alternatives prior
783        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
784        backtracked to try for "b" instead of moving to the next alternative branch
785        at the same level (in this case, to look for "c"). The Perl documentation
786        is clear that when (*THEN) is backtracked onto, it goes to the "next
787        alternative in the innermost enclosing group".
788    
789    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
790        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
791        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
792        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
793        (*THEN).
794    
795    3.  If \s appeared in a character class, it removed the VT character from
796        the class, even if it had been included by some previous item, for example
797        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
798        of \s, but is part of the POSIX "space" class.)
799    
800    4.  A partial match never returns an empty string (because you can always
801        match an empty string at the end of the subject); however the checking for
802        an empty string was starting at the "start of match" point. This has been
803        changed to the "earliest inspected character" point, because the returned
804        data for a partial match starts at this character. This means that, for
805        example, /(?<=abc)def/ gives a partial match for the subject "abc"
806        (previously it gave "no match").
807    
808    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
809        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
810        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
811        has an implication that the given string is incomplete (because a partial
812        match is preferred over a full match). For this reason, these items now
813        give a partial match in this situation. [Aside: previously, the one case
814        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
815        match rather than a full match, which was wrong by the old rules, but is
816        now correct.]
817    
818    6.  There was a bug in the handling of #-introduced comments, recognized when
819        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
820        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
821        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
822        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
823        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
824        places in pcre_compile().
825    
826    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
827        comments when looking ahead for named forward references to subpatterns,
828        the only newline sequence it recognized was NL. It now handles newlines
829        according to the set newline convention.
830    
831    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
832        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
833        cater for a lack of strerror(). These oversights have been fixed.
834    
835    9.  Added --match-limit and --recursion-limit to pcregrep.
836    
837    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
838    
839    11. When the -o option was used, pcregrep was setting a return code of 1, even
840        when matches were found, and --line-buffered was not being honoured.
841    
842    12. Added an optional parentheses number to the -o and --only-matching options
843        of pcregrep.
844    
845    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
846        can match an empty string. The code to do it in pcretest and pcredemo
847        needed fixing:
848    
849        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
850            only one byte after an empty string match just before CRLF (this case
851            just got forgotten; "any" and "anycrlf" were OK).
852    
853        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
854            mode when an empty string match preceded an ASCII character followed by
855            a non-ASCII character. (The code for advancing by one character rather
856            than one byte was nonsense.)
857    
858        (c) The pcredemo.c sample program did not have any code at all to handle
859            the cases when CRLF is a valid newline sequence.
860    
861    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
862        as a starting offset was within the subject string. There is now a new
863        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
864        negative or greater than the length of the string. In order to test this,
865        pcretest is extended to allow the setting of negative starting offsets.
866    
867    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
868        starting offset points to the beginning of a UTF-8 character was
869        unnecessarily clumsy. I tidied it up.
870    
871    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
872        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
873    
874    17. Nobody had reported that the --include_dir option, which was added in
875        release 7.7 should have been called --include-dir (hyphen, not underscore)
876        for compatibility with GNU grep. I have changed it to --include-dir, but
877        left --include_dir as an undocumented synonym, and the same for
878        --exclude-dir, though that is not available in GNU grep, at least as of
879        release 2.5.4.
880    
881    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
882        characters from a string of bytes) have been redefined so as not to use
883        loops, in order to improve performance in some environments. At the same
884        time, I abstracted some of the common code into auxiliary macros to save
885        repetition (this should not affect the compiled code).
886    
887    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
888        compile-time error is now given if \c is not followed by an ASCII
889        character, that is, a byte less than 128. (In EBCDIC mode, the code is
890        different, and any byte value is allowed.)
891    
892    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
893        START_OPTIMIZE option, which is now allowed at compile time - but just
894        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
895        to pcregrep and other applications that have no direct access to PCRE
896        options. The new /Y option in pcretest sets this option when calling
897        pcre_compile().
898    
899    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
900        back references. Groups containing recursive back references were forced to
901        be atomic by that change, but in the case of named groups, the amount of
902        memory required was incorrectly computed, leading to "Failed: internal
903        error: code overflow". This has been fixed.
904    
905    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
906        pcretest.c, to avoid build problems in some Borland environments.
907    
908    
909    Version 8.10 25-Jun-2010
910    ------------------------
911    
912    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
913        THEN.
914    
915    2.  (*ACCEPT) was not working when inside an atomic group.
916    
917    3.  Inside a character class, \B is treated as a literal by default, but
918        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
919        causes the error). The code is unchanged, but I tidied the documentation.
920    
921    4.  Inside a character class, PCRE always treated \R and \X as literals,
922        whereas Perl faults them if its -w option is set. I have changed PCRE so
923        that it faults them when PCRE_EXTRA is set.
924    
925    5.  Added support for \N, which always matches any character other than
926        newline. (It is the same as "." when PCRE_DOTALL is not set.)
927    
928    6.  When compiling pcregrep with newer versions of gcc which may have
929        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
930        declared with attribute warn_unused_result" were given. Just casting the
931        result to (void) does not stop the warnings; a more elaborate fudge is
932        needed. I've used a macro to implement this.
933    
934    7.  Minor change to pcretest.c to avoid a compiler warning.
935    
936    8.  Added four artifical Unicode properties to help with an option to make
937        \s etc use properties (see next item). The new properties are: Xan
938        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
939    
940    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
941        use Unicode properties. (*UCP) at the start of a pattern can be used to set
942        this option. Modified pcretest to add /W to test this facility. Added
943        REG_UCP to make it available via the POSIX interface.
944    
945    10. Added --line-buffered to pcregrep.
946    
947    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
948        studied, and the match started with a letter with a code point greater than
949        127 whose first byte was different to the first byte of the other case of
950        the letter, the other case of this starting letter was not recognized
951        (#976).
952    
953    12. If a pattern that was studied started with a repeated Unicode property
954        test, for example, \p{Nd}+, there was the theoretical possibility of
955        setting up an incorrect bitmap of starting bytes, but fortunately it could
956        not have actually happened in practice until change 8 above was made (it
957        added property types that matched character-matching opcodes).
958    
959    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
960        possible starting bytes for non-anchored patterns.
961    
962    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
963        \R, and also a number of cases that involve Unicode properties, both
964        explicit and implicit when PCRE_UCP is set.
965    
966    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
967        input, it could crash or give wrong results if characters with values
968        greater than 0xc0 were present in the subject string. (Detail: it assumed
969        UTF-8 input when processing these items.)
970    
971    16. Added a lot of (int) casts to avoid compiler warnings in systems where
972        size_t is 64-bit (#991).
973    
974    17. Added a check for running out of memory when PCRE is compiled with
975        --disable-stack-for-recursion (#990).
976    
977    18. If the last data line in a file for pcretest does not have a newline on
978        the end, a newline was missing in the output.
979    
980    19. The default pcre_chartables.c file recognizes only ASCII characters (values
981        less than 128) in its various bitmaps. However, there is a facility for
982        generating tables according to the current locale when PCRE is compiled. It
983        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
984        characters, are recognized by isspace() and therefore were getting set in
985        these tables, and indeed these tables seem to approximate to ISO 8859. This
986        caused a problem in UTF-8 mode when pcre_study() was used to create a list
987        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
988        which of course cannot start UTF-8 characters. I have changed the code so
989        that only real ASCII characters (less than 128) and the correct starting
990        bytes for UTF-8 encodings are set for characters greater than 127 when in
991        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
992        altogether.)
993    
994    20. Added the /T option to pcretest so as to be able to run tests with non-
995        standard character tables, thus making it possible to include the tests
996        used for 19 above in the standard set of tests.
997    
998    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
999        reference to a subpattern the other side of a comment that contains an
1000        opening parenthesis caused either an internal compiling error, or a
1001        reference to the wrong subpattern.
1002    
1003    
1004    Version 8.02 19-Mar-2010
1005    ------------------------
1006    
1007    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1008    
1009    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1010        configured.
1011    
1012    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1013        original author of that file, following a query about its status.
1014    
1015    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1016        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1017    
1018    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1019        quantifier applied to a forward-referencing subroutine call, could compile
1020        incorrect code or give the error "internal error: previously-checked
1021        referenced subpattern not found".
1022    
1023    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1024        variables to point to external functions. For these systems, therefore,
1025        pcre_malloc etc. are now initialized to local functions that call the
1026        relevant global functions.
1027    
1028    7.  There were two entries missing in the vectors called coptable and poptable
1029        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1030        I've fixed the data, and added a kludgy way of testing at compile time that
1031        the lengths are correct (equal to the number of opcodes).
1032    
1033    8.  Following on from 7, I added a similar kludge to check the length of the
1034        eint vector in pcreposix.c.
1035    
1036    9.  Error texts for pcre_compile() are held as one long string to avoid too
1037        much relocation at load time. To find a text, the string is searched,
1038        counting zeros. There was no check for running off the end of the string,
1039        which could happen if a new error number was added without updating the
1040        string.
1041    
1042    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1043    
1044    11. \K was not working if it appeared in an atomic group or in a group that
1045        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1046        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1047        the assertion is positive, but not if it is negative.
1048    
1049    12. Change 11 fortuitously reduced the size of the stack frame used in the
1050        "match()" function of pcre_exec.c by one pointer. Forthcoming
1051        implementation of support for (*MARK) will need an extra pointer on the
1052        stack; I have reserved it now, so that the stack frame size does not
1053        decrease.
1054    
1055    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1056        item in branch that calls a recursion is a subroutine call - as in the
1057        second branch in the above example - was incorrectly given the compile-
1058        time error "recursive call could loop indefinitely" because pcre_compile()
1059        was not correctly checking the subroutine for matching a non-empty string.
1060    
1061    14. The checks for overrunning compiling workspace could trigger after an
1062        overrun had occurred. This is a "should never occur" error, but it can be
1063        triggered by pathological patterns such as hundreds of nested parentheses.
1064        The checks now trigger 100 bytes before the end of the workspace.
1065    
1066    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1067    
1068    
1069    Version 8.01 19-Jan-2010
1070    ------------------------
1071    
1072    1.  If a pattern contained a conditional subpattern with only one branch (in
1073        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1074        computed the wrong minimum data length (which is of course zero for such
1075        subpatterns). This could cause incorrect "no match" results.
1076    
1077    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1078        the pattern is reset in the first branch, pcre_compile() failed with
1079        "internal error: code overflow at offset...". This happened only when
1080        the reset was to the original external option setting. (An optimization
1081        abstracts leading options settings into an external setting, which was the
1082        cause of this.)
1083    
1084    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1085        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1086        assertion pattern did not match (meaning that the assertion was true), it
1087        was incorrectly treated as false if the SKIP had been reached during the
1088        matching. This also applied to assertions used as conditions.
1089    
1090    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1091        assertion subpattern, including such a pattern used as a condition,
1092        unpredictable results occurred, instead of the error return
1093        PCRE_ERROR_DFA_UITEM.
1094    
1095    5.  The C++ GlobalReplace function was not working like Perl for the special
1096        situation when an empty string is matched. It now does the fancy magic
1097        stuff that is necessary.
1098    
1099    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1100        removed. (These were left over from very, very early versions of PCRE.)
1101    
1102    7.  Some cosmetic changes to the code to make life easier when compiling it
1103        as part of something else:
1104    
1105        (a) Change DEBUG to PCRE_DEBUG.
1106    
1107        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1108            called "current" as "current_branch", to prevent a collision with the
1109            Linux macro when compiled as a kernel module.
1110    
1111        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1112            prevent a collision with the Linux macro when compiled as a kernel
1113            module.
1114    
1115    8.  In pcre_compile() there are some checks for integer overflows that used to
1116        cast potentially large values to (double). This has been changed to that
1117        when building, a check for int64_t is made, and if it is found, it is used
1118        instead, thus avoiding the use of floating point arithmetic. (There is no
1119        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1120        double.
1121    
1122    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1123        2005 (difference between two addresses compared to an unsigned value).
1124    
1125    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1126        custom one, because of the following reported problem in Windows:
1127    
1128          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1129              under Win32.
1130          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1131              therefore missing the function definition.
1132          - The compiler thus generates a "C" signature for the test function.
1133          - The linker fails to find the "C" function.
1134          - PCRE fails to configure if asked to do so against libbz2.
1135    
1136    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1137        messages were output:
1138    
1139          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1140          rerunning libtoolize, to keep the correct libtool macros in-tree.
1141          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1142    
1143        I have done both of these things.
1144    
1145    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1146        most of the time, it *can* run out if it is given a pattern that contains a
1147        runaway infinite recursion. I updated the discussion in the pcrestack man
1148        page.
1149    
1150    13. Now that we have gone to the x.xx style of version numbers, the minor
1151        version may start with zero. Using 08 or 09 is a bad idea because users
1152        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1153        interpreted as invalid octal numbers. I've updated the previous comment in
1154        configure.ac, and also added a check that gives an error if 08 or 09 are
1155        used.
1156    
1157    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1158        causing partial matching to fail when the end of the subject matched \W
1159        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1160    
1161    15. There were some discrepancies between the declarations in pcre_internal.h
1162        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1163        their definitions. The declarations used "const uschar *" and the
1164        definitions used USPTR. Even though USPTR is normally defined as "const
1165        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1166        reported that: "This difference in casting confuses some C++ compilers, for
1167        example, SunCC recognizes above declarations as different functions and
1168        generates broken code for hbpcre." I have changed the declarations to use
1169        USPTR.
1170    
1171    16. GNU libtool is named differently on some systems. The autogen.sh script now
1172        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1173        (FreeBSD).
1174    
1175    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1176        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1177        comment: "Figure out how to create a longlong from a string: strtoll and
1178        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1179        instance, but it only takes 2 args instead of 3!"
1180    
1181    18. A subtle bug concerned with back references has been fixed by a change of
1182        specification, with a corresponding code fix. A pattern such as
1183        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1184        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1185        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1186        same bug. Such groups have to be quantified to be useful, or contained
1187        inside another quantified group. (If there's no repetition, the reference
1188        can never match.) The problem arises because, having left the group and
1189        moved on to the rest of the pattern, a later failure that backtracks into
1190        the group uses the captured value from the final iteration of the group
1191        rather than the correct earlier one. I have fixed this in PCRE by forcing
1192        any group that contains a reference to itself to be an atomic group; that
1193        is, there cannot be any backtracking into it once it has completed. This is
1194        similar to recursive and subroutine calls.
1195    
1196    
1197    Version 8.00 19-Oct-09
1198    ----------------------
1199    
1200    1.  The table for translating pcre_compile() error codes into POSIX error codes
1201        was out-of-date, and there was no check on the pcre_compile() error code
1202        being within the table. This could lead to an OK return being given in
1203        error.
1204    
1205    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1206        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1207        in a Windows environment.
1208    
1209    3.  The pcregrep --count option prints the count for each file even when it is
1210        zero, as does GNU grep. However, pcregrep was also printing all files when
1211        --files-with-matches was added. Now, when both options are given, it prints
1212        counts only for those files that have at least one match. (GNU grep just
1213        prints the file name in this circumstance, but including the count seems
1214        more useful - otherwise, why use --count?) Also ensured that the
1215        combination -clh just lists non-zero counts, with no names.
1216    
1217    4.  The long form of the pcregrep -F option was incorrectly implemented as
1218        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1219        but it seems right to fix it, and I didn't think it was worth preserving
1220        the old behaviour.
1221    
1222    5.  The command line items --regex=pattern and --regexp=pattern were not
1223        recognized by pcregrep, which required --regex pattern or --regexp pattern
1224        (with a space rather than an '='). The man page documented the '=' forms,
1225        which are compatible with GNU grep; these now work.
1226    
1227    6.  No libpcreposix.pc file was created for pkg-config; there was just
1228        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1229    
1230    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1231        when UCP support is not needed, by modifying the Python script that
1232        generates it from Unicode data files. This should not matter if the module
1233        is correctly used as a library, but I received one complaint about 50K of
1234        unwanted data. My guess is that the person linked everything into his
1235        program rather than using a library. Anyway, it does no harm.
1236    
1237    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1238        was a minimum greater than 1 for a wide character in a possessive
1239        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1240        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1241        character. Chaos in the form of incorrect output or a compiling loop could
1242        result.
1243    
1244    9.  The restrictions on what a pattern can contain when partial matching is
1245        requested for pcre_exec() have been removed. All patterns can now be
1246        partially matched by this function. In addition, if there are at least two
1247        slots in the offset vector, the offset of the earliest inspected character
1248        for the match and the offset of the end of the subject are set in them when
1249        PCRE_ERROR_PARTIAL is returned.
1250    
1251    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1252        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1253        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1254        and may be more useful for multi-segment matching.
1255    
1256    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1257        used to be given if ever the end of the subject was reached; now it is
1258        given only if matching could not proceed because another character was
1259        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1260        string "Z", which now yields "no match" instead of "partial match". In the
1261        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1262        final character ended with (*FAIL).
1263    
1264    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1265        if the pattern had a "must contain" character that was already found in the
1266        earlier partial match, unless partial matching was again requested. For
1267        example, with the pattern /dog.(body)?/, the "must contain" character is
1268        "g". If the first part-match was for the string "dog", restarting with
1269        "sbody" failed. This bug has been fixed.
1270    
1271    13. The string returned by pcre_dfa_exec() after a partial match has been
1272        changed so that it starts at the first inspected character rather than the
1273        first character of the match. This makes a difference only if the pattern
1274        starts with a lookbehind assertion or \b or \B (\K is not supported by
1275        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1276        matching functions compatible, and I think it's the right thing to do.
1277    
1278    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1279        so that the demonstration program is easily available in environments where
1280        PCRE has not been installed from source.
1281    
1282    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1283        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1284        library.
1285    
1286    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1287        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1288        is not the first non-POSIX option to be added. Clearly some people find
1289        these options useful.
1290    
1291    17. If a caller to the POSIX matching function regexec() passes a non-zero
1292        value for nmatch with a NULL value for pmatch, the value of
1293        nmatch is forced to zero.
1294    
1295    18. RunGrepTest did not have a test for the availability of the -u option of
1296        the diff command, as RunTest does. It now checks in the same way as
1297        RunTest, and also checks for the -b option.
1298    
1299    19. If an odd number of negated classes containing just a single character
1300        interposed, within parentheses, between a forward reference to a named
1301        subpattern and the definition of the subpattern, compilation crashed with
1302        an internal error, complaining that it could not find the referenced
1303        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1304        [The bug was that it was starting one character too far in when skipping
1305        over the character class, thus treating the ] as data rather than
1306        terminating the class. This meant it could skip too much.]
1307    
1308    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1309        /g option in pcretest when the pattern contains \K, which makes it possible
1310        to have an empty string match not at the start, even when the pattern is
1311        anchored. Updated pcretest and pcredemo to use this option.
1312    
1313    21. If the maximum number of capturing subpatterns in a recursion was greater
1314        than the maximum at the outer level, the higher number was returned, but
1315        with unset values at the outer level. The correct (outer level) value is
1316        now given.
1317    
1318    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1319        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1320        make it do so. The string so far is captured, making this feature
1321        compatible with Perl.
1322    
1323    23. The tests have been re-organized, adding tests 11 and 12, to make it
1324        possible to check the Perl 5.10 features against Perl 5.10.
1325    
1326    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1327        pattern matches a fixed length string. PCRE did not allow this; now it
1328        does. Neither allows recursion.
1329    
1330    25. I finally figured out how to implement a request to provide the minimum
1331        length of subject string that was needed in order to match a given pattern.
1332        (It was back references and recursion that I had previously got hung up
1333        on.) This code has now been added to pcre_study(); it finds a lower bound
1334        to the length of subject needed. It is not necessarily the greatest lower
1335        bound, but using it to avoid searching strings that are too short does give
1336        some useful speed-ups. The value is available to calling programs via
1337        pcre_fullinfo().
1338    
1339    26. While implementing 25, I discovered to my embarrassment that pcretest had
1340        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1341        study optimizations had never been tested with that matching function.
1342        Oops. What is worse, even when it was passed study data, there was a bug in
1343        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1344        were also very few tests of studied patterns with pcre_dfa_exec().
1345    
1346    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1347        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1348        on the other side of the coin, they are no longer allowed to have different
1349        names, because these cannot be distinguished in PCRE, and this has caused
1350        confusion. (This is a difference from Perl.)
1351    
1352    28. When duplicate subpattern names are present (necessarily with different
1353        numbers, as required by 27 above), and a test is made by name in a
1354        conditional pattern, either for a subpattern having been matched, or for
1355        recursion in such a pattern, all the associated numbered subpatterns are
1356        tested, and the overall condition is true if the condition is true for any
1357        one of them. This is the way Perl works, and is also more like the way
1358        testing by number works.
1359    
1360    
1361    Version 7.9 11-Apr-09
1362    ---------------------
1363    
1364    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1365        (pcretest), all targets were linked against these libraries. This included
1366        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1367        libraries. This caused unwanted dependencies to be created. This problem
1368        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1369        pcretest is linked with readline.
1370    
1371    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1372        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1373        moved outside it again, because FALSE and TRUE are already defined in AIX,
1374        but BOOL is not.
1375    
1376    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1377        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1378    
1379    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1380        hyphens) following file names and line numbers when outputting matching
1381        lines. This is not true; no spaces are inserted. I have also clarified the
1382        wording for the --colour (or --color) option.
1383    
1384    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1385        was not coloured; this is different to GNU grep, so I have changed it to be
1386        the same.
1387    
1388    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1389        each matching line was coloured. Now it goes on to look for further matches
1390        of any of the test patterns, which is the same behaviour as GNU grep.
1391    
1392    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1393        doesn't make sense to accept an empty string match in pcregrep, so I have
1394        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1395        seems to be how GNU grep behaves.
1396    
1397    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1398        start or after a newline", because the conditional assertion was not being
1399        correctly handled. The rule now is that both the assertion and what follows
1400        in the first alternative must satisfy the test.
1401    
1402    9.  If auto-callout was enabled in a pattern with a conditional group whose
1403        condition was an assertion, PCRE could crash during matching, both with
1404        pcre_exec() and pcre_dfa_exec().
1405    
1406    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1407        used for matching.
1408    
1409    11. Unicode property support in character classes was not working for
1410        characters (bytes) greater than 127 when not in UTF-8 mode.
1411    
1412    12. Added the -M command line option to pcretest.
1413    
1414    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1415    
1416    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1417    
1418    16. Added comments and documentation about mis-use of no_arg in the C++
1419        wrapper.
1420    
1421    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1422        from Martin Jerabek that uses macro names for all relevant character and
1423        string constants.
1424    
1425    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1426        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1427        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1428        these, but not everybody uses configure.
1429    
1430    19. A conditional group that had only one branch was not being correctly
1431        recognized as an item that could match an empty string. This meant that an
1432        enclosing group might also not be so recognized, causing infinite looping
1433        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1434        with the subject "ab", where knowledge that the repeated group can match
1435        nothing is needed in order to break the loop.
1436    
1437    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1438        exec(), but without supplying a callout function, matching went wrong.
1439    
1440    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1441        leak if the size of the offset vector was greater than 30. When the vector
1442        is smaller, the saved offsets during recursion go onto a local stack
1443        vector, but for larger vectors malloc() is used. It was failing to free
1444        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1445        error, in fact).
1446    
1447    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1448        heapframe that is used only when UTF-8 support is enabled. This caused no
1449        problem, but was untidy.
1450    
1451    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1452        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1453        included within another project.
1454    
1455    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1456        slightly modified by me:
1457    
1458          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1459              not building pcregrep.
1460    
1461          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1462              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1463    
1464    25. Forward references, both numeric and by name, in patterns that made use of
1465        duplicate group numbers, could behave incorrectly or give incorrect errors,
1466        because when scanning forward to find the reference group, PCRE was not
1467        taking into account the duplicate group numbers. A pattern such as
1468        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1469    
1470    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1471        the feature of a custom pointer more persuasive (as requested by a user).
1472    
1473    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1474        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1475        pre-defined.
1476    
1477    28. Added support for (*UTF8) at the start of a pattern.
1478    
1479    29. Arrange for flags added by the "release type" setting in CMake to be shown
1480        in the configuration summary.
1481    
1482    
1483    Version 7.8 05-Sep-08
1484    ---------------------
1485    
1486    1.  Replaced UCP searching code with optimized version as implemented for Ad
1487        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1488        stage table and inline lookup instead of a function, giving speed ups of 2
1489        to 5 times on some simple patterns that I tested. Permission was given to
1490        distribute the MultiStage2.py script that generates the tables (it's not in
1491        the tarball, but is in the Subversion repository).
1492    
1493    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1494        scripts.
1495    
1496    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1497        a group with a zero qualifier. The result of the study could be incorrect,
1498        or the function might crash, depending on the pattern.
1499    
1500    4.  Caseless matching was not working for non-ASCII characters in back
1501        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1502        It now works when Unicode Property Support is available.
1503    
1504    5.  In pcretest, an escape such as \x{de} in the data was always generating
1505        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1506        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1507        truncation.
1508    
1509    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1510    
1511    7.  Added two (int) casts to pcregrep when printing the difference of two
1512        pointers, in case they are 64-bit values.
1513    
1514    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1515        test 2 if it fails.
1516    
1517    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1518        and a #define of that name to empty if it is not externally set. This is to
1519        allow users of MSVC to set it if necessary.
1520    
1521    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1522        the convenience functions in the pcre_get.c source file.
1523    
1524    11. An option change at the start of a pattern that had top-level alternatives
1525        could cause overwriting and/or a crash. This command provoked a crash in
1526        some environments:
1527    
1528          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1529    
1530        This potential security problem was recorded as CVE-2008-2371.
1531    
1532    12. For a pattern where the match had to start at the beginning or immediately
1533        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1534        pcre_dfa_exec() could read past the end of the passed subject if there was
1535        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1536        pcretest so that it places the subject at the end of its malloc-ed buffer.
1537    
1538    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1539        exec() might read past the end of the data buffer in UTF-8 mode.
1540    
1541    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1542        the data contained the byte 0x85 as part of a UTF-8 character within its
1543        first line. This applied both to normal and DFA matching.
1544    
1545    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1546        /^[^d]*?$/8 failed to match "abc".
1547    
1548    16. Added a missing copyright notice to pcrecpp_internal.h.
1549    
1550    17. Make it more clear in the documentation that values returned from
1551        pcre_exec() in ovector are byte offsets, not character counts.
1552    
1553    18. Tidied a few places to stop certain compilers from issuing warnings.
1554    
1555    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1556        supplied by Stefan Weber. I made a further small update for 7.8 because
1557        there is a change of source arrangements: the pcre_searchfuncs.c module is
1558        replaced by pcre_ucd.c.
1559    
1560    
1561    Version 7.7 07-May-08
1562    ---------------------
1563    
1564    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1565        a string to a long long, pretend we don't even have a long long." This is
1566        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1567    
1568    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1569        pre-7.6 versions, which defined a global no_arg variable instead of putting
1570        it in the RE class. (See also #8 below.)
1571    
1572    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1573        Lopes.
1574    
1575    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1576    
1577        (1) The include/exclude patterns were being applied to the whole pathnames
1578            of files, instead of just to the final components.
1579    
1580        (2) If there was more than one level of directory, the subdirectories were
1581            skipped unless they satisfied the include/exclude conditions. This is
1582            inconsistent with GNU grep (and could even be seen as contrary to the
1583            pcregrep specification - which I improved to make it absolutely clear).
1584            The action now is always to scan all levels of directory, and just
1585            apply the include/exclude patterns to regular files.
1586    
1587    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1588        --exclude_dir in the tests to avoid scanning .svn directories.
1589    
1590    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1591        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1592        doesn't support NULs in patterns.
1593    
1594    7.  Added some missing "const"s to declarations of static tables in
1595        pcre_compile.c and pcre_dfa_exec.c.
1596    
1597    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1598        caused by fix #2  above. (Subsequently also a second patch to fix the
1599        first patch. And a third patch - this was a messy problem.)
1600    
1601    9.  Applied Craig's patch to remove the use of push_back().
1602    
1603    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1604        matching function regexec().
1605    
1606    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1607        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1608        references. PCRE supports relative numbers with this syntax (I don't think
1609        Oniguruma does).
1610    
1611    12. Previously, a group with a zero repeat such as (...){0} was completely
1612        omitted from the compiled regex. However, this means that if the group
1613        was called as a subroutine from elsewhere in the pattern, things went wrong
1614        (an internal error was given). Such groups are now left in the compiled
1615        pattern, with a new opcode that causes them to be skipped at execution
1616        time.
1617    
1618    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1619        to the way PCRE behaves:
1620    
1621        (a) A lone ] character is dis-allowed (Perl treats it as data).
1622    
1623        (b) A back reference to an unmatched subpattern matches an empty string
1624            (Perl fails the current match path).
1625    
1626        (c) A data ] in a character class must be notated as \] because if the
1627            first data character in a class is ], it defines an empty class. (In
1628            Perl it is not possible to have an empty class.) The empty class []
1629            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1630            The negative empty class [^] matches any one character, independently
1631            of the DOTALL setting.
1632    
1633    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1634        non-existent subpattern following a character class starting with ']' and
1635        containing () gave an internal compiling error instead of "reference to
1636        non-existent subpattern". Fortunately, when the pattern did exist, the
1637        compiled code was correct. (When scanning forwards to check for the
1638        existencd of the subpattern, it was treating the data ']' as terminating
1639        the class, so got the count wrong. When actually compiling, the reference
1640        was subsequently set up correctly.)
1641    
1642    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1643        it was being rejected as not supported by pcre_dfa_exec(), even though
1644        other assertions are supported. I have made pcre_dfa_exec() support
1645        (*FAIL).
1646    
1647    16. The implementation of 13c above involved the invention of a new opcode,
1648        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1649        cannot be changed at match time, I realized I could make a small
1650        improvement to matching performance by compiling OP_ALLANY instead of
1651        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1652        on the OP_ANY path.
1653    
1654    17. Compiling pcretest on Windows with readline support failed without the
1655        following two fixes: (1) Make the unistd.h include conditional on
1656        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1657    
1658    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1659        ncurses library to be included for pcretest when ReadLine support is
1660        requested, but also to allow for it to be overridden. This patch came from
1661        Daniel Bergström.
1662    
1663    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1664        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1665        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1666        spotting this.
1667    
1668    
1669    Version 7.6 28-Jan-08
1670    ---------------------
1671    
1672    1.  A character class containing a very large number of characters with
1673        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1674        overflow.
1675    
1676    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1677        HAVE_LONG_LONG is not defined.
1678    
1679    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1680        bring them up to date and include new features. This patch includes:
1681    
1682        - Fixed PH's badly added libz and libbz2 support.
1683        - Fixed a problem with static linking.
1684        - Added pcredemo. [But later removed - see 7 below.]
1685        - Fixed dftables problem and added an option.
1686        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1687            HAVE_LONG_LONG.
1688        - Added readline support for pcretest.
1689        - Added an listing of the option settings after cmake has run.
1690    
1691    4.  A user submitted a patch to Makefile that makes it easy to create
1692        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1693        Makefile.am that cause it to include this special target, without
1694        affecting anything else. Note that the same mingw target plus all
1695        the other distribution libraries and programs are now supported
1696        when configuring with CMake (see 6 below) instead of with
1697        Configure/Make.
1698    
1699    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1700        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1701        exported in the Windows port". It has not yet been confirmed that the patch
1702        solves the problem, but it does no harm.
1703    
1704    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1705        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1706        with CMake, and also correct the comment about stack recursion.
1707    
1708    7.  Remove the automatic building of pcredemo from the ./configure system and
1709        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1710        of a program that users should build themselves after PCRE is installed, so
1711        building it automatically is not really right. What is more, it gave
1712        trouble in some build environments.
1713    
1714    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1715    
1716    
1717    Version 7.5 10-Jan-08
1718    ---------------------
1719    
1720    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1721        values in parens when parsing an RE using the C++ wrapper."
1722    
1723    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1724        Characters greater than 255 were excluded from the class instead of being
1725        included.
1726    
1727    3.  The same bug as (2) above applied to negated POSIX classes such as
1728        [:^space:].
1729    
1730    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1731        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1732        I have changed it.
1733    
1734    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1735        first named subpattern) and a construct such as (?&a) would reference the
1736        first named subpattern whose name started with "a" (in other words, the
1737        length check was missing). Both these problems are fixed. "Subpattern name
1738        expected" is now given for (?&) (a zero-length name), and this patch also
1739        makes it give the same error for \k'' (previously it complained that that
1740        was a reference to a non-existent subpattern).
1741    
1742    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1743        this is right because (?- can be followed by option settings as well as by
1744        digits. I have, however, made the messages clearer.
1745    
1746    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1747        than the number used in the conditional) now cause a compile-time error.
1748        This is actually not compatible with Perl, which accepts such patterns, but
1749        treats the conditional as always being FALSE (as PCRE used to), but it
1750        seems to me that giving a diagnostic is better.
1751    
1752    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1753        and messages.
1754    
1755    9.  Fix two occurrences of "backslash" in comments that should have been
1756        "backspace".
1757    
1758    10. Remove two redundant lines of code that can never be obeyed (their function
1759        was moved elsewhere).
1760    
1761    11. The program that makes PCRE's Unicode character property table had a bug
1762        which caused it to generate incorrect table entries for sequences of
1763        characters that have the same character type, but are in different scripts.
1764        It amalgamated them into a single range, with the script of the first of
1765        them. In other words, some characters were in the wrong script. There were
1766        thirteen such cases, affecting characters in the following ranges:
1767    
1768          U+002b0 - U+002c1
1769          U+0060c - U+0060d
1770          U+0061e - U+00612
1771          U+0064b - U+0065e
1772          U+0074d - U+0076d
1773          U+01800 - U+01805
1774          U+01d00 - U+01d77
1775          U+01d9b - U+01dbf
1776          U+0200b - U+0200f
1777          U+030fc - U+030fe
1778          U+03260 - U+0327f
1779          U+0fb46 - U+0fbb1
1780          U+10450 - U+1049d
1781    
1782    12. The -o option (show only the matching part of a line) for pcregrep was not
1783        compatible with GNU grep in that, if there was more than one match in a
1784        line, it showed only the first of them. It now behaves in the same way as
1785        GNU grep.
1786    
1787    13. If the -o and -v options were combined for pcregrep, it printed a blank
1788        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1789        does the same. The return code can be used to tell if there were any
1790        non-matching lines.
1791    
1792    14. Added --file-offsets and --line-offsets to pcregrep.
1793    
1794    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1795        infinitely looping recursion. The bug was that positive lookaheads were not
1796        being skipped when checking for a possible empty match (negative lookaheads
1797        and both kinds of lookbehind were skipped).
1798    
1799    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1800        inclusion of <windows.h> to before rather than after the definition of
1801        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1802    
1803    17. Specifying a possessive quantifier with a specific limit for a Unicode
1804        character property caused pcre_compile() to compile bad code, which led at
1805        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1806        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1807        caused the error; without that there was no problem.
1808    
1809    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1810    
1811    19. Added --enable-pcretest-libreadline.
1812    
1813    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1814        RE::GlobalReplace(). As a result, the number of replacements returned was
1815        double what it should be. I removed one of the increments, but Craig sent a
1816        later patch that removed the other one (the right fix) and added unit tests
1817        that check the return values (which was not done before).
1818    
1819    21. Several CMake things:
1820    
1821        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1822            the names libpcre and libpcreposix, not just pcre and pcreposix.
1823    
1824        (2) The above change means that pcretest and pcregrep are now correctly
1825            linked with the newly-built libraries, not previously installed ones.
1826    
1827        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1828    
1829    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1830        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1831        UTF-8 newline character). The key issue is that the pattern starts .*;
1832        this means that the match must be either at the beginning, or after a
1833        newline. The bug was in the code for advancing after a failed match and
1834        checking that the new position followed a newline. It was not taking
1835        account of UTF-8 characters correctly.
1836    
1837    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1838        character classes. PCRE was not treating the sequence [:...:] as a
1839        character class unless the ... were all letters. Perl, however, seems to
1840        allow any characters between [: and :], though of course it rejects as
1841        unknown any "names" that contain non-letters, because all the known class
1842        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1843        for example, whereas PCRE did not - it did not recognize a POSIX character
1844        class. This seemed a bit dangerous, so the code has been changed to be
1845        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1846        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1847        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1848        Perl does, and where it didn't before.
1849    
1850    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1851        Windows environments %n is disabled by default.
1852    
1853    
1854    Version 7.4 21-Sep-07
1855    ---------------------
1856    
1857    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1858        means that a class such as [\s] counted as "explicit reference to CR or
1859        LF". That isn't really right - the whole point of the change was to try to
1860        help when there was an actual mention of one of the two characters. So now
1861        the change happens only if \r or \n (or a literal CR or LF) character is
1862        encountered.
1863    
1864    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1865        of both had grown to the point where there were only 3 bits left.
1866        Fortunately, there was spare space in the data structure, and so I have
1867        moved the internal flags into a new 16-bit field to free up more option
1868        bits.
1869    
1870    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1871        but did not set the internal JCHANGED flag - either of these is enough to
1872        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1873        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1874        start sets both bits.
1875    
1876    4.  Added options (at build time, compile time, exec time) to change \R from
1877        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1878    
1879    5.  doc/pcresyntax.html was missing from the distribution.
1880    
1881    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1882        compatibility, even though it is no longer used.
1883    
1884    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1885        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1886        windows.h file is present (where different names are used). [This was
1887        reversed later after testing - see 16 below.]
1888    
1889    8.  Changed all #include <config.h> to #include "config.h". There were also
1890        some further <pcre.h> cases that I changed to "pcre.h".
1891    
1892    9.  When pcregrep was used with the --colour option, it missed the line ending
1893        sequence off the lines that it output.
1894    
1895    10. It was pointed out to me that arrays of string pointers cause lots of
1896        relocations when a shared library is dynamically loaded. A technique of
1897        using a single long string with a table of offsets can drastically reduce
1898        these. I have refactored PCRE in four places to do this. The result is
1899        dramatic:
1900    
1901          Originally:                          290
1902          After changing UCP table:            187
1903          After changing error message table:   43
1904          After changing table of "verbs"       36
1905          After changing table of Posix names   22
1906    
1907        Thanks to the folks working on Gregex for glib for this insight.
1908    
1909    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1910        unicode-properties was also set.
1911    
1912    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1913    
1914    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1915        checked only for CRLF.
1916    
1917    14. Added casts to pcretest.c to avoid compiler warnings.
1918    
1919    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1920    
1921    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1922        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1923        entirely. This removes changes made in 7 above.
1924    
1925    17. The CMake files have been updated, and there is now more information about
1926        building with CMake in the NON-UNIX-USE document.
1927    
1928    
1929    Version 7.3 28-Aug-07
1930    ---------------------
1931    
1932     1. In the rejigging of the build system that eventually resulted in 7.1, the
1933        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1934        brackets there is not right, since it causes compilers to look for an
1935        installed pcre.h, not the version that is in the source that is being
1936        compiled (which of course may be different). I have changed it back to:
1937    
1938          #include "pcre.h"
1939    
1940        I have a vague recollection that the change was concerned with compiling in
1941        different directories, but in the new build system, that is taken care of
1942        by the VPATH setting the Makefile.
1943    
1944     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1945        when the subject happened to end in the byte 0x85 (e.g. if the last
1946        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1947        characters but of course it shouldn't be taken as a newline when it is part
1948        of another character. The bug was that, for an unlimited repeat of . in
1949        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1950        characters when looking for a newline.
1951    
1952     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1953    
1954     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1955        in debug output.
1956    
1957     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1958        long printing in the pcrecpp unittest when running under MinGW.
1959    
1960     6. ESC_K was left out of the EBCDIC table.
1961    
1962     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1963        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1964        limit also applies to "virtual nesting" when a pattern is recursive, and in
1965        this case 1000 isn't so big. I have been able to remove this limit at the
1966        expense of backing off one optimization in certain circumstances. Normally,
1967        when pcre_exec() would call its internal match() function recursively and
1968        immediately return the result unconditionally, it uses a "tail recursion"
1969        feature to save stack. However, when a subpattern that can match an empty
1970        string has an unlimited repetition quantifier, it no longer makes this
1971        optimization. That gives it a stack frame in which to save the data for
1972        checking that an empty string has been matched. Previously this was taken
1973        from the 1000-entry workspace that had been reserved. So now there is no
1974        explicit limit, but more stack is used.
1975    
1976     8. Applied Daniel's patches to solve problems with the import/export magic
1977        syntax that is required for Windows, and which was going wrong for the
1978        pcreposix and pcrecpp parts of the library. These were overlooked when this
1979        problem was solved for the main library.
1980    
1981     9. There were some crude static tests to avoid integer overflow when computing
1982        the size of patterns that contain repeated groups with explicit upper
1983        limits. As the maximum quantifier is 65535, the maximum group length was
1984        set at 30,000 so that the product of these two numbers did not overflow a
1985        32-bit integer. However, it turns out that people want to use groups that
1986        are longer than 30,000 bytes (though not repeat them that many times).
1987        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1988        made it possible to implement the integer overflow checks in a much more
1989        dynamic way, which I have now done. The artificial limitation on group
1990        length has been removed - we now have only the limit on the total length of
1991        the compiled pattern, which depends on the LINK_SIZE setting.
1992    
1993    10. Fixed a bug in the documentation for get/copy named substring when
1994        duplicate names are permitted. If none of the named substrings are set, the
1995        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1996        empty string.
1997    
1998    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1999        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2000        because the ] is interpreted as the first data character and the
2001        terminating ] is not found. PCRE has been made compatible with Perl in this
2002        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2003        cause memory overwriting.
2004    
2005    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2006        string has been matched (to stop an infinite loop). It was not recognizing
2007        a conditional subpattern that could match an empty string if that
2008        subpattern was within another subpattern. For example, it looped when
2009        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2010        condition was not nested. This bug has been fixed.
2011    
2012    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2013        past the start of the subject in the presence of bytes with the top bit
2014        set, for example "\x8aBCD".
2015    
2016    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2017        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2018    
2019    14. Optimized (?!) to (*FAIL).
2020    
2021    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2022        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2023        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2024        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2025        does: it's just the validity check that is more restrictive.
2026    
2027    16. Inserted checks for integer overflows during escape sequence (backslash)
2028        processing, and also fixed erroneous offset values for syntax errors during
2029        backslash processing.
2030    
2031    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2032        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2033    
2034    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2035        caused an overrun.
2036    
2037    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2038        something other than just ASCII characters) inside a group that had an
2039        unlimited repeat caused a loop at compile time (while checking to see
2040        whether the group could match an empty string).
2041    
2042    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2043        [\P{Any}] did so. (Error in the code for printing property names.)
2044    
2045    21. An orphan \E inside a character class could cause a crash.
2046    
2047    22. A repeated capturing bracket such as (A)? could cause a wild memory
2048        reference during compilation.
2049    
2050    23. There are several functions in pcre_compile() that scan along a compiled
2051        expression for various reasons (e.g. to see if it's fixed length for look
2052        behind). There were bugs in these functions when a repeated \p or \P was
2053        present in the pattern. These operators have additional parameters compared
2054        with \d, etc, and these were not being taken into account when moving along
2055        the compiled data. Specifically:
2056    
2057        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2058            length.
2059    
2060        (b) An item such as \pL+ within a repeated group could cause crashes or
2061            loops.
2062    
2063        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2064            "reference to non-existent subpattern" error.
2065    
2066        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2067    
2068    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2069        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2070    
2071    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2072    
2073    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2074        character were causing crashes (broken optimization).
2075    
2076    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2077        \p or \P) caused a compile-time loop.
2078    
2079    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2080        break. For example, the unstudied pattern [\r\n]A does not match the string
2081        "\r\nA" because change 7.0/46 below moves the current point on by two
2082        characters after failing to match at the start. However, the pattern \nA
2083        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2084        the same is true. There doesn't seem any very clean way out of this, but
2085        what I have chosen to do makes the common cases work: PCRE now takes note
2086        of whether there can be an explicit match for \r or \n anywhere in the
2087        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2088        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2089        pattern has explicit CR or LF references.
2090    
2091    29. Added (*CR) etc for changing newline setting at start of pattern.
2092    
2093    
2094    Version 7.2 19-Jun-07
2095    ---------------------
2096    
2097     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2098        which is apparently normally available under Windows.
2099    
2100     2. Re-jig the pcregrep tests with different newline settings in an attempt
2101        to make them independent of the local environment's newline setting.
2102    
2103     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2104    
2105     4. Some of the "internals" tests were previously cut out when the link size
2106        was not 2, because the output contained actual offsets. The recent new
2107        "Z" feature of pcretest means that these can be cut out, making the tests
2108        usable with all link sizes.
2109    
2110     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2111        stack recursion. This gives a massive performance boost under BSD, but just
2112        a small improvement under Linux. However, it saves one field in the frame
2113        in all cases.
2114    
2115     6. Added more features from the forthcoming Perl 5.10:
2116    
2117        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2118            recursion call. It refers to the nth most recently opened parentheses.
2119    
2120        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2121            to be opened parentheses.
2122    
2123        (c) Conditions that refer to capturing parentheses can be specified
2124            relatively, for example, (?(-2)... or (?(+3)...
2125    
2126        (d) \K resets the start of the current match so that everything before
2127            is not part of it.
2128    
2129        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2130    
2131        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2132            reference syntax.
2133    
2134        (g) (?| introduces a group in which the numbering of parentheses in each
2135            alternative starts with the same number.
2136    
2137        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2138    
2139     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2140        PCRE_INFO_JCHANGED.
2141    
2142     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2143        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2144        for detecting groups that can match an empty string.
2145    
2146     9. A pattern with a very large number of alternatives (more than several
2147        hundred) was running out of internal workspace during the pre-compile
2148        phase, where pcre_compile() figures out how much memory will be needed. A
2149        bit of new cunning has reduced the workspace needed for groups with
2150        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2151        workspace instead of running out of the 4096 that are available.
2152    
2153    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2154    
2155    11. Applied patch from Google to remove an optimization that didn't quite work.
2156        The report of the bug said:
2157    
2158          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2159          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2160          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2161    
2162    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2163        it matched the wrong number of bytes.
2164    
2165    
2166    Version 7.1 24-Apr-07
2167    ---------------------
2168    
2169     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2170        that is more "standard", making use of automake and other Autotools. There
2171        is some re-arrangement of the files and adjustment of comments consequent
2172        on this.
2173    
2174     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2175        for recursive directory scanning broke on some systems because the files
2176        are not scanned in any specific order and on different systems the order
2177        was different. A call to "sort" has been inserted into RunGrepTest for the
2178        approprate test as a short-term fix. In the longer term there may be an
2179        alternative.
2180    
2181     3. I had an email from Eric Raymond about problems translating some of PCRE's
2182        man pages to HTML (despite the fact that I distribute HTML pages, some
2183        people do their own conversions for various reasons). The problems
2184        concerned the use of low-level troff macros .br and .in. I have therefore
2185        removed all such uses from the man pages (some were redundant, some could
2186        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2187        HTML has been updated to handle .nf/.fi and to complain if it encounters
2188        .br or .in.
2189    
2190     4. Updated comments in configure.ac that get placed in config.h.in and also
2191        arranged for config.h to be included in the distribution, with the name
2192        config.h.generic, for the benefit of those who have to compile without
2193        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2194    
2195     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2196        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2197        makevp.bat for the current PCRE, using the additional files
2198        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2199    
2200     6. A Windows user reported a minor discrepancy with test 2, which turned out
2201        to be caused by a trailing space on an input line that had got lost in his
2202        copy. The trailing space was an accident, so I've just removed it.
2203    
2204     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2205        that is needed.
2206    
2207     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2208        as "const" (a) because they are and (b) because it helps the PHP
2209        maintainers who have recently made a script to detect big data structures
2210        in the php code that should be moved to the .rodata section. I remembered
2211        to update Builducptable as well, so it won't revert if ucptable.h is ever
2212        re-created.
2213    
2214     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2215        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2216        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2217        support is not required. This saves 1.5-2K of code, which is important in
2218        some applications.
2219    
2220        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2221        so as not to refer to the tables, even though these functions will never be
2222        called when UTF-8 support is disabled. Otherwise there are problems with a
2223        shared library.
2224    
2225    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2226    
2227        (a) It was defining its arguments as char * instead of void *.
2228    
2229        (b) It was assuming that all moves were upwards in memory; this was true
2230            a long time ago when I wrote it, but is no longer the case.
2231    
2232        The emulated memove() is provided for those environments that have neither
2233        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2234        is clearly not the case, as these two bugs were recently reported.
2235    
2236    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2237        and Detrail to create the HTML documentation, the .txt form of the man
2238        pages, and it removes trailing spaces from listed files. It also creates
2239        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2240        case, it wraps all the #defines with #ifndefs. This script should be run
2241        before "make dist".
2242    
2243    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2244        with Unicode property support.
2245    
2246        (a) For a maximizing quantifier, if the two different cases of the
2247            character were of different lengths in their UTF-8 codings (there are
2248            some cases like this - I found 11), and the matching function had to
2249            back up over a mixture of the two cases, it incorrectly assumed they
2250            were both the same length.
2251    
2252        (b) When PCRE was configured to use the heap rather than the stack for
2253            recursion during matching, it was not correctly preserving the data for
2254            the other case of a UTF-8 character when checking ahead for a match
2255            while processing a minimizing repeat. If the check also involved
2256            matching a wide character, but failed, corruption could cause an
2257            erroneous result when trying to check for a repeat of the original
2258            character.
2259    
2260    13. Some tidying changes to the testing mechanism:
2261    
2262        (a) The RunTest script now detects the internal link size and whether there
2263            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2264            values substituted by "configure". (The RunGrepTest script already did
2265            this for UTF-8.) The configure.ac script no longer substitutes the
2266            relevant variables.
2267    
2268        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2269            with length and offset values. This means that the output is different
2270            for different internal link sizes. Test 2 is skipped for link sizes
2271            other than 2 because of this, bypassing the problem. Unfortunately,
2272            there was also a test in test 3 (the locale tests) that used /B and
2273            failed for link sizes other than 2. Rather than cut the whole test out,
2274            I have added a new /Z option to pcretest that replaces the length and
2275            offset values with spaces. This is now used to make test 3 independent
2276            of link size. (Test 2 will be tidied up later.)
2277    
2278    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2279        segmentation fault instead of returning the appropriate error message.
2280    
2281    15. In multiline mode when the newline sequence was set to "any", the pattern
2282        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2283        This doesn't seem right; it now treats the CRLF combination as the line
2284        ending, and so does not match in that case. It's only a pattern such as ^$
2285        that would hit this one: something like ^ABC$ would have failed after \r
2286        and then tried again after \r\n.
2287    
2288    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2289        in an attempt to make files that differ only in their line terminators
2290        compare equal. This works on Linux.
2291    
2292    17. Under certain error circumstances pcregrep might try to free random memory
2293        as it exited. This is now fixed, thanks to valgrind.
2294    
2295    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2296        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2297        was because its rules for how to advance for /g after matching an empty
2298        string at the end of a line did not allow for this case. They now check for
2299        it specially.
2300    
2301    20. pcretest is supposed to handle patterns and data of any length, by
2302        extending its buffers when necessary. It was getting this wrong when the
2303        buffer for a data line had to be extended.
2304    
2305    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2306        CRLF as a newline sequence.
2307    
2308    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2309        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2310        I have nevertheless tidied it up.
2311    
2312    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2313    
2314    24. Added a man page for pcre-config.
2315    
2316    
2317    Version 7.0 19-Dec-06
2318    ---------------------
2319    
2320     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2321        moving to gcc 4.1.1.
2322    
2323     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2324        sys/time.h, which is documented as needed for this function. It doesn't
2325        seem to matter on Linux, but it showed up on some releases of OS X.
2326    
2327     3. It seems that there are systems where bytes whose values are greater than
2328        127 match isprint() in the "C" locale. The "C" locale should be the
2329        default when a C program starts up. In most systems, only ASCII printing
2330        characters match isprint(). This difference caused the output from pcretest
2331        to vary, making some of the tests fail. I have changed pcretest so that:
2332    
2333        (a) When it is outputting text in the compiled version of a pattern, bytes
2334            other than 32-126 are always shown as hex escapes.
2335    
2336        (b) When it is outputting text that is a matched part of a subject string,
2337            it does the same, unless a different locale has been set for the match
2338            (using the /L modifier). In this case, it uses isprint() to decide.
2339    
2340     4. Fixed a major bug that caused incorrect computation of the amount of memory
2341        required for a compiled pattern when options that changed within the
2342        pattern affected the logic of the preliminary scan that determines the
2343        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2344        that the computed length was too small. The symptoms of this bug were
2345        either the PCRE error "internal error: code overflow" from pcre_compile(),
2346        or a glibc crash with a message such as "pcretest: free(): invalid next
2347        size (fast)". Examples of patterns that provoked this bug (shown in
2348        pcretest format) are:
2349    
2350          /(?-x: )/x
2351          /(?x)(?-x: \s*#\s*)/
2352          /((?i)[\x{c0}])/8
2353          /(?i:[\x{c0}])/8
2354    
2355        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2356        is now done differently.
2357    
2358     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2359        wrapper classes; (b) implement a new function in the C++ scanner that is
2360        more efficient than the old way of doing things because it avoids levels of
2361        recursion in the regex matching; (c) add a paragraph to the documentation
2362        for the FullMatch() function.
2363    
2364     6. The escape sequence \n was being treated as whatever was defined as
2365        "newline". Not only was this contrary to the documentation, which states
2366        that \n is character 10 (hex 0A), but it also went horribly wrong when
2367        "newline" was defined as CRLF. This has been fixed.
2368    
2369     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2370        was being set to -1 for the "end of line" case (supposedly a value that no
2371        character can have). Though this value is never used (the check for end of
2372        line is "zero bytes in current character"), it caused compiler complaints.
2373        I've changed it to 0xffffffff.
2374    
2375     8. In pcre_version.c, the version string was being built by a sequence of
2376        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2377        string (as it is for production releases) called a macro with an empty
2378        argument. The C standard says the result of this is undefined. The gcc
2379        compiler treats it as an empty string (which was what was wanted) but it is
2380        reported that Visual C gives an error. The source has been hacked around to
2381        avoid this problem.
2382    
2383     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2384        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2385        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2386        of them did).
2387    
2388    10. Originally, pcretest opened its input and output without "b"; then I was
2389        told that "b" was needed in some environments, so it was added for release
2390        5.0 to both the input and output. (It makes no difference on Unix-like
2391        systems.) Later I was told that it is wrong for the input on Windows. I've
2392        now abstracted the modes into two macros, to make it easier to fiddle with
2393        them, and removed "b" from the input mode under Windows.
2394    
2395    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2396    
2397    12. Added -help and --help to pcretest as an official way of being reminded
2398        of the options.
2399    
2400    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2401        and pcrecpp.cc because they annoy compilers at high warning levels.
2402    
2403    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2404    
2405    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2406        scripts are not C programs :-) and which was not noticed because it works
2407        on Linux.
2408    
2409    16. pcretest is supposed to handle any length of pattern and data line (as one
2410        line or as a continued sequence of lines) by extending its input buffer if
2411        necessary. This feature was broken for very long pattern lines, leading to
2412        a string of junk being passed to pcre_compile() if the pattern was longer
2413        than about 50K.
2414    
2415    17. I have done a major re-factoring of the way pcre_compile() computes the
2416        amount of memory needed for a compiled pattern. Previously, there was code
2417        that made a preliminary scan of the pattern in order to do this. That was
2418        OK when PCRE was new, but as the facilities have expanded, it has become
2419        harder and harder to keep it in step with the real compile phase, and there
2420        have been a number of bugs (see for example, 4 above). I have now found a
2421        cunning way of running the real compile function in a "fake" mode that
2422        enables it to compute how much memory it would need, while actually only
2423        ever using a few hundred bytes of working memory and without too many
2424        tests of the mode. This should make future maintenance and development
2425        easier. A side effect of this work is that the limit of 200 on the nesting
2426        depth of parentheses has been removed (though this was never a serious
2427        limitation, I suspect). However, there is a downside: pcre_compile() now
2428        runs more slowly than before (30% or more, depending on the pattern). I
2429        hope this isn't a big issue. There is no effect on runtime performance.
2430    
2431    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2432        newline (only possible for the last line of a file) and it was a
2433        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2434    
2435    19. Added additional timing features to pcretest. (1) The -tm option now times
2436        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2437        separate command line item, by a number that specifies the number of
2438        repeats to use when timing. The default is 50000; this gives better
2439        precision, but takes uncomfortably long for very large patterns.
2440    
2441    20. Extended pcre_study() to be more clever in cases where a branch of a
2442        subpattern has no definite first character. For example, (a*|b*)[cd] would
2443        previously give no result from pcre_study(). Now it recognizes that the
2444        first character must be a, b, c, or d.
2445    
2446    21. There was an incorrect error "recursive call could loop indefinitely" if
2447        a subpattern (or the entire pattern) that was being tested for matching an
2448        empty string contained only one non-empty item after a nested subpattern.
2449        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2450        incorrectly, because the \d was being skipped in the check.
2451    
2452    22. The pcretest program now has a new pattern option /B and a command line
2453        option -b, which is equivalent to adding /B to every pattern. This causes
2454        it to show the compiled bytecode, without the additional information that
2455        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2456        is the same as /B/I).
2457    
2458    23. A new optimization is now able automatically to treat some sequences such
2459        as a*b as a*+b. More specifically, if something simple (such as a character
2460        or a simple class like \d) has an unlimited quantifier, and is followed by
2461        something that cannot possibly match the quantified thing, the quantifier
2462        is automatically "possessified".
2463    
2464    24. A recursive reference to a subpattern whose number was greater than 39
2465        went wrong under certain circumstances in UTF-8 mode. This bug could also
2466        have affected the operation of pcre_study().
2467    
2468    25. Realized that a little bit of performance could be had by replacing
2469        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2470    
2471    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2472    
2473    27. Possessive quantifiers such as a++ were previously implemented by turning
2474        them into atomic groups such as ($>a+). Now they have their own opcodes,
2475        which improves performance. This includes the automatically created ones
2476        from 23 above.
2477    
2478    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2479        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2480        the first matched character to be a colon. This applied both to named and
2481        numbered groups.
2482    
2483    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2484    
2485    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2486        building PCRE on Windows easier, so I have included it in the distribution.
2487    
2488    31. There is now a check in pcretest against a ridiculously large number being
2489        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2490        loop, the loop is abandoned.
2491    
2492    32. Forward references to subpatterns in conditions such as (?(2)...) where
2493        subpattern 2 is defined later cause pcre_compile() to search forwards in
2494        the pattern for the relevant set of parentheses. This search went wrong
2495        when there were unescaped parentheses in a character class, parentheses
2496        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2497    
2498    33. "Subroutine" calls and backreferences were previously restricted to
2499        referencing subpatterns earlier in the regex. This restriction has now
2500        been removed.
2501    
2502    34. Added a number of extra features that are going to be in Perl 5.10. On the
2503        whole, these are just syntactic alternatives for features that PCRE had
2504        previously implemented using the Python syntax or my own invention. The
2505        other formats are all retained for compatibility.
2506    
2507        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2508            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2509            also .NET compatible.
2510    
2511        (b) A recursion or subroutine call to a named group can now be defined as
2512            (?&name) as well as (?P>name).
2513    
2514        (c) A backreference to a named group can now be defined as \k<name> or
2515            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2516            5.10, are also .NET compatible.
2517    
2518        (d) A conditional reference to a named group can now use the syntax
2519            (?(<name>) or (?('name') as well as (?(name).
2520    
2521        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2522            groups (named and numbered) that are never evaluated inline, but can be
2523            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2524            is always false. There may be only one alternative in such a group.
2525    
2526        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2527            as the simple (?(R). The condition is true only if the most recent
2528            recursion is that of the given number or name. It does not search out
2529            through the entire recursion stack.
2530    
2531        (g) The escape \gN or \g{N} has been added, where N is a positive or
2532            negative number, specifying an absolute or relative reference.
2533    
2534    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2535        some "unreachable code" warnings.
2536    
2537    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2538        things, this adds five new scripts.
2539    
2540    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2541        There were also incompatibilities regarding the handling of \Q..\E inside
2542        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2543        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2544    
2545    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2546        matches an empty string, and forcibly breaks the loop. There were bugs in
2547        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2548        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2549        separate and independent bugs (that affected different cases) have been
2550        fixed.
2551    
2552    39. Refactored the code to abolish the use of different opcodes for small
2553        capturing bracket numbers. This is a tidy that I avoided doing when I
2554        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2555        The new approach is not only tidier, it makes it possible to reduce the
2556        memory needed to fix the previous bug (38).
2557    
2558    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2559        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2560        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2561        mode.
2562    
2563    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2564        report.
2565    
2566    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2567        copy construction and assignment in the C++ wrapper.
2568    
2569    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2570        couple of bugs that could have given wrong results in the "--newline=crlf"
2571        case.
2572    
2573    44. Added a number of casts and did some reorganization of signed/unsigned int
2574        variables following suggestions from Dair Grant. Also renamed the variable
2575        "this" as "item" because it is a C++ keyword.
2576    
2577    45. Arranged for dftables to add
2578    
2579          #include "pcre_internal.h"
2580    
2581        to pcre_chartables.c because without it, gcc 4.x may remove the array
2582        definition from the final binary if PCRE is built into a static library and
2583        dead code stripping is activated.
2584    
2585    46. For an unanchored pattern, if a match attempt fails at the start of a
2586        newline sequence, and the newline setting is CRLF or ANY, and the next two
2587        characters are CRLF, advance by two characters instead of one.
2588    
2589    
2590    Version 6.7 04-Jul-06
2591    ---------------------
2592    
2593     1. In order to handle tests when input lines are enormously long, pcretest has
2594        been re-factored so that it automatically extends its buffers when
2595        necessary. The code is crude, but this _is_ just a test program. The
2596        default size has been increased from 32K to 50K.
2597    
2598     2. The code in pcre_study() was using the value of the re argument before
2599        testing it for NULL. (Of course, in any sensible call of the function, it
2600        won't be NULL.)
2601    
2602     3. The memmove() emulation function in pcre_internal.h, which is used on
2603        systems that lack both memmove() and bcopy() - that is, hardly ever -
2604        was missing a "static" storage class specifier.
2605    
2606     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2607        containing an extended class (one that cannot be represented by a bitmap
2608        because it contains high-valued characters or Unicode property items, e.g.
2609        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2610        pattern, but PCRE should not loop if you do not (it no longer does).
2611        [Detail: two cases were found: (a) a repeated subpattern containing an
2612        extended class; (b) a recursive reference to a subpattern that followed a
2613        previous extended class. It wasn't skipping over the extended class
2614        correctly when UTF-8 mode was not set.]
2615    
2616     5. A negated single-character class was not being recognized as fixed-length
2617        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2618        compile error "lookbehind assertion is not fixed length".
2619    
2620     6. The RunPerlTest auxiliary script was showing an unexpected difference
2621        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2622        write a Perl script that can interpret lines of an input file either as
2623        byte characters or as UTF-8, which is what "perltest" was being required to
2624        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2625        can't do is switch easily at run time between having the "use utf8;" pragma
2626        or not. In the end, I fudged it by using the RunPerlTest script to insert
2627        "use utf8;" explicitly for the UTF-8 tests.
2628    
2629     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2630        the end of the subject string, contrary to the documentation and to what
2631        Perl does. This was true of both matching functions. Now it matches only at
2632        the start of the subject and immediately after *internal* newlines.
2633    
2634     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2635        a pointer to an int instead of a pointer to an unsigned long int. This
2636        caused problems on 64-bit systems.
2637    
2638     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2639        instance of the 'standard' template library not being so standard".
2640    
2641    10. There was no check on the number of named subpatterns nor the maximum
2642        length of a subpattern name. The product of these values is used to compute
2643        the size of the memory block for a compiled pattern. By supplying a very
2644        long subpattern name and a large number of named subpatterns, the size
2645        computation could be caused to overflow. This is now prevented by limiting
2646        the length of names to 32 characters, and the number of named subpatterns
2647        to 10,000.
2648    
2649    11. Subpatterns that are repeated with specific counts have to be replicated in
2650        the compiled pattern. The size of memory for this was computed from the
2651        length of the subpattern and the repeat count. The latter is limited to
2652        65535, but there was no limit on the former, meaning that integer overflow
2653        could in principle occur. The compiled length of a repeated subpattern is
2654        now limited to 30,000 bytes in order to prevent this.
2655    
2656    12. Added the optional facility to have named substrings with the same name.
2657    
2658    13. Added the ability to use a named substring as a condition, using the
2659        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2660        are numbers (not recommended). Forward references are permitted.
2661    
2662    14. Added forward references in named backreferences (if you see what I mean).
2663    
2664    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2665        pattern could run off the end of the subject. For example, the pattern
2666        "(?s)(.{1,5})"8 did this with the subject "ab".
2667    
2668    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2669        PCRE_CASELESS was set when matching characters that were quantified with ?
2670        or *.
2671    
2672    17. A character class other than a single negated character that had a minimum
2673        but no maximum quantifier - for example [ab]{6,} - was not handled
2674        correctly by pce_dfa_exec(). It would match only one character.
2675    
2676    18. A valid (though odd) pattern that looked like a POSIX character
2677        class but used an invalid character after [ (for example [[,abc,]]) caused
2678        pcre_compile() to give the error "Failed: internal error: code overflow" or
2679        in some cases to crash with a glibc free() error. This could even happen if
2680        the pattern terminated after [[ but there just happened to be a sequence of
2681        letters, a binary zero, and a closing ] in the memory that followed.
2682    
2683    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2684        over the years. Originally (before any Unicode support), just the bottom 8
2685        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2686        output from "man perlunicode" includes this:
2687    
2688          The regular expression compiler produces polymorphic opcodes.  That
2689          is, the pattern adapts to the data and automatically switches to
2690          the Unicode character scheme when presented with Unicode data--or
2691          instead uses a traditional byte scheme when presented with byte
2692          data.
2693    
2694        Sadly, a wide octal escape does not cause a switch, and in a string with
2695        no other multibyte characters, these octal escapes are treated as before.
2696        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2697        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2698        Unicode string.
2699    
2700        I have not perpetrated such confusion in PCRE. Up till now, it took just
2701        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2702        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2703        translate to the appropriate multibyte character.
2704    
2705    29. Applied some refactoring to reduce the number of warnings from Microsoft
2706        and Borland compilers. This has included removing the fudge introduced
2707        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2708        a warning about an unused variable.
2709    
2710    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2711        characters since release 4.0, because Perl (from release 5.004) does not.
2712        [Or at least, is documented not to: some releases seem to be in conflict
2713        with the documentation.] However, when a pattern was studied with
2714        pcre_study() and all its branches started with \s, PCRE still included VT
2715        as a possible starting character. Of course, this did no harm; it just
2716        caused an unnecessary match attempt.
2717    
2718    22. Removed a now-redundant internal flag bit that recorded the fact that case
2719        dependency changed within the pattern. This was once needed for "required
2720        byte" processing, but is no longer used. This recovers a now-scarce options
2721        bit. Also moved the least significant internal flag bit to the most-
2722        significant bit of the word, which was not previously used (hangover from
2723        the days when it was an int rather than a uint) to free up another bit for
2724        the future.
2725    
2726    23. Added support for CRLF line endings as well as CR and LF. As well as the
2727        default being selectable at build time, it can now be changed at runtime
2728        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2729        specify that it is scanning data with non-default line endings.
2730    
2731    24. Changed the definition of CXXLINK to make it agree with the definition of
2732        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2733    
2734    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2735        recursions. This makes a big different to stack usage for some patterns.
2736    
2737    26. If a subpattern containing a named recursion or subroutine reference such
2738        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2739        the space required for the compiled pattern went wrong and gave too small a
2740        value. Depending on the environment, this could lead to "Failed: internal
2741        error: code overflow at offset 49" or "glibc detected double free or
2742        corruption" errors.
2743    
2744    27. Applied patches from Google (a) to support the new newline modes and (b) to
2745        advance over multibyte UTF-8 characters in GlobalReplace.
2746    
2747    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2748        difference for some implementation of PCRE in some Windows version.
2749    
2750    29. Added some extra testing facilities to pcretest:
2751    
2752        \q<number>   in a data line sets the "match limit" value
2753        \Q<number>   in a data line sets the "match recursion limt" value
2754        -S <number>  sets the stack size, where <number> is in megabytes
2755    
2756        The -S option isn't available for Windows.
2757    
2758    
2759    Version 6.6 06-Feb-06
2760    ---------------------
2761    
2762     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2763        in pcreposix.h. I have copied the definition from pcre.h.
2764    
2765     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2766        because pcre.h is no longer a built file.
2767    
2768     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2769        not normally included in the compiled code.
2770    
2771    
2772    Version 6.5 01-Feb-06
2773    ---------------------
2774    
2775     1. When using the partial match feature with pcre_dfa_exec(), it was not
2776        anchoring the second and subsequent partial matches at the new starting
2777        point. This could lead to incorrect results. For example, with the pattern
2778        /1234/, partially matching against "123" and then "a4" gave a match.
2779    
2780     2. Changes to pcregrep:
2781    
2782        (a) All non-match returns from pcre_exec() were being treated as failures
2783            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2784            error message is output. Some extra information is given for the
2785            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2786            probably the only errors that are likely to be caused by users (by
2787            specifying a regex that has nested indefinite repeats, for instance).
2788            If there are more than 20 of these errors, pcregrep is abandoned.
2789    
2790        (b) A binary zero was treated as data while matching, but terminated the
2791            output line if it was written out. This has been fixed: binary zeroes
2792            are now no different to any other data bytes.
2793    
2794        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2795            used to set a locale for matching. The --locale=xxxx long option has
2796            been added (no short equivalent) to specify a locale explicitly on the
2797            pcregrep command, overriding the environment variables.
2798    
2799        (d) When -B was used with -n, some line numbers in the output were one less
2800            than they should have been.
2801    
2802        (e) Added the -o (--only-matching) option.
2803    
2804        (f) If -A or -C was used with -c (count only), some lines of context were
2805            accidentally printed for the final match.
2806    
2807        (g) Added the -H (--with-filename) option.
2808    
2809        (h) The combination of options -rh failed to suppress file names for files
2810            that were found from directory arguments.
2811    
2812        (i) Added the -D (--devices) and -d (--directories) options.
2813    
2814        (j) Added the -F (--fixed-strings) option.
2815    
2816        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2817    
2818        (l) Added the --colo(u)r option.
2819    
2820        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2821            is not present by default.
2822    
2823     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2824        items such as (?R) or (?1), when the recursion could match a number of
2825        alternatives. If it matched one of the alternatives, but subsequently,
2826        outside the recursion, there was a failure, the code tried to back up into
2827        the recursion. However, because of the way PCRE is implemented, this is not
2828        possible, and the result was an incorrect result from the match.
2829    
2830        In order to prevent this happening, the specification of recursion has
2831        been changed so that all such subpatterns are automatically treated as
2832        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2833    
2834     4. I had overlooked the fact that, in some locales, there are characters for
2835        which isalpha() is true but neither isupper() nor islower() are true. In
2836        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2837        and ordfeminine) are like this. This affected the treatment of \w and \W
2838        when they appeared in character classes, but not when they appeared outside
2839        a character class. The bit map for "word" characters is now created
2840        separately from the results of isalnum() instead of just taking it from the
2841        upper, lower, and digit maps. (Plus the underscore character, of course.)
2842    
2843     5. The above bug also affected the handling of POSIX character classes such as
2844        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2845        permanent tables. Instead, the bit maps for such a class were previously
2846        created as the appropriate unions of the upper, lower, and digit bitmaps.
2847        Now they are created by subtraction from the [[:word:]] class, which has
2848        its own bitmap.
2849    
2850     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2851        It is created by subtracting the vertical space characters (\x09, \x0a,
2852        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2853        subtraction was done in the overall bitmap for a character class, meaning
2854        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2855        be recognized. This bug has been fixed.
2856    
2857     7. Patches from the folks at Google:
2858    
2859          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2860          real life, but is still worth protecting against".
2861    
2862          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2863          regular expressions".
2864    
2865          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2866          have it.
2867    
2868          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2869          "configure" and the latter not, in order to fix a problem somebody had
2870          with compiling the Arg class on HP-UX.
2871    
2872          (e) Improve the error-handling of the C++ wrapper a little bit.
2873    
2874          (f) New tests for checking recursion limiting.
2875    
2876     8. The pcre_memmove() function, which is used only if the environment does not
2877        have a standard memmove() function (and is therefore rarely compiled),
2878        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2879        returning a result (though PCRE never actually uses the result).
2880    
2881     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2882        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2883        returned instead of calling malloc() with an overflowing number that would
2884        most likely cause subsequent chaos.
2885    
2886    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2887    
2888    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2889        with this option is matched, the nmatch and pmatch options of regexec() are
2890        ignored.
2891    
2892    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2893        provided in case anyone wants to the the POSIX interface with UTF-8
2894        strings.
2895    
2896    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2897        C++ linking (needed for some HP-UX environments).
2898    
2899    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2900        (unused parameter) and in the pcre_printint() function (omitted "default"
2901        switch label when the default is to do nothing).
2902    
2903    15. Added some code to make it possible, when PCRE is compiled as a C++
2904        library, to replace subject pointers for pcre_exec() with a smart pointer
2905        class, thus making it possible to process discontinuous strings.
2906    
2907    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2908        much the same function. They were added by different people who were trying
2909        to make PCRE easy to compile on non-Unix systems. It has been suggested
2910        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2911        for compiling on Windows systems. I have therefore replaced it with
2912        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2913        defaults to "extern" for C or "extern C" for C++, which works fine on
2914        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2915        SCOPE with something explicit in config.h. In addition:
2916    
2917        (a) pcreposix.h still had just "extern" instead of either of these macros;
2918            I have replaced it with PCRE_DATA_SCOPE.
2919    
2920        (b) Functions such as _pcre_xclass(), which are internal to the library,
2921            but external in the C sense, all had PCRE_EXPORT in their definitions.
2922            This is apparently wrong for the Windows case, so I have removed it.
2923            (It makes no difference on Unix-like systems.)
2924    
2925    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2926        of recursive calls to match(). This is different to MATCH_LIMIT because
2927        that limits the total number of calls to match(), not all of which increase
2928        the depth of recursion. Limiting the recursion depth limits the amount of
2929        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2930        when PCRE is compiled, and changed at run time. A patch from Google adds
2931        this functionality to the C++ interface.
2932    
2933    18. Changes to the handling of Unicode character properties:
2934    
2935        (a) Updated the table to Unicode 4.1.0.
2936    
2937        (b) Recognize characters that are not in the table as "Cn" (undefined).
2938    
2939        (c) I revised the way the table is implemented to a much improved format
2940            which includes recognition of ranges. It now supports the ranges that
2941            are defined in UnicodeData.txt, and it also amalgamates other
2942            characters into ranges. This has reduced the number of entries in the
2943            table from around 16,000 to around 3,000, thus reducing its size
2944            considerably. I realized I did not need to use a tree structure after
2945            all - a binary chop search is just as efficient. Having reduced the
2946            number of entries, I extended their size from 6 bytes to 8 bytes to
2947            allow for more data.
2948    
2949        (d) Added support for Unicode script names via properties such as \p{Han}.
2950    
2951    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2952        matching that character.
2953    
2954    20. When matching a repeated Unicode property with a minimum greater than zero,
2955        (for example \pL{2,}), PCRE could look past the end of the subject if it
2956        reached it while seeking the minimum number of characters. This could
2957        happen only if some of the characters were more than one byte long, because
2958        there is a check for at least the minimum number of bytes.
2959    
2960    21. Refactored the implementation of \p and \P so as to be more general, to
2961        allow for more different types of property in future. This has changed the
2962        compiled form incompatibly. Anybody with saved compiled patterns that use
2963        \p or \P will have to recompile them.
2964    
2965    22. Added "Any" and "L&" to the supported property types.
2966    
2967    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2968        but give a compile time error if the value is greater than 0xff.
2969    
2970    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2971        accidentally not being installed or uninstalled.
2972    
2973    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2974        made were to insert the current release number. This seemed silly, because
2975        it made things harder for people building PCRE on systems that don't run
2976        "configure". I have turned pcre.h into a distributed file, no longer built
2977        by "configure", with the version identification directly included. There is
2978        no longer a pcre.h.in file.
2979    
2980        However, this change necessitated a change to the pcre-config script as
2981        well. It is built from pcre-config.in, and one of the substitutions was the
2982        release number. I have updated configure.ac so that ./configure now finds
2983        the release number by grepping pcre.h.
2984    
2985    26. Added the ability to run the tests under valgrind.
2986    
2987    
2988    Version 6.4 05-Sep-05
2989    ---------------------
2990    
2991     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2992        "--" to be printed when multiple files were scanned, even when none of the
2993        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2994        consider it to be a bug, and have restored the previous behaviour.
2995    
2996     2. A couple of code tidies to get rid of compiler warnings.
2997    
2998     3. The pcretest program used to cheat by referring to symbols in the library
2999        whose names begin with _pcre_. These are internal symbols that are not
3000        really supposed to be visible externally, and in some environments it is
3001        possible to suppress them. The cheating is now confined to including
3002        certain files from the library's source, which is a bit cleaner.
3003    
3004     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3005        file's purpose clearer.
3006    
3007     5. Reorganized pcre_ucp_findchar().
3008    
3009    
3010    Version 6.3 15-Aug-05
3011    ---------------------
3012    
3013     1. The file libpcre.pc.in did not have general read permission in the tarball.
3014    
3015     2. There were some problems when building without C++ support:
3016    
3017        (a) If C++ support was not built, "make install" and "make test" still
3018            tried to test it.
3019    
3020        (b) There were problems when the value of CXX was explicitly set. Some
3021            changes have been made to try to fix these, and ...
3022    
3023        (c) --disable-cpp can now be used to explicitly disable C++ support.
3024    
3025        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3026            backslash in a target when C++ was disabled. This confuses some
3027            versions of "make", apparently. Using an intermediate variable solves
3028            this. (Same for CPP_LOBJ.)
3029    
3030     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3031        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3032        necessary on certain architectures.
3033    
3034     4. Added a setting of -export-symbols-regex to the link command to remove
3035        those symbols that are exported in the C sense, but actually are local
3036        within the library, and not documented. Their names all begin with
3037        "_pcre_". This is not a perfect job, because (a) we have to except some
3038        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3039        available (and never for static libraries). I have made a note to try to
3040        find a way round (a) in the future.
3041    
3042    
3043    Version 6.2 01-Aug-05
3044    ---------------------
3045    
3046     1. There was no test for integer overflow of quantifier values. A construction
3047        such as {1111111111111111} would give undefined results. What is worse, if
3048        a minimum quantifier for a parenthesized subpattern overflowed and became
3049        negative, the calculation of the memory size went wrong. This could have
3050        led to memory overwriting.
3051    
3052     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3053    
3054     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3055        operating environments where this matters.
3056    
3057     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3058        PCRE options from within the C++ wrapper.
3059    
3060     5. Named capturing subpatterns were not being correctly counted when a pattern
3061        was compiled. This caused two problems: (a) If there were more than 100
3062        such subpatterns, the calculation of the memory needed for the whole
3063        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3064        back references of the form \12, where the number was greater than 9, were
3065        not recognized as back references, even though there were sufficient
3066        previous subpatterns.
3067    
3068     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3069        versions of gcc, e.g. 2.95.4.
3070    
3071    
3072    Version 6.1 21-Jun-05
3073    ---------------------
3074    
3075     1. There was one reference to the variable "posix" in pcretest.c that was not
3076        surrounded by "#if !defined NOPOSIX".
3077    
3078     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3079        the cross-check on the old pcre_info() function, for the benefit of the
3080        cut-down version of PCRE that is currently imported into Exim.
3081    
3082     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3083        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3084        patterns that start (?i) but I don't think that matters. The use of (?i) is
3085        just an example; this all applies to the other options as well.
3086    
3087     4. Since libtool seems to echo the compile commands it is issuing, the output
3088        from "make" can be reduced a bit by putting "@" in front of each libtool
3089        compile command.
3090    
3091     5. Patch from the folks at Google for configure.in to be a bit more thorough
3092        in checking for a suitable C++ installation before trying to compile the
3093        C++ stuff. This should fix a reported problem when a compiler was present,
3094        but no suitable headers.
3095    
3096     6. The man pages all had just "PCRE" as their title. I have changed them to
3097        be the relevant file name. I have also arranged that these names are
3098        retained in the file doc/pcre.txt, which is a concatenation in text format
3099        of all the man pages except the little individual ones for each function.
3100    
3101     7. The NON-UNIX-USE file had not been updated for the different set of source
3102        files that come with release 6. I also added a few comments about the C++
3103        wrapper.
3104    
3105    
3106    Version 6.0 07-Jun-05
3107    ---------------------
3108    
3109     1. Some minor internal re-organization to help with my DFA experiments.
3110    
3111     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3112        didn't matter for the library itself when fully configured, but did matter
3113        when compiling without UCP support, or within Exim, where the ucp files are
3114        not imported.
3115    
3116     3. Refactoring of the library code to split up the various functions into
3117        different source modules. The addition of the new DFA matching code (see
3118        below) to a single monolithic source would have made it really too
3119        unwieldy, quite apart from causing all the code to be include in a
3120        statically linked application, when only some functions are used. This is
3121        relevant even without the DFA addition now that patterns can be compiled in
3122        one application and matched in another.
3123    
3124        The downside of splitting up is that there have to be some external
3125        functions and data tables that are used internally in different modules of
3126        the library but which are not part of the API. These have all had their
3127        names changed to start with "_pcre_" so that they are unlikely to clash
3128        with other external names.
3129    
3130     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3131        a different (DFA) algorithm. Although it is slower than the original
3132        function, it does have some advantages for certain types of matching
3133        problem.
3134    
3135     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3136        including restarting after a partial match.
3137    
3138     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3139        defined when compiling for Windows was sent to me. I have put it into the
3140        code, though I have no means of testing or verifying it.
3141    
3142     7. Added the pcre_refcount() auxiliary function.
3143    
3144     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3145        match before or at the first newline in the subject string. In pcretest,
3146        the /f option on a pattern can be used to set this.
3147    
3148     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3149        would behave wrongly. This has been present in PCRE since release 4.0.
3150    
3151    10. A number of changes to the pcregrep command:
3152    
3153        (a) Refactored how -x works; insert ^(...)$ instead of setting
3154            PCRE_ANCHORED and checking the length, in preparation for adding
3155            something similar for -w.
3156    
3157        (b) Added the -w (match as a word) option.
3158    
3159        (c) Refactored the way lines are read and buffered so as to have more
3160            than one at a time available.
3161    
3162        (d) Implemented a pcregrep test script.
3163    
3164        (e) Added the -M (multiline match) option. This allows patterns to match
3165            over several lines of the subject. The buffering ensures that at least
3166            8K, or the rest of the document (whichever is the shorter) is available
3167            for matching (and similarly the previous 8K for lookbehind assertions).
3168    
3169        (f) Changed the --help output so that it now says
3170    
3171              -w, --word-regex(p)
3172    
3173            instead of two lines, one with "regex" and the other with "regexp"
3174            because that confused at least one person since the short forms are the
3175            same. (This required a bit of code, as the output is generated
3176            automatically from a table. It wasn't just a text change.)
3177    
3178        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3179            option but starts with a hyphen. Could be a pattern or a path name
3180            starting with a hyphen, for instance.
3181    
3182        (h) "-" can be given as a file name to represent stdin.
3183    
3184        (i) When file names are being printed, "(standard input)" is used for
3185            the standard input, for compatibility with GNU grep. Previously
3186            "<stdin>" was used.
3187    
3188        (j) The option --label=xxx can be used to supply a name to be used for
3189            stdin when file names are being printed. There is no short form.
3190    
3191        (k) Re-factored the options decoding logic because we are going to add
3192            two more options that take data. Such options can now be given in four
3193            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3194    
3195        (l) Added the -A, -B, and -C options for requesting that lines of context
3196            around matches be printed.
3197    
3198        (m) Added the -L option to print the names of files that do not contain
3199            any matching lines, that is, the complement of -l.
3200    
3201        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3202            continue to scan other files.
3203    
3204        (o) The -s option was incorrectly implemented. For compatibility with other
3205            greps, it now suppresses the error message for a non-existent or non-
3206            accessible file (but not the return code). There is a new option called
3207            -q that suppresses the output of matching lines, which was what -s was
3208            previously doing.
3209    
3210        (p) Added --include and --exclude options to specify files for inclusion
3211            and exclusion when recursing.
3212    
3213    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3214        Hopefully, it now does.
3215    
3216    12. Missing cast in pcre_study().
3217    
3218    13. Added an "uninstall" target to the makefile.
3219    
3220    14. Replaced "extern" in the function prototypes in Makefile.in with
3221        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3222        world, but is set differently for Windows.
3223    
3224    15. Added a second compiling function called pcre_compile2(). The only
3225        difference is that it has an extra argument, which is a pointer to an
3226        integer error code. When there is a compile-time failure, this is set
3227        non-zero, in addition to the error test pointer being set to point to an
3228        error message. The new argument may be NULL if no error number is required
3229        (but then you may as well call pcre_compile(), which is now just a
3230        wrapper). This facility is provided because some applications need a
3231        numeric error indication, but it has also enabled me to tidy up the way
3232        compile-time errors are handled in the POSIX wrapper.
3233    
3234    16. Added VPATH=.libs to the makefile; this should help when building with one
3235        prefix path and installing with another. (Or so I'm told by someone who
3236        knows more about this stuff than I do.)
3237    
3238    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3239        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3240        match everything, including newlines. This is not POSIX-compatible, but
3241        somebody wanted the feature. From pcretest it can be activated by using
3242        both the P and the s flags.
3243    
3244    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3245    
3246    19. libpcre.pc was being incorrectly installed as executable.
3247    
3248    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3249        it now also looks for '\r' so that it will work unmodified on Windows.
3250    
3251    21. Added Google's contributed C++ wrapper to the distribution.
3252    
3253    22. Added some untidy missing memory free() calls in pcretest, to keep
3254        Electric Fence happy when testing.
3255    
3256    
3257    
3258    Version 5.0 13-Sep-04
3259    ---------------------
3260    
3261     1. Internal change: literal characters are no longer packed up into items
3262        containing multiple characters in a single byte-string. Each character
3263        is now matched using a separate opcode. However, there may be more than one
3264        byte in the character in UTF-8 mode.
3265    
3266     2. The pcre_callout_block structure has two new fields: pattern_position and
3267        next_item_length. These contain the offset in the pattern to the next match
3268        item, and its length, respectively.
3269    
3270     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3271        insertion of callouts before each pattern item. Added the /C option to
3272        pcretest to make use of this.
3273    
3274     4. On the advice of a Windows user, the lines
3275    
3276          #if defined(_WIN32) || defined(WIN32)
3277          _setmode( _fileno( stdout ), 0x8000 );
3278          #endif  /* defined(_WIN32) || defined(WIN32) */
3279    
3280        have been added to the source of pcretest. This apparently does useful
3281        magic in relation to line terminators.
3282    
3283     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3284        for the benefit of those environments where the "b" makes a difference.
3285    
3286     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3287        to know about it. I have put a hack into configure.in that adds in code
3288        to set GCC=yes if CC=icc. This seems to end up at a point in the
3289        generated configure script that is early enough to affect the setting of
3290        compiler options, which is what is needed, but I have no means of testing
3291        whether it really works. (The user who reported this had patched the
3292        generated configure script, which of course I cannot do.)
3293    
3294        LATER: After change 22 below (new libtool files), the configure script
3295        seems to know about icc (and also ecc). Therefore, I have commented out
3296        this hack in configure.in.
3297    
3298     7. Added support for pkg-config (2 patches were sent in).
3299    
3300     8. Negated POSIX character classes that used a combination of internal tables
3301        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3302        [[:^ascii]]. Typically, they would match almost any characters. The other
3303        POSIX classes were not broken in this way.
3304    
3305     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3306        to find the match, as PCRE was deluded into thinking that the match had to
3307        start at the start point or following a newline. The same bug applied to
3308        patterns with negative forward assertions or any backward assertions
3309        preceding ".*" at the start, unless the pattern required a fixed first
3310        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3311    
3312    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3313        starting at the last subject character, bytes beyond the end of the subject
3314        string were read.
3315    
3316    11. Renamed the variable "class" as "classbits" to make life easier for C++
3317        users. (Previously there was a macro definition, but it apparently wasn't
3318        enough.)
3319    
3320    12. Added the new field "tables" to the extra data so that tables can be passed
3321        in at exec time, or the internal tables can be re-selected. This allows
3322        a compiled regex to be saved and re-used at a later time by a different
3323        program that might have everything at different addresses.
3324    
3325    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3326        -R library as well as a -L library.
3327    
3328    14. The debugging options of pcretest (-d on the command line or D on a
3329        pattern) showed incorrect output for anything following an extended class
3330        that contained multibyte characters and which was followed by a quantifier.
3331    
3332    15. Added optional support for general category Unicode character properties
3333        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3334        support. It adds about 90K to the size of the library. The meanings of the
3335        inbuilt class escapes such as \d and \s have NOT been changed.
3336    
3337    16. Updated pcredemo.c to include calls to free() to release the memory for the
3338        compiled pattern.
3339    
3340    17. The generated file chartables.c was being created in the source directory
3341        instead of in the building directory. This caused the build to fail if the
3342        source directory was different from the building directory, and was
3343        read-only.
3344    
3345    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3346        file. No doubt somebody will tell me if they don't make sense... Also added
3347        Dan Mooney's comments about building on OpenVMS.
3348    
3349    19. Added support for partial matching via the PCRE_PARTIAL option for
3350        pcre_exec() and the \P data escape in pcretest.
3351    
3352    20. Extended pcretest with 3 new pattern features:
3353    
3354        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3355              write the compiled pattern to the file whose name is "rest-of-line".
3356              This is a straight binary dump of the data, with the saved pointer to
3357              the character tables forced to be NULL. The study data, if any, is
3358              written too. After writing, pcretest reads a new pattern.
3359    
3360        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3361              compiled pattern from the given file. There must not be any
3362              occurrences of "<" in the file name (pretty unlikely); if there are,
3363              pcretest will instead treat the initial "<" as a pattern delimiter.
3364              After reading in the pattern, pcretest goes on to read data lines as
3365              usual.
3366    
3367        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3368              and 16-bit fields in a compiled pattern, to simulate a pattern that
3369              was compiled on a host of opposite endianness.
3370    
3371    21. The pcre-exec() function can now cope with patterns that were compiled on
3372        hosts of opposite endianness, with this restriction:
3373    
3374          As for any compiled expression that is saved and used later, the tables
3375          pointer field cannot be preserved; the extra_data field in the arguments
3376          to pcre_exec() should be used to pass in a tables address if a value
3377          other than the default internal tables were used at compile time.
3378    
3379    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3380        now diagnosed as an error. Previously, most of the time, a negative number
3381        would have been treated as zero, but if in addition "ovector" was passed as
3382        NULL, a crash could occur.
3383    
3384    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3385        new versions from the libtool 1.5 distribution (the last one is a copy of
3386        a file called libtool.m4). This seems to have fixed the need to patch
3387        "configure" to support Darwin 1.3 (which I used to do). However, I still
3388        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3389        workstation).
3390    
3391    24. Changed the PCRE licence to be the more standard "BSD" licence.
3392    
3393    
3394    Version 4.5 01-Dec-03
3395    ---------------------
3396    
3397     1. There has been some re-arrangement of the code for the match() function so
3398        that it can be compiled in a version that does not call itself recursively.
3399        Instead, it keeps those local variables that need separate instances for
3400        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3401        needs to "recurse". Keeping track of where control must go is done by means
3402        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3403        hide most of the details from the main code, and operates only if
3404        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3405        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3406        operating.
3407    
3408        To make it easier for callers to provide specially tailored get/free
3409        functions for this usage, two new functions, pcre_stack_malloc, and
3410        pcre_stack_free, are used. They are always called in strict stacking order,
3411        and the size of block requested is always the same.
3412    
3413        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3414        PCRE has been compiled to use the stack or the heap for recursion. The
3415        -C option of pcretest uses this to show which version is compiled.
3416    
3417        A new data escape \S, is added to pcretest; it causes the amounts of store
3418        obtained and freed by both kinds of malloc/free at match time to be added
3419        to the output.
3420    
3421     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3422        what's available on my current Linux desktop machine.
3423    
3424     3. When matching a UTF-8 string, the test for a valid string at the start has
3425        been extended. If start_offset is not zero, PCRE now checks that it points
3426        to a byte that is the start of a UTF-8 character. If not, it returns
3427        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3428        this is necessary because there may be backward assertions in the pattern.
3429        When matching the same subject several times, it may save resources to use
3430        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3431    
3432     4. The code for checking the validity of UTF-8 strings has been tightened so
3433        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3434        containing "overlong sequences".
3435    
3436     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3437        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3438        should have been "&", but it just so happened that all the cases this let
3439        through by mistake were picked up later in the function.
3440    
3441     6. I had used a variable called "isblank" - this is a C99 function, causing
3442        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3443    
3444     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3445        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3446        so the version is shown; (c) stop "make test" echoing "./RunTest".
3447    
3448     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3449    
3450     9. The prototype for memmove() for systems that don't have it was using
3451        size_t, but the inclusion of the header that defines size_t was later. I've
3452        moved the #includes for the C headers earlier to avoid this.
3453    
3454    10. Added some adjustments to the code to make it easier to compiler on certain
3455        special systems:
3456    
3457          (a) Some "const" qualifiers were missing.
3458          (b) Added the macro EXPORT before all exported functions; by default this
3459              is defined to be empty.
3460          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3461              that it reads its output file name as an argument instead of writing
3462              to the standard output and assuming this can be redirected.
3463    
3464    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3465        class containing characters with values greater than 255, PCRE compilation
3466        went into a loop.
3467    
3468    12. A recursive reference to a subpattern that was within another subpattern
3469        that had a minimum quantifier of zero caused PCRE to crash. For example,
3470        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3471        recursion. If the recursively-called subpattern itself had a zero repeat,
3472        that was OK.
3473    
3474    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3475        buffer into which it was copied (for escape processing) was still set at
3476        1024, so long lines caused crashes.
3477    
3478    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3479        "internal error: code overflow...". This applied to any character class
3480        that was followed by a possessive quantifier.
3481    
3482    15. Modified the Makefile to add libpcre.la as a prerequisite for
3483        libpcreposix.la because I was told this is needed for a parallel build to
3484        work.
3485    
3486    16. If a pattern that contained .* following optional items at the start was
3487        studied, the wrong optimizing data was generated, leading to matching
3488        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3489        matching string must start with a or b or c. The correct conclusion for
3490        this pattern is that a match can start with any character.
3491    
3492    
3493    Version 4.4 13-Aug-03
3494    ---------------------
3495    
3496     1. In UTF-8 mode, a character class containing characters with values between
3497        127 and 255 was not handled correctly if the compiled pattern was studied.
3498        In fixing this, I have also improved the studying algorithm for such
3499        classes (slightly).
3500    
3501     2. Three internal functions had redundant arguments passed to them. Removal
3502        might give a very teeny performance improvement.
3503    
3504     3. Documentation bug: the value of the capture_top field in a callout is *one
3505        more than* the number of the hightest numbered captured substring.
3506    
3507     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3508        in incorrectly linking with a previously installed version. They now link
3509        explicitly with libpcre.la.
3510    
3511     5. configure.in no longer needs to recognize Cygwin specially.
3512    
3513     6. A problem in pcre.in for Windows platforms is fixed.
3514    
3515     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3516        pcretest, it used to include the size of the study block as part of its
3517        output. Unfortunately, the structure contains a field that has a different
3518        size on different hardware architectures. This meant that the tests that
3519        showed this size failed. As the block is currently always of a fixed size,
3520        this information isn't actually particularly useful in pcretest output, so
3521        I have just removed it.
3522    
3523     8. Three pre-processor statements accidentally did not start in column 1.
3524        Sadly, there are *still* compilers around that complain, even though
3525        standard C has not required this for well over a decade. Sigh.
3526    
3527     9. In pcretest, the code for checking callouts passed small integers in the
3528        callout_data field, which is a void * field. However, some picky compilers
3529        complained about the casts involved for this on 64-bit systems. Now
3530        pcretest passes the address of the small integer instead, which should get
3531        rid of the warnings.
3532    
3533    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3534        both compile and run time, and gives an error if an invalid UTF-8 sequence
3535        is found. There is a option for disabling this check in cases where the
3536        string is known to be correct and/or the maximum performance is wanted.
3537    
3538    11. In response to a bug report, I changed one line in Makefile.in from
3539    
3540            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3541        to
3542            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3543    
3544        to look similar to other lines, but I have no way of telling whether this
3545        is the right thing to do, as I do not use Windows. No doubt I'll get told
3546        if it's wrong...
3547    
3548    
3549    Version 4.3 21-May-03
3550    ---------------------
3551    
3552    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3553       Makefile.
3554    
3555    2. Some refactoring to improve the quality of the code:
3556    
3557       (i)   The utf8_table... variables are now declared "const".
3558    
3559       (ii)  The code for \cx, which used the "case flipping" table to upper case
3560             lower case letters, now just substracts 32. This is ASCII-specific,
3561             but the whole concept of \cx is ASCII-specific, so it seems
3562             reasonable.
3563    
3564       (iii) PCRE was using its character types table to recognize decimal and
3565             hexadecimal digits in the pattern. This is silly, because it handles
3566             only 0-9, a-f, and A-F, but the character types table is locale-
3567             specific, which means strange things might happen. A private
3568             table is now used for this - though it costs 256 bytes, a table is
3569             much faster than multiple explicit tests. Of course, the standard
3570             character types table is still used for matching digits in subject
3571             strings against \d.
3572    
3573       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3574             ending in _t are). So I've renamed it as ESC_tee.
3575    
3576    3. The first argument for regexec() in the POSIX wrapper should have been
3577       defined as "const".
3578    
3579    4. Changed pcretest to use malloc() for its buffers so that they can be
3580       Electric Fenced for debugging.
3581    
3582    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3583       to read one or more bytes before the start of the subject string. Often this
3584       had no effect on PCRE's behaviour, but in some circumstances it could
3585       provoke a segmentation fault.
3586    
3587    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3588       to try to read one or more bytes before the start of the subject string.
3589    
3590    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3591       UTF-8 support could misbehave in various ways if the subject string
3592       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3593       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3594       back over UTF-8 characters.)
3595    
3596    
3597    Version 4.2 14-Apr-03
3598    ---------------------
3599    
3600    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3601    
3602    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3603         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3604         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3605         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3606         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3607           and BUILD_EXEEXT
3608         Note: automatic setting of the BUILD variables is not yet working
3609         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3610           compile-time but not at link-time
3611         [LINK]: use for linking executables only
3612         make different versions for Windows and non-Windows
3613         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3614           libraries
3615         [LINK_FOR_BUILD]: new variable
3616         [OBJEXT]: use throughout
3617         [EXEEXT]: use throughout
3618         <winshared>: new target
3619         <wininstall>: new target
3620         <dftables.o>: use native compiler
3621         <dftables>: use native linker
3622         <install>: handle Windows platform correctly
3623         <clean>: ditto
3624         <check>: ditto
3625         copy DLL to top builddir before testing
3626    
3627       As part of these changes, -no-undefined was removed again. This was reported
3628       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3629       in any case.
3630    
3631    3. Some tidies to get rid of compiler warnings:
3632    
3633       . In the match_data structure, match_limit was an unsigned long int, whereas
3634         match_call_count was an int. I've made them both unsigned long ints.
3635    
3636       . In pcretest the fact that a const uschar * doesn't automatically cast to
3637         a void * provoked a warning.
3638    
3639       . Turning on some more compiler warnings threw up some "shadow" variables
3640         and a few more missing casts.
3641    
3642    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3643       option, a class that contained a single character with a value between 128
3644       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3645    
3646    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3647       option, a class that contained several characters, but with at least one
3648       whose value was between 128 and 255 caused PCRE to crash.
3649    
3650    
3651    Version 4.1 12-Mar-03
3652    ---------------------
3653    
3654    1. Compiling with gcc -pedantic found a couple of places where casts were
3655    needed, and a string in dftables.c that was longer than standard compilers are
3656    required to support.
3657    
3658    2. Compiling with Sun's compiler found a few more places where the code could
3659    be tidied up in order to avoid warnings.
3660    
3661    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3662    first of these names is deprecated in the latest Autoconf in favour of the name
3663    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3664    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3665    analogy I have changed it to CFLAGS_FOR_BUILD.
3666    
3667    4. Added -no-undefined to the linking command in the Makefile, because this is
3668    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3669    linking step for the pcreposix library.
3670    
3671    5. PCRE was failing to diagnose the case of two named groups with the same
3672    name.
3673    
3674    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3675    literal character that is needed in the subject for a match, and scans along to
3676    ensure that it is present before embarking on the full matching process. This
3677    saves time in cases of nested unlimited repeats that are never going to match.
3678    Problem: the scan can take a lot of time if the subject is very long (e.g.
3679    megabytes), thus penalizing straightforward matches. It is now done only if the
3680    amount of subject to be scanned is less than 1000 bytes.
3681    
3682    7. A lesser problem with the same optimization is that it was recording the
3683    first character of an anchored pattern as "needed", thus provoking a search
3684    right along the subject, even when the first match of the pattern was going to
3685    fail. The "needed" character is now not set for anchored patterns, unless it
3686    follows something in the pattern that is of non-fixed length. Thus, it still
3687    fulfils its original purpose of finding quick non-matches in cases of nested
3688    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3689    
3690    
3691    Version 4.0 17-Feb-03
3692    ---------------------
3693    
3694    1. If a comment in an extended regex that started immediately after a meta-item
3695    extended to the end of string, PCRE compiled incorrect data. This could lead to
3696    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3697    
3698    2. Moved to autoconf 2.53 and libtool 1.4.2.
3699    
3700    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3701    the special perltest8 script is no longer needed - all the tests can be run
3702    from a single perltest script.
3703    
3704    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3705    by \s. It has now been removed in PCRE. This means it isn't recognized as
3706    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3707    class [:space:] *does* include VT, thereby creating a mess.
3708    
3709    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3710    space and tab.
3711    
3712    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3713    its new features into the main test script, reducing the number of scripts.
3714    
3715    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3716    were backward compatible, and made the (?i) apply to the whole pattern, as if
3717    /i were given. Now it behaves more logically, and applies the option setting
3718    only to what follows. PCRE has been changed to follow suit. However, if it
3719    finds options settings right at the start of the pattern, it extracts them into
3720    the global options, as before. Thus, they show up in the info data.
3721    
3722    8. Added support for the \Q...\E escape sequence. Characters in between are
3723    treated as literals. This is slightly different from Perl in that $ and @ are
3724    also handled as literals inside the quotes. In Perl, they will cause variable
3725    interpolation. Note the following examples:
3726    
3727        Pattern            PCRE matches      Perl matches
3728    
3729        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3730        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3731        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3732    
3733    For compatibility with Perl, \Q...\E sequences are recognized inside character
3734    classes as well as outside them.
3735    
3736    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3737    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3738    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3739    signed/unsigned warnings.
3740    
3741    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3742    option for pcretest, so I've replaced it by a simple function that does just
3743    that job.
3744    
3745    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3746    "pcregrep -".
3747    
3748    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3749    Java package. This provides some syntactic sugar for simple cases of what my
3750    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3751    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3752    item, you can use this simplified notation. Note that only makes sense with
3753    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3754    greediness, whatever the setting of the PCRE_UNGREEDY option.
3755    
3756    13. A change of greediness default within a pattern was not taking effect at
3757    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3758    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3759    was abstracted outside.
3760    
3761    14. PCRE now supports the \G assertion. It is true when the current matching
3762    position is at the start point of the match. This differs from \A when the
3763    starting offset is non-zero. Used with the /g option of pcretest (or similar
3764    code), it works in the same way as it does for Perl's /g option. If all
3765    alternatives of a regex begin with \G, the expression is anchored to the start
3766    match position, and the "anchored" flag is set in the compiled expression.
3767    
3768    15. Some bugs concerning the handling of certain option changes within patterns
3769    have been fixed. These applied to options other than (?ims). For example,
3770    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3771    been the other way round. Some of this was related to change 7 above.
3772    
3773    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3774    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3775    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3776    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3777    
3778    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3779    mode. Unlike ".", it always matches newline, whatever the setting of
3780    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3781    assertions. Perl allows it, but it doesn't (in general) work because it can't
3782    calculate the length of the lookbehind. At least, that's the case for Perl
3783    5.8.0 - I've been told they are going to document that it doesn't work in
3784    future.
3785    
3786    18. Added an error diagnosis for escapes that PCRE does not support: these are
3787