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

Diff of /code/trunk/ChangeLog

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

revision 65 by nigel, Sat Feb 24 21:40:08 2007 UTC revision 1242 by zherczeg, Sat Jan 26 17:51:43 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.33 xx-xxxx-201x
5    -------------------------
6    
7    1.  Added 'U' to some constants that are compared to unsigned integers, to
8        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9        variables that are added to signed variables, to ensure the result is
10        signed and can be negated.
11    
12    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
13        CMake config files.
14    
15    3.  Revise the creation of config.h.generic so that all boolean macros are
16        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17        overriding via -D on the command line possible.
18    
19    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20        to unsigned int is reported to make a quite noticeable speed difference in
21        a specific Windows environment. Testing on Linux did also appear to show
22        some benefit (and it is clearly not harmful). Also fixed the definition of
23        Xop which should be unsigned.
24    
25    5.  Related to (4), changing the definition of the intermediate variable cc
26        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27        improvements.
28    
29    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
30        unnecessary spaces.
31    
32    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33        and later.
34    
35    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36    
37    
38    Version 8.32 30-November-2012
39    -----------------------------
40    
41    1.  Improved JIT compiler optimizations for first character search and single
42        character iterators.
43    
44    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
45        Patch by Daniel Richard G.
46    
47    3.  Single character iterator optimizations in the JIT compiler.
48    
49    4.  Improved JIT compiler optimizations for character ranges.
50    
51    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
52        Reported by Giuseppe D'Angelo.
53    
54    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
55        of a line, was being set incorrectly in cases where .* appeared inside
56        atomic brackets at the start of a pattern, or where there was a subsequent
57        *PRUNE or *SKIP.
58    
59    7.  Improved instruction cache flush for POWER/PowerPC.
60        Patch by Daniel Richard G.
61    
62    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
63        grep:
64    
65        (a) There is now no limit to the number of patterns to be matched.
66    
67        (b) An error is given if a pattern is too long.
68    
69        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
70            are now supported.
71    
72        (d) --exclude-from and --include-from (multiple use) have been added.
73    
74        (e) Exclusions and inclusions now apply to all files and directories, not
75            just to those obtained from scanning a directory recursively.
76    
77        (f) Multiple uses of -f and --file-list are now supported.
78    
79        (g) In a Windows environment, the default for -d has been changed from
80            "read" (the GNU grep default) to "skip", because otherwise the presence
81            of a directory in the file list provokes an error.
82    
83        (h) The documentation has been revised and clarified in places.
84    
85    9.  Improve the matching speed of capturing brackets.
86    
87    10. Changed the meaning of \X so that it now matches a Unicode extended
88        grapheme cluster.
89    
90    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
91        out POSIX threads when JIT support is configured.
92    
93    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
94    
95    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
96        structure could go wrong in environments where size_t is not the same size
97        as int.
98    
99    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
100        set.
101    
102    15. The EBCDIC support had decayed; later updates to the code had included
103        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
104        general tidy up of EBCDIC-related issues, and the documentation was also
105        not quite right. There is now a test that can be run on ASCII systems to
106        check some of the EBCDIC-related things (but is it not a full test).
107    
108    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
109        in a small tidy to the code.
110    
111    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
112    
113    18. If the --only-matching (-o) option in pcregrep is specified multiple
114        times, each one causes appropriate output. For example, -o1 -o2 outputs the
115        substrings matched by the 1st and 2nd capturing parentheses. A separating
116        string can be specified by --om-separator (default empty).
117    
118    19. Improving the first n character searches.
119    
120    20. Turn case lists for horizontal and vertical white space into macros so that
121        they are defined only once.
122    
123    21. This set of changes together give more compatible Unicode case-folding
124        behaviour for characters that have more than one other case when UCP
125        support is available.
126    
127        (a) The Unicode property table now has offsets into a new table of sets of
128            three or more characters that are case-equivalent. The MultiStage2.py
129            script that generates these tables (the pcre_ucd.c file) now scans
130            CaseFolding.txt instead of UnicodeData.txt for character case
131            information.
132    
133        (b) The code for adding characters or ranges of characters to a character
134            class has been abstracted into a generalized function that also handles
135            case-independence. In UTF-mode with UCP support, this uses the new data
136            to handle characters with more than one other case.
137    
138        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
139            whose other case is greater than 256 are now correctly matched
140            caselessly. Previously, the high codepoint matched the low one, but not
141            vice versa.
142    
143        (d) The processing of \h, \H, \v, and \ in character classes now makes use
144            of the new class addition function, using character lists defined as
145            macros alongside the case definitions of 20 above.
146    
147        (e) Caseless back references now work with characters that have more than
148            one other case.
149    
150        (f) General caseless matching of characters with more than one other case
151            is supported.
152    
153    22. Unicode character properties were updated from Unicode 6.2.0
154    
155    23. Improved CMake support under Windows. Patch by Daniel Richard G.
156    
157    24. Add support for 32-bit character strings, and UTF-32
158    
159    25. Major JIT compiler update (code refactoring and bugfixing).
160        Experimental Sparc 32 support is added.
161    
162    26. Applied a modified version of Daniel Richard G's patch to create
163        pcre.h.generic and config.h.generic by "make" instead of in the
164        PrepareRelease script.
165    
166    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
167        pcre_compile.c when checking for a zero character.
168    
169    28. Introducing a native interface for JIT. Through this interface, the compiled
170        machine code can be directly executed. The purpose of this interface is to
171        provide fast pattern matching, so several sanity checks are not performed.
172        However, feature tests are still performed. The new interface provides
173        1.4x speedup compared to the old one.
174    
175    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
176        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
177        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
178    
179    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
180        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
181        the "old" RFC 2279). Instead, it ended up passing a negative length to
182        pcre_exec().
183    
184    31. Add support for GCC's visibility feature to hide internal functions.
185    
186    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
187        "unknown -C option" after outputting 0 or 1.
188    
189    33. There is now support for generating a code coverage report for the test
190        suite in environments where gcc is the compiler and lcov is installed. This
191        is mainly for the benefit of the developers.
192    
193    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
194        unaddressable using valgrind annotations, allowing valgrind to detect
195        invalid memory accesses. This is mainly for the benefit of the developers.
196    
197    25. (*UTF) can now be used to start a pattern in any of the three libraries.
198    
199    26. Give configure error if --enable-cpp but no C++ compiler found.
200    
201    
202    Version 8.31 06-July-2012
203    -------------------------
204    
205    1.  Fixing a wrong JIT test case and some compiler warnings.
206    
207    2.  Removed a bashism from the RunTest script.
208    
209    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
210        to unsigned type, result still unsigned" that was given by an MS compiler
211        on encountering the code "-sizeof(xxx)".
212    
213    4.  Partial matching support is added to the JIT compiler.
214    
215    5.  Fixed several bugs concerned with partial matching of items that consist
216        of more than one character:
217    
218        (a) /^(..)\1/ did not partially match "aba" because checking references was
219            done on an "all or nothing" basis. This also applied to repeated
220            references.
221    
222        (b) \R did not give a hard partial match if \r was found at the end of the
223            subject.
224    
225        (c) \X did not give a hard partial match after matching one or more
226            characters at the end of the subject.
227    
228        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
229            a partial match for the string "\r".
230    
231        (e) When newline was set to CRLF, the metacharacter "." did not recognize
232            a partial match for a CR character at the end of the subject string.
233    
234    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
235        running pcretest, the text "(JIT)" added to the output whenever JIT is
236        actually used to run the match.
237    
238    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
239        or /S+[+] with a digit between 1 and 7.
240    
241    8.  OP_NOT now supports any UTF character not just single-byte ones.
242    
243    9.  (*MARK) control verb is now supported by the JIT compiler.
244    
245    10. The command "./RunTest list" lists the available tests without actually
246        running any of them. (Because I keep forgetting what they all are.)
247    
248    11. Add PCRE_INFO_MAXLOOKBEHIND.
249    
250    12. Applied a (slightly modified) user-supplied patch that improves performance
251        when the heap is used for recursion (compiled with --disable-stack-for-
252        recursion). Instead of malloc and free for each heap frame each time a
253        logical recursion happens, frames are retained on a chain and re-used where
254        possible. This sometimes gives as much as 30% improvement.
255    
256    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
257        call.
258    
259    14. As documented, (*COMMIT) is now confined to within a positive assertion.
260    
261    15. It is now possible to link pcretest with libedit as an alternative to
262        libreadline.
263    
264    16. (*COMMIT) control verb is now supported by the JIT compiler.
265    
266    17. The Unicode data tables have been updated to Unicode 6.1.0.
267    
268    18. Added --file-list option to pcregrep.
269    
270    19. Added binary file support to pcregrep, including the -a, --binary-files,
271        -I, and --text options.
272    
273    20. The madvise function is renamed for posix_madvise for QNX compatibility
274        reasons. Fixed by Giuseppe D'Angelo.
275    
276    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
277    
278    22. Changed the option for creating symbolic links for 16-bit man pages from
279        -s to -sf so that re-installing does not cause issues.
280    
281    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
282    
283    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
284        matches in certain environments (the workspace was not being correctly
285        retained). Also added to pcre_dfa_exec() a simple plausibility check on
286        some of the workspace data at the beginning of a restart.
287    
288    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
289        was not doing so when it should - probably a typo introduced by SVN 528
290        (change 8.10/14).
291    
292    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
293        \w+ when the character tables indicated that \x{c4} was a word character.
294        There were several related cases, all because the tests for doing a table
295        lookup were testing for characters less than 127 instead of 255.
296    
297    27. If a pattern contains capturing parentheses that are not used in a match,
298        their slots in the ovector are set to -1. For those that are higher than
299        any matched groups, this happens at the end of processing. In the case when
300        there were back references that the ovector was too small to contain
301        (causing temporary malloc'd memory to be used during matching), and the
302        highest capturing number was not used, memory off the end of the ovector
303        was incorrectly being set to -1. (It was using the size of the temporary
304        memory instead of the true size.)
305    
306    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
307        ovector size, it uses memory at the end of the block that it has got.
308    
309    29. Check for an overlong MARK name and give an error at compile time. The
310        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
311    
312    30. JIT compiler update.
313    
314    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
315        Rill for the patch.
316    
317    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
318    
319    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
320    
321    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
322        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
323        was enabled.
324    
325    35. Improve JIT code generation for greedy plus quantifier.
326    
327    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
328        1 to "aa" instead of to an empty string. The bug affected repeated groups
329        that could potentially match an empty string.
330    
331    37. Optimizing single character iterators in JIT.
332    
333    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
334        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
335        codepoints that are too big for the mode are faulted, and in a UTF mode,
336        disallowed codepoints are also faulted.
337    
338    39. If PCRE was compiled with UTF support, in three places in the DFA
339        matcher there was code that should only have been obeyed in UTF mode, but
340        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
341        processing when bytes with values greater than 127 were present. In 16-bit
342        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
343        both cases the values are those that cannot be the first data item in a UTF
344        character. The three items that might have provoked this were recursions,
345        possessively repeated groups, and atomic groups.
346    
347    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
348        and pcregrep, because some OS require shared objects to be explicitly
349        passed to ld, causing the link step to fail if they are not.
350    
351    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
352        mode, patterns that started with \h* or \R* might be incorrectly matched.
353    
354    
355    Version 8.30 04-February-2012
356    -----------------------------
357    
358    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
359        name is defined in ctype.h.
360    
361    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
362        only in quite long subpatterns.
363    
364    3.  Removed the function pcre_info(), which has been obsolete and deprecated
365        since it was replaced by pcre_fullinfo() in February 2000.
366    
367    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
368        match a (*MARK), and the match failed at the start of the subject, a
369        reference to memory before the start of the subject could occur. This bug
370        was introduced by fix 17 of release 8.21.
371    
372    5.  A reference to an unset group with zero minimum repetition was giving
373        totally wrong answers (in non-JavaScript-compatibility mode). For example,
374        /(another)?(\1?)test/ matched against "hello world test". This bug was
375        introduced in release 8.13.
376    
377    6.  Add support for 16-bit character strings (a large amount of work involving
378        many changes and refactorings).
379    
380    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
381        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
382        from a file.
383    
384    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
385        rounding is not applied in this particular case).
386    
387    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
388        if they appear, or are escaped, in patterns.
389    
390    10. Get rid of a number of -Wunused-but-set-variable warnings.
391    
392    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
393        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
394        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
395        also returns the mark "x". This bug applied to capturing parentheses,
396        non-capturing parentheses, and atomic parentheses. It also applied to some
397        assertions.
398    
399    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
400        information out of configure.ac instead of relying on pcre.h.generic, which
401        is not stored in the repository.
402    
403    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
404        -lreadline.
405    
406    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
407    
408    15. Applied Graycode's patch to put the top-level frame on the stack rather
409        than the heap when not using the stack for recursion. This gives a
410        performance improvement in many cases when recursion is not deep.
411    
412    16. Experimental code added to "pcretest -C" to output the stack frame size.
413    
414    
415    Version 8.21 12-Dec-2011
416    ------------------------
417    
418    1.  Updating the JIT compiler.
419    
420    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
421        are added as well.
422    
423    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
424        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
425        calling _pcre_jit_exec. Some extra comments are added.
426    
427    4.  (*MARK) settings inside atomic groups that do not contain any capturing
428        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
429        was introduced by change 18 for 8.20.
430    
431    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
432        ECMA-262 standard.
433    
434    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
435        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
436        This bug was probably introduced by change 9 of 8.13.
437    
438    7.  While fixing 6 above, I noticed that a number of other items were being
439        incorrectly rejected as "not fixed length". This arose partly because newer
440        opcodes had not been added to the fixed-length checking code. I have (a)
441        corrected the bug and added tests for these items, and (b) arranged for an
442        error to occur if an unknown opcode is encountered while checking for fixed
443        length instead of just assuming "not fixed length". The items that were
444        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
445        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
446        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
447    
448    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
449        being incorrectly compiled and would have given unpredicatble results.
450    
451    9.  A possessively repeated subpattern with minimum repeat count greater than
452        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
453        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
454        the first (A) could occur when it should not.
455    
456    10. Add a cast and remove a redundant test from the code.
457    
458    11. JIT should use pcre_malloc/pcre_free for allocation.
459    
460    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
461        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
462        is anything other than /usr, -L is still shown).
463    
464    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
465    
466    14. Perl does not support \N without a following name in a [] class; PCRE now
467        also gives an error.
468    
469    15. If a forward reference was repeated with an upper limit of around 2000,
470        it caused the error "internal error: overran compiling workspace". The
471        maximum number of forward references (including repeats) was limited by the
472        internal workspace, and dependent on the LINK_SIZE. The code has been
473        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
474        the default depends on LINK_SIZE. There is a new upper limit (for safety)
475        of around 200,000 forward references. While doing this, I also speeded up
476        the filling in of repeated forward references.
477    
478    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
479        incorrectly expecting the subject to contain another "a" after the start.
480    
481    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
482        in the match, the SKIP should be ignored. This was not happening; instead
483        the SKIP was being treated as NOMATCH. For patterns such as
484        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
485        tested.
486    
487    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
488        now much more compatible with Perl, in particular in cases where the result
489        is a non-match for a non-anchored pattern. For example, if
490        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
491        "m", where previously it did not return a name. A side effect of this
492        change is that for partial matches, the last encountered mark name is
493        returned, as for non matches. A number of tests that were previously not
494        Perl-compatible have been moved into the Perl-compatible test files. The
495        refactoring has had the pleasing side effect of removing one argument from
496        the match() function, thus reducing its stack requirements.
497    
498    19. If the /S+ option was used in pcretest to study a pattern using JIT,
499        subsequent uses of /S (without +) incorrectly behaved like /S+.
500    
501    21. Retrieve executable code size support for the JIT compiler and fixing
502        some warnings.
503    
504    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
505        not work when the shorter character appeared right at the end of the
506        subject string.
507    
508    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
509        systems.
510    
511    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
512        output it when the /M option is used in pcretest.
513    
514    25. The CheckMan script was not being included in the distribution. Also, added
515        an explicit "perl" to run Perl scripts from the PrepareRelease script
516        because this is reportedly needed in Windows.
517    
518    26. If study data was being save in a file and studying had not found a set of
519        "starts with" bytes for the pattern, the data written to the file (though
520        never used) was taken from uninitialized memory and so caused valgrind to
521        complain.
522    
523    27. Updated RunTest.bat as provided by Sheri Pierce.
524    
525    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
526    
527    29. Computation of memory usage for the table of capturing group names was
528        giving an unnecessarily large value.
529    
530    
531    Version 8.20 21-Oct-2011
532    ------------------------
533    
534    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
535        a POSIX class. After further experiments with Perl, which convinced me that
536        Perl has bugs and confusions, a closing square bracket is no longer allowed
537        in a POSIX name. This bug also affected patterns with classes that started
538        with full stops.
539    
540    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
541        captured substring, but while checking the failing first alternative,
542        substring 1 is temporarily captured. If the output vector supplied to
543        pcre_exec() was not big enough for this capture, the yield of the function
544        was still zero ("insufficient space for captured substrings"). This cannot
545        be totally fixed without adding another stack variable, which seems a lot
546        of expense for a edge case. However, I have improved the situation in cases
547        such as /(a)(b)x|abc/ matched against "abc", where the return code
548        indicates that fewer than the maximum number of slots in the ovector have
549        been set.
550    
551    3.  Related to (2) above: when there are more back references in a pattern than
552        slots in the output vector, pcre_exec() uses temporary memory during
553        matching, and copies in the captures as far as possible afterwards. It was
554        using the entire output vector, but this conflicts with the specification
555        that only 2/3 is used for passing back captured substrings. Now it uses
556        only the first 2/3, for compatibility. This is, of course, another edge
557        case.
558    
559    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
560        main code base, and can be used by building with --enable-jit. When this is
561        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
562        runtime --no-jit option is given.
563    
564    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
565        ovector, the return from the function was zero, implying that there were
566        other matches that did not fit. The correct "exactly full" value is now
567        returned.
568    
569    6.  If a subpattern that was called recursively or as a subroutine contained
570        (*PRUNE) or any other control that caused it to give a non-standard return,
571        invalid errors such as "Error -26 (nested recursion at the same subject
572        position)" or even infinite loops could occur.
573    
574    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
575        computing the minimum length on reaching *ACCEPT, and so ended up with the
576        wrong value of 1 rather than 0. Further investigation indicates that
577        computing a minimum subject length in the presence of *ACCEPT is difficult
578        (think back references, subroutine calls), and so I have changed the code
579        so that no minimum is registered for a pattern that contains *ACCEPT.
580    
581    8.  If (*THEN) was present in the first (true) branch of a conditional group,
582        it was not handled as intended. [But see 16 below.]
583    
584    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
585        Sheri Pierce.
586    
587    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
588        the first byte in a match must be "a".
589    
590    11. Change 17 for 8.13 increased the recursion depth for patterns like
591        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
592        pattern contains any instances of (*THEN). If it does not, the old
593        optimizations are restored. It would be nice to do this on a per-group
594        basis, but at the moment that is not feasible.
595    
596    12. In some environments, the output of pcretest -C is CRLF terminated. This
597        broke RunTest's code that checks for the link size. A single white space
598        character after the value is now allowed for.
599    
600    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
601        For "fr", it uses the Windows-specific input and output files.
602    
603    14. If (*THEN) appeared in a group that was called recursively or as a
604        subroutine, it did not work as intended. [But see next item.]
605    
606    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
607        pattern fragments (but not containing any | characters). If A and B are
608        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
609        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
610        D. In other words, Perl considers parentheses that do not contain any |
611        characters to be part of a surrounding alternative, whereas PCRE was
612        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
613        differently. PCRE now behaves in the same way as Perl, except in the case
614        of subroutine/recursion calls such as (?1) which have in any case always
615        been different (but PCRE had them first :-).
616    
617    16. Related to 15 above: Perl does not treat the | in a conditional group as
618        creating alternatives. Such a group is treated in the same way as an
619        ordinary group without any | characters when processing (*THEN). PCRE has
620        been changed to match Perl's behaviour.
621    
622    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
623        RunGrepTest script failed.
624    
625    18. Change 22 for version 13 caused atomic groups to use more stack. This is
626        inevitable for groups that contain captures, but it can lead to a lot of
627        stack use in large patterns. The old behaviour has been restored for atomic
628        groups that do not contain any capturing parentheses.
629    
630    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
631        suppress the check for a minimum subject length at run time. (If it was
632        given to pcre_exec() or pcre_dfa_exec() it did work.)
633    
634    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
635        fail to work when decoding hex characters in data strings in EBCDIC
636        environments.
637    
638    21. It appears that in at least one Mac OS environment, the isxdigit() function
639        is implemented as a macro that evaluates to its argument more than once,
640        contravening the C 90 Standard (I haven't checked a later standard). There
641        was an instance in pcretest which caused it to go wrong when processing
642        \x{...} escapes in subject strings. The has been rewritten to avoid using
643        things like p++ in the argument of isxdigit().
644    
645    
646    Version 8.13 16-Aug-2011
647    ------------------------
648    
649    1.  The Unicode data tables have been updated to Unicode 6.0.0.
650    
651    2.  Two minor typos in pcre_internal.h have been fixed.
652    
653    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
654        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
655        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
656    
657    4.  There were a number of related bugs in the code for matching backrefences
658        caselessly in UTF-8 mode when codes for the characters concerned were
659        different numbers of bytes. For example, U+023A and U+2C65 are an upper
660        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
661        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
662        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
663        2-byte code at the end of the subject (it thought there wasn't enough data
664        left).
665    
666    5.  Comprehensive information about what went wrong is now returned by
667        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
668        as the output vector has at least 2 elements. The offset of the start of
669        the failing character and a reason code are placed in the vector.
670    
671    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
672        now returned is for the first byte of the failing character, instead of the
673        last byte inspected. This is an incompatible change, but I hope it is small
674        enough not to be a problem. It makes the returned offset consistent with
675        pcre_exec() and pcre_dfa_exec().
676    
677    7.  pcretest now gives a text phrase as well as the error number when
678        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
679        failure, the offset and reason code are output.
680    
681    8.  When \R was used with a maximizing quantifier it failed to skip backwards
682        over a \r\n pair if the subsequent match failed. Instead, it just skipped
683        back over a single character (\n). This seems wrong (because it treated the
684        two characters as a single entity when going forwards), conflicts with the
685        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
686        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
687        has been changed.
688    
689    9.  Some internal refactoring has changed the processing so that the handling
690        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
691        time (the PCRE_DOTALL option was changed this way some time ago: version
692        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
693        which was always a bit of a fudge. It also means that there is one less
694        argument for the match() function, which reduces its stack requirements
695        slightly. This change also fixes an incompatibility with Perl: the pattern
696        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
697    
698    10. More internal refactoring has drastically reduced the number of recursive
699        calls to match() for possessively repeated groups such as (abc)++ when
700        using pcre_exec().
701    
702    11. While implementing 10, a number of bugs in the handling of groups were
703        discovered and fixed:
704    
705        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
706        (a|)*(?1) gave a compile-time internal error.
707        ((a|)+)+  did not notice that the outer group could match an empty string.
708        (^a|^)+   was not marked as anchored.
709        (.*a|.*)+ was not marked as matching at start or after a newline.
710    
711    12. Yet more internal refactoring has removed another argument from the match()
712        function. Special calls to this function are now indicated by setting a
713        value in a variable in the "match data" data block.
714    
715    13. Be more explicit in pcre_study() instead of relying on "default" for
716        opcodes that mean there is no starting character; this means that when new
717        ones are added and accidentally left out of pcre_study(), testing should
718        pick them up.
719    
720    14. The -s option of pcretest has been documented for ages as being an old
721        synonym of -m (show memory usage). I have changed it to mean "force study
722        for every regex", that is, assume /S for every regex. This is similar to -i
723        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
724        using it. It makes it easier to run collections of tests with and without
725        study enabled, and thereby test pcre_study() more easily. All the standard
726        tests are now run with and without -s (but some patterns can be marked as
727        "never study" - see 20 below).
728    
729    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
730        restoration of the capturing data to the outer values was not happening
731        correctly.
732    
733    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
734        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
735        pattern had matched an empty string, and so incorrectly returned a no
736        match.
737    
738    17. There was optimizing code for the last branch of non-capturing parentheses,
739        and also for the obeyed branch of a conditional subexpression, which used
740        tail recursion to cut down on stack usage. Unfortunately, now that there is
741        the possibility of (*THEN) occurring in these branches, tail recursion is
742        no longer possible because the return has to be checked for (*THEN). These
743        two optimizations have therefore been removed. [But see 8.20/11 above.]
744    
745    18. If a pattern containing \R was studied, it was assumed that \R always
746        matched two bytes, thus causing the minimum subject length to be
747        incorrectly computed because \R can also match just one byte.
748    
749    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
750        was incorrectly computed.
751    
752    20. If /S is present twice on a test pattern in pcretest input, it now
753        *disables* studying, thereby overriding the use of -s on the command line
754        (see 14 above). This is necessary for one or two tests to keep the output
755        identical in both cases.
756    
757    21. When (*ACCEPT) was used in an assertion that matched an empty string and
758        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
759    
760    22. When an atomic group that contained a capturing parenthesis was
761        successfully matched, but the branch in which it appeared failed, the
762        capturing was not being forgotten if a higher numbered group was later
763        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
764        group 1 to "a", when in fact it should be unset. This applied to multi-
765        branched capturing and non-capturing groups, repeated or not, and also to
766        positive assertions (capturing in negative assertions does not happen
767        in PCRE) and also to nested atomic groups.
768    
769    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
770        subject after a captured substring, to make it easier to tell which of a
771        number of identical substrings has been captured.
772    
773    24. The way atomic groups are processed by pcre_exec() has been changed so that
774        if they are repeated, backtracking one repetition now resets captured
775        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
776        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
777        "aaa". Previously, it would have been "a". As part of this code
778        refactoring, the way recursive calls are handled has also been changed.
779    
780    25. If an assertion condition captured any substrings, they were not passed
781        back unless some other capturing happened later. For example, if
782        (?(?=(a))a) was matched against "a", no capturing was returned.
783    
784    26. When studying a pattern that contained subroutine calls or assertions,
785        the code for finding the minimum length of a possible match was handling
786        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
787        group 1 called group 2 while simultaneously a separate group 2 called group
788        1). A stack overflow occurred in this case. I have fixed this by limiting
789        the recursion depth to 10.
790    
791    27. Updated RunTest.bat in the distribution to the version supplied by Tom
792        Fortmann. This supports explicit test numbers on the command line, and has
793        argument validation and error reporting.
794    
795    28. An instance of \X with an unlimited repeat could fail if at any point the
796        first character it looked at was a mark character.
797    
798    29. Some minor code refactoring concerning Unicode properties and scripts
799        should reduce the stack requirement of match() slightly.
800    
801    30. Added the '=' option to pcretest to check the setting of unused capturing
802        slots at the end of the pattern, which are documented as being -1, but are
803        not included in the return count.
804    
805    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
806        compiled something random. Now it gives a compile-time error (as does
807        Perl).
808    
809    32. A *MARK encountered during the processing of a positive assertion is now
810        recorded and passed back (compatible with Perl).
811    
812    33. If --only-matching or --colour was set on a pcregrep call whose pattern
813        had alternative anchored branches, the search for a second match in a line
814        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
815        matched the line "0102" twice. The same bug affected patterns that started
816        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
817        twice.
818    
819    34. Previously, PCRE did not allow quantification of assertions. However, Perl
820        does, and because of capturing effects, quantifying parenthesized
821        assertions may at times be useful. Quantifiers are now allowed for
822        parenthesized assertions.
823    
824    35. A minor code tidy in pcre_compile() when checking options for \R usage.
825    
826    36. \g was being checked for fancy things in a character class, when it should
827        just be a literal "g".
828    
829    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
830        appearance of a nested POSIX class supersedes an apparent external class.
831        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
832        unescaped square brackets may also appear as part of class names. For
833        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
834        more like Perl. (But see 8.20/1 above.)
835    
836    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
837        was because it thought it was \N{name}, which is not supported).
838    
839    39. Add minix to OS list not supporting the -S option in pcretest.
840    
841    40. PCRE tries to detect cases of infinite recursion at compile time, but it
842        cannot analyze patterns in sufficient detail to catch mutual recursions
843        such as ((?1))((?2)). There is now a runtime test that gives an error if a
844        subgroup is called recursively as a subpattern for a second time at the
845        same position in the subject string. In previous releases this might have
846        been caught by the recursion limit, or it might have run out of stack.
847    
848    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
849        happen only once. PCRE was, however incorrectly giving a compile time error
850        "recursive call could loop indefinitely" because it cannot analyze the
851        pattern in sufficient detail. The compile time test no longer happens when
852        PCRE is compiling a conditional subpattern, but actual runaway loops are
853        now caught at runtime (see 40 above).
854    
855    42. It seems that Perl allows any characters other than a closing parenthesis
856        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
857        has been changed to be the same.
858    
859    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
860        as not to get warnings when autogen.sh is called. Also changed
861        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
862    
863    44. To help people who use pcregrep to scan files containing exceedingly long
864        lines, the following changes have been made:
865    
866        (a) The default value of the buffer size parameter has been increased from
867            8K to 20K. (The actual buffer used is three times this size.)
868    
869        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
870            PCRE is built.
871    
872        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
873            to be set at run time.
874    
875        (d) Numerical values in pcregrep options can be followed by K or M, for
876            example --buffer-size=50K.
877    
878        (e) If a line being scanned overflows pcregrep's buffer, an error is now
879            given and the return code is set to 2.
880    
881    45. Add a pointer to the latest mark to the callout data block.
882    
883    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
884        partial match of an empty string instead of no match. This was specific to
885        the use of ".".
886    
887    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
888        complete match instead of a partial match. This bug was dependent on both
889        the PCRE_UTF8 and PCRE_DOTALL options being set.
890    
891    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
892        starting byte set, because \b was not being ignored.
893    
894    
895    Version 8.12 15-Jan-2011
896    ------------------------
897    
898    1.  Fixed some typos in the markup of the man pages, and wrote a script that
899        checks for such things as part of the documentation building process.
900    
901    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
902        --match-limit and --recursion-limit options (added for 8.11). In
903        particular, this made one of the standard tests fail. (The integer value
904        went into the wrong half of a long int.)
905    
906    3.  If the --colour option was given to pcregrep with -v (invert match), it
907        did strange things, either producing crazy output, or crashing. It should,
908        of course, ignore a request for colour when reporting lines that do not
909        match.
910    
911    4.  Another pcregrep bug caused similar problems if --colour was specified with
912        -M (multiline) and the pattern match finished with a line ending.
913    
914    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
915        matched in multiline mode, the following line was shown as part of the
916        match. This seems wrong, so I have changed it.
917    
918    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
919        the check for further matches in the same line (so they could be coloured)
920        to overrun the end of the current line. If another match was found, it was
921        incorrectly shown (and then shown again when found in the next line).
922    
923    7.  If pcregrep was compiled under Windows, there was a reference to the
924        function pcregrep_exit() before it was defined. I am assuming this was
925        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
926        reported by a user. I've moved the definition above the reference.
927    
928    
929    Version 8.11 10-Dec-2010
930    ------------------------
931    
932    1.  (*THEN) was not working properly if there were untried alternatives prior
933        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
934        backtracked to try for "b" instead of moving to the next alternative branch
935        at the same level (in this case, to look for "c"). The Perl documentation
936        is clear that when (*THEN) is backtracked onto, it goes to the "next
937        alternative in the innermost enclosing group".
938    
939    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
940        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
941        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
942        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
943        (*THEN).
944    
945    3.  If \s appeared in a character class, it removed the VT character from
946        the class, even if it had been included by some previous item, for example
947        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
948        of \s, but is part of the POSIX "space" class.)
949    
950    4.  A partial match never returns an empty string (because you can always
951        match an empty string at the end of the subject); however the checking for
952        an empty string was starting at the "start of match" point. This has been
953        changed to the "earliest inspected character" point, because the returned
954        data for a partial match starts at this character. This means that, for
955        example, /(?<=abc)def/ gives a partial match for the subject "abc"
956        (previously it gave "no match").
957    
958    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
959        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
960        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
961        has an implication that the given string is incomplete (because a partial
962        match is preferred over a full match). For this reason, these items now
963        give a partial match in this situation. [Aside: previously, the one case
964        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
965        match rather than a full match, which was wrong by the old rules, but is
966        now correct.]
967    
968    6.  There was a bug in the handling of #-introduced comments, recognized when
969        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
970        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
971        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
972        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
973        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
974        places in pcre_compile().
975    
976    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
977        comments when looking ahead for named forward references to subpatterns,
978        the only newline sequence it recognized was NL. It now handles newlines
979        according to the set newline convention.
980    
981    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
982        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
983        cater for a lack of strerror(). These oversights have been fixed.
984    
985    9.  Added --match-limit and --recursion-limit to pcregrep.
986    
987    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
988    
989    11. When the -o option was used, pcregrep was setting a return code of 1, even
990        when matches were found, and --line-buffered was not being honoured.
991    
992    12. Added an optional parentheses number to the -o and --only-matching options
993        of pcregrep.
994    
995    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
996        can match an empty string. The code to do it in pcretest and pcredemo
997        needed fixing:
998    
999        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1000            only one byte after an empty string match just before CRLF (this case
1001            just got forgotten; "any" and "anycrlf" were OK).
1002    
1003        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1004            mode when an empty string match preceded an ASCII character followed by
1005            a non-ASCII character. (The code for advancing by one character rather
1006            than one byte was nonsense.)
1007    
1008        (c) The pcredemo.c sample program did not have any code at all to handle
1009            the cases when CRLF is a valid newline sequence.
1010    
1011    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1012        as a starting offset was within the subject string. There is now a new
1013        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1014        negative or greater than the length of the string. In order to test this,
1015        pcretest is extended to allow the setting of negative starting offsets.
1016    
1017    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1018        starting offset points to the beginning of a UTF-8 character was
1019        unnecessarily clumsy. I tidied it up.
1020    
1021    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1022        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1023    
1024    17. Nobody had reported that the --include_dir option, which was added in
1025        release 7.7 should have been called --include-dir (hyphen, not underscore)
1026        for compatibility with GNU grep. I have changed it to --include-dir, but
1027        left --include_dir as an undocumented synonym, and the same for
1028        --exclude-dir, though that is not available in GNU grep, at least as of
1029        release 2.5.4.
1030    
1031    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1032        characters from a string of bytes) have been redefined so as not to use
1033        loops, in order to improve performance in some environments. At the same
1034        time, I abstracted some of the common code into auxiliary macros to save
1035        repetition (this should not affect the compiled code).
1036    
1037    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1038        compile-time error is now given if \c is not followed by an ASCII
1039        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1040        different, and any byte value is allowed.)
1041    
1042    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1043        START_OPTIMIZE option, which is now allowed at compile time - but just
1044        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1045        to pcregrep and other applications that have no direct access to PCRE
1046        options. The new /Y option in pcretest sets this option when calling
1047        pcre_compile().
1048    
1049    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1050        back references. Groups containing recursive back references were forced to
1051        be atomic by that change, but in the case of named groups, the amount of
1052        memory required was incorrectly computed, leading to "Failed: internal
1053        error: code overflow". This has been fixed.
1054    
1055    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1056        pcretest.c, to avoid build problems in some Borland environments.
1057    
1058    
1059    Version 8.10 25-Jun-2010
1060    ------------------------
1061    
1062    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1063        THEN.
1064    
1065    2.  (*ACCEPT) was not working when inside an atomic group.
1066    
1067    3.  Inside a character class, \B is treated as a literal by default, but
1068        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1069        causes the error). The code is unchanged, but I tidied the documentation.
1070    
1071    4.  Inside a character class, PCRE always treated \R and \X as literals,
1072        whereas Perl faults them if its -w option is set. I have changed PCRE so
1073        that it faults them when PCRE_EXTRA is set.
1074    
1075    5.  Added support for \N, which always matches any character other than
1076        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1077    
1078    6.  When compiling pcregrep with newer versions of gcc which may have
1079        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1080        declared with attribute warn_unused_result" were given. Just casting the
1081        result to (void) does not stop the warnings; a more elaborate fudge is
1082        needed. I've used a macro to implement this.
1083    
1084    7.  Minor change to pcretest.c to avoid a compiler warning.
1085    
1086    8.  Added four artifical Unicode properties to help with an option to make
1087        \s etc use properties (see next item). The new properties are: Xan
1088        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1089    
1090    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1091        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1092        this option. Modified pcretest to add /W to test this facility. Added
1093        REG_UCP to make it available via the POSIX interface.
1094    
1095    10. Added --line-buffered to pcregrep.
1096    
1097    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1098        studied, and the match started with a letter with a code point greater than
1099        127 whose first byte was different to the first byte of the other case of
1100        the letter, the other case of this starting letter was not recognized
1101        (#976).
1102    
1103    12. If a pattern that was studied started with a repeated Unicode property
1104        test, for example, \p{Nd}+, there was the theoretical possibility of
1105        setting up an incorrect bitmap of starting bytes, but fortunately it could
1106        not have actually happened in practice until change 8 above was made (it
1107        added property types that matched character-matching opcodes).
1108    
1109    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1110        possible starting bytes for non-anchored patterns.
1111    
1112    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1113        \R, and also a number of cases that involve Unicode properties, both
1114        explicit and implicit when PCRE_UCP is set.
1115    
1116    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1117        input, it could crash or give wrong results if characters with values
1118        greater than 0xc0 were present in the subject string. (Detail: it assumed
1119        UTF-8 input when processing these items.)
1120    
1121    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1122        size_t is 64-bit (#991).
1123    
1124    17. Added a check for running out of memory when PCRE is compiled with
1125        --disable-stack-for-recursion (#990).
1126    
1127    18. If the last data line in a file for pcretest does not have a newline on
1128        the end, a newline was missing in the output.
1129    
1130    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1131        less than 128) in its various bitmaps. However, there is a facility for
1132        generating tables according to the current locale when PCRE is compiled. It
1133        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1134        characters, are recognized by isspace() and therefore were getting set in
1135        these tables, and indeed these tables seem to approximate to ISO 8859. This
1136        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1137        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1138        which of course cannot start UTF-8 characters. I have changed the code so
1139        that only real ASCII characters (less than 128) and the correct starting
1140        bytes for UTF-8 encodings are set for characters greater than 127 when in
1141        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1142        altogether.)
1143    
1144    20. Added the /T option to pcretest so as to be able to run tests with non-
1145        standard character tables, thus making it possible to include the tests
1146        used for 19 above in the standard set of tests.
1147    
1148    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1149        reference to a subpattern the other side of a comment that contains an
1150        opening parenthesis caused either an internal compiling error, or a
1151        reference to the wrong subpattern.
1152    
1153    
1154    Version 8.02 19-Mar-2010
1155    ------------------------
1156    
1157    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1158    
1159    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1160        configured.
1161    
1162    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1163        original author of that file, following a query about its status.
1164    
1165    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1166        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1167    
1168    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1169        quantifier applied to a forward-referencing subroutine call, could compile
1170        incorrect code or give the error "internal error: previously-checked
1171        referenced subpattern not found".
1172    
1173    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1174        variables to point to external functions. For these systems, therefore,
1175        pcre_malloc etc. are now initialized to local functions that call the
1176        relevant global functions.
1177    
1178    7.  There were two entries missing in the vectors called coptable and poptable
1179        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1180        I've fixed the data, and added a kludgy way of testing at compile time that
1181        the lengths are correct (equal to the number of opcodes).
1182    
1183    8.  Following on from 7, I added a similar kludge to check the length of the
1184        eint vector in pcreposix.c.
1185    
1186    9.  Error texts for pcre_compile() are held as one long string to avoid too
1187        much relocation at load time. To find a text, the string is searched,
1188        counting zeros. There was no check for running off the end of the string,
1189        which could happen if a new error number was added without updating the
1190        string.
1191    
1192    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1193    
1194    11. \K was not working if it appeared in an atomic group or in a group that
1195        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1196        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1197        the assertion is positive, but not if it is negative.
1198    
1199    12. Change 11 fortuitously reduced the size of the stack frame used in the
1200        "match()" function of pcre_exec.c by one pointer. Forthcoming
1201        implementation of support for (*MARK) will need an extra pointer on the
1202        stack; I have reserved it now, so that the stack frame size does not
1203        decrease.
1204    
1205    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1206        item in branch that calls a recursion is a subroutine call - as in the
1207        second branch in the above example - was incorrectly given the compile-
1208        time error "recursive call could loop indefinitely" because pcre_compile()
1209        was not correctly checking the subroutine for matching a non-empty string.
1210    
1211    14. The checks for overrunning compiling workspace could trigger after an
1212        overrun had occurred. This is a "should never occur" error, but it can be
1213        triggered by pathological patterns such as hundreds of nested parentheses.
1214        The checks now trigger 100 bytes before the end of the workspace.
1215    
1216    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1217    
1218    
1219    Version 8.01 19-Jan-2010
1220    ------------------------
1221    
1222    1.  If a pattern contained a conditional subpattern with only one branch (in
1223        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1224        computed the wrong minimum data length (which is of course zero for such
1225        subpatterns). This could cause incorrect "no match" results.
1226    
1227    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1228        the pattern is reset in the first branch, pcre_compile() failed with
1229        "internal error: code overflow at offset...". This happened only when
1230        the reset was to the original external option setting. (An optimization
1231        abstracts leading options settings into an external setting, which was the
1232        cause of this.)
1233    
1234    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1235        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1236        assertion pattern did not match (meaning that the assertion was true), it
1237        was incorrectly treated as false if the SKIP had been reached during the
1238        matching. This also applied to assertions used as conditions.
1239    
1240    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1241        assertion subpattern, including such a pattern used as a condition,
1242        unpredictable results occurred, instead of the error return
1243        PCRE_ERROR_DFA_UITEM.
1244    
1245    5.  The C++ GlobalReplace function was not working like Perl for the special
1246        situation when an empty string is matched. It now does the fancy magic
1247        stuff that is necessary.
1248    
1249    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1250        removed. (These were left over from very, very early versions of PCRE.)
1251    
1252    7.  Some cosmetic changes to the code to make life easier when compiling it
1253        as part of something else:
1254    
1255        (a) Change DEBUG to PCRE_DEBUG.
1256    
1257        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1258            called "current" as "current_branch", to prevent a collision with the
1259            Linux macro when compiled as a kernel module.
1260    
1261        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1262            prevent a collision with the Linux macro when compiled as a kernel
1263            module.
1264    
1265    8.  In pcre_compile() there are some checks for integer overflows that used to
1266        cast potentially large values to (double). This has been changed to that
1267        when building, a check for int64_t is made, and if it is found, it is used
1268        instead, thus avoiding the use of floating point arithmetic. (There is no
1269        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1270        double.
1271    
1272    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1273        2005 (difference between two addresses compared to an unsigned value).
1274    
1275    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1276        custom one, because of the following reported problem in Windows:
1277    
1278          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1279              under Win32.
1280          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1281              therefore missing the function definition.
1282          - The compiler thus generates a "C" signature for the test function.
1283          - The linker fails to find the "C" function.
1284          - PCRE fails to configure if asked to do so against libbz2.
1285    
1286    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1287        messages were output:
1288    
1289          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1290          rerunning libtoolize, to keep the correct libtool macros in-tree.
1291          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1292    
1293        I have done both of these things.
1294    
1295    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1296        most of the time, it *can* run out if it is given a pattern that contains a
1297        runaway infinite recursion. I updated the discussion in the pcrestack man
1298        page.
1299    
1300    13. Now that we have gone to the x.xx style of version numbers, the minor
1301        version may start with zero. Using 08 or 09 is a bad idea because users
1302        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1303        interpreted as invalid octal numbers. I've updated the previous comment in
1304        configure.ac, and also added a check that gives an error if 08 or 09 are
1305        used.
1306    
1307    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1308        causing partial matching to fail when the end of the subject matched \W
1309        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1310    
1311    15. There were some discrepancies between the declarations in pcre_internal.h
1312        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1313        their definitions. The declarations used "const uschar *" and the
1314        definitions used USPTR. Even though USPTR is normally defined as "const
1315        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1316        reported that: "This difference in casting confuses some C++ compilers, for
1317        example, SunCC recognizes above declarations as different functions and
1318        generates broken code for hbpcre." I have changed the declarations to use
1319        USPTR.
1320    
1321    16. GNU libtool is named differently on some systems. The autogen.sh script now
1322        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1323        (FreeBSD).
1324    
1325    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1326        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1327        comment: "Figure out how to create a longlong from a string: strtoll and
1328        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1329        instance, but it only takes 2 args instead of 3!"
1330    
1331    18. A subtle bug concerned with back references has been fixed by a change of
1332        specification, with a corresponding code fix. A pattern such as
1333        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1334        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1335        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1336        same bug. Such groups have to be quantified to be useful, or contained
1337        inside another quantified group. (If there's no repetition, the reference
1338        can never match.) The problem arises because, having left the group and
1339        moved on to the rest of the pattern, a later failure that backtracks into
1340        the group uses the captured value from the final iteration of the group
1341        rather than the correct earlier one. I have fixed this in PCRE by forcing
1342        any group that contains a reference to itself to be an atomic group; that
1343        is, there cannot be any backtracking into it once it has completed. This is
1344        similar to recursive and subroutine calls.
1345    
1346    
1347    Version 8.00 19-Oct-09
1348    ----------------------
1349    
1350    1.  The table for translating pcre_compile() error codes into POSIX error codes
1351        was out-of-date, and there was no check on the pcre_compile() error code
1352        being within the table. This could lead to an OK return being given in
1353        error.
1354    
1355    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1356        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1357        in a Windows environment.
1358    
1359    3.  The pcregrep --count option prints the count for each file even when it is
1360        zero, as does GNU grep. However, pcregrep was also printing all files when
1361        --files-with-matches was added. Now, when both options are given, it prints
1362        counts only for those files that have at least one match. (GNU grep just
1363        prints the file name in this circumstance, but including the count seems
1364        more useful - otherwise, why use --count?) Also ensured that the
1365        combination -clh just lists non-zero counts, with no names.
1366    
1367    4.  The long form of the pcregrep -F option was incorrectly implemented as
1368        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1369        but it seems right to fix it, and I didn't think it was worth preserving
1370        the old behaviour.
1371    
1372    5.  The command line items --regex=pattern and --regexp=pattern were not
1373        recognized by pcregrep, which required --regex pattern or --regexp pattern
1374        (with a space rather than an '='). The man page documented the '=' forms,
1375        which are compatible with GNU grep; these now work.
1376    
1377    6.  No libpcreposix.pc file was created for pkg-config; there was just
1378        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1379    
1380    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1381        when UCP support is not needed, by modifying the Python script that
1382        generates it from Unicode data files. This should not matter if the module
1383        is correctly used as a library, but I received one complaint about 50K of
1384        unwanted data. My guess is that the person linked everything into his
1385        program rather than using a library. Anyway, it does no harm.
1386    
1387    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1388        was a minimum greater than 1 for a wide character in a possessive
1389        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1390        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1391        character. Chaos in the form of incorrect output or a compiling loop could
1392        result.
1393    
1394    9.  The restrictions on what a pattern can contain when partial matching is
1395        requested for pcre_exec() have been removed. All patterns can now be
1396        partially matched by this function. In addition, if there are at least two
1397        slots in the offset vector, the offset of the earliest inspected character
1398        for the match and the offset of the end of the subject are set in them when
1399        PCRE_ERROR_PARTIAL is returned.
1400    
1401    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1402        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1403        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1404        and may be more useful for multi-segment matching.
1405    
1406    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1407        used to be given if ever the end of the subject was reached; now it is
1408        given only if matching could not proceed because another character was
1409        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1410        string "Z", which now yields "no match" instead of "partial match". In the
1411        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1412        final character ended with (*FAIL).
1413    
1414    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1415        if the pattern had a "must contain" character that was already found in the
1416        earlier partial match, unless partial matching was again requested. For
1417        example, with the pattern /dog.(body)?/, the "must contain" character is
1418        "g". If the first part-match was for the string "dog", restarting with
1419        "sbody" failed. This bug has been fixed.
1420    
1421    13. The string returned by pcre_dfa_exec() after a partial match has been
1422        changed so that it starts at the first inspected character rather than the
1423        first character of the match. This makes a difference only if the pattern
1424        starts with a lookbehind assertion or \b or \B (\K is not supported by
1425        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1426        matching functions compatible, and I think it's the right thing to do.
1427    
1428    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1429        so that the demonstration program is easily available in environments where
1430        PCRE has not been installed from source.
1431    
1432    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1433        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1434        library.
1435    
1436    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1437        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1438        is not the first non-POSIX option to be added. Clearly some people find
1439        these options useful.
1440    
1441    17. If a caller to the POSIX matching function regexec() passes a non-zero
1442        value for nmatch with a NULL value for pmatch, the value of
1443        nmatch is forced to zero.
1444    
1445    18. RunGrepTest did not have a test for the availability of the -u option of
1446        the diff command, as RunTest does. It now checks in the same way as
1447        RunTest, and also checks for the -b option.
1448    
1449    19. If an odd number of negated classes containing just a single character
1450        interposed, within parentheses, between a forward reference to a named
1451        subpattern and the definition of the subpattern, compilation crashed with
1452        an internal error, complaining that it could not find the referenced
1453        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1454        [The bug was that it was starting one character too far in when skipping
1455        over the character class, thus treating the ] as data rather than
1456        terminating the class. This meant it could skip too much.]
1457    
1458    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1459        /g option in pcretest when the pattern contains \K, which makes it possible
1460        to have an empty string match not at the start, even when the pattern is
1461        anchored. Updated pcretest and pcredemo to use this option.
1462    
1463    21. If the maximum number of capturing subpatterns in a recursion was greater
1464        than the maximum at the outer level, the higher number was returned, but
1465        with unset values at the outer level. The correct (outer level) value is
1466        now given.
1467    
1468    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1469        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1470        make it do so. The string so far is captured, making this feature
1471        compatible with Perl.
1472    
1473    23. The tests have been re-organized, adding tests 11 and 12, to make it
1474        possible to check the Perl 5.10 features against Perl 5.10.
1475    
1476    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1477        pattern matches a fixed length string. PCRE did not allow this; now it
1478        does. Neither allows recursion.
1479    
1480    25. I finally figured out how to implement a request to provide the minimum
1481        length of subject string that was needed in order to match a given pattern.
1482        (It was back references and recursion that I had previously got hung up
1483        on.) This code has now been added to pcre_study(); it finds a lower bound
1484        to the length of subject needed. It is not necessarily the greatest lower
1485        bound, but using it to avoid searching strings that are too short does give
1486        some useful speed-ups. The value is available to calling programs via
1487        pcre_fullinfo().
1488    
1489    26. While implementing 25, I discovered to my embarrassment that pcretest had
1490        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1491        study optimizations had never been tested with that matching function.
1492        Oops. What is worse, even when it was passed study data, there was a bug in
1493        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1494        were also very few tests of studied patterns with pcre_dfa_exec().
1495    
1496    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1497        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1498        on the other side of the coin, they are no longer allowed to have different
1499        names, because these cannot be distinguished in PCRE, and this has caused
1500        confusion. (This is a difference from Perl.)
1501    
1502    28. When duplicate subpattern names are present (necessarily with different
1503        numbers, as required by 27 above), and a test is made by name in a
1504        conditional pattern, either for a subpattern having been matched, or for
1505        recursion in such a pattern, all the associated numbered subpatterns are
1506        tested, and the overall condition is true if the condition is true for any
1507        one of them. This is the way Perl works, and is also more like the way
1508        testing by number works.
1509    
1510    
1511    Version 7.9 11-Apr-09
1512    ---------------------
1513    
1514    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1515        (pcretest), all targets were linked against these libraries. This included
1516        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1517        libraries. This caused unwanted dependencies to be created. This problem
1518        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1519        pcretest is linked with readline.
1520    
1521    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1522        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1523        moved outside it again, because FALSE and TRUE are already defined in AIX,
1524        but BOOL is not.
1525    
1526    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1527        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1528    
1529    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1530        hyphens) following file names and line numbers when outputting matching
1531        lines. This is not true; no spaces are inserted. I have also clarified the
1532        wording for the --colour (or --color) option.
1533    
1534    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1535        was not coloured; this is different to GNU grep, so I have changed it to be
1536        the same.
1537    
1538    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1539        each matching line was coloured. Now it goes on to look for further matches
1540        of any of the test patterns, which is the same behaviour as GNU grep.
1541    
1542    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1543        doesn't make sense to accept an empty string match in pcregrep, so I have
1544        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1545        seems to be how GNU grep behaves.
1546    
1547    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1548        start or after a newline", because the conditional assertion was not being
1549        correctly handled. The rule now is that both the assertion and what follows
1550        in the first alternative must satisfy the test.
1551    
1552    9.  If auto-callout was enabled in a pattern with a conditional group whose
1553        condition was an assertion, PCRE could crash during matching, both with
1554        pcre_exec() and pcre_dfa_exec().
1555    
1556    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1557        used for matching.
1558    
1559    11. Unicode property support in character classes was not working for
1560        characters (bytes) greater than 127 when not in UTF-8 mode.
1561    
1562    12. Added the -M command line option to pcretest.
1563    
1564    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1565    
1566    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1567    
1568    16. Added comments and documentation about mis-use of no_arg in the C++
1569        wrapper.
1570    
1571    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1572        from Martin Jerabek that uses macro names for all relevant character and
1573        string constants.
1574    
1575    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1576        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1577        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1578        these, but not everybody uses configure.
1579    
1580    19. A conditional group that had only one branch was not being correctly
1581        recognized as an item that could match an empty string. This meant that an
1582        enclosing group might also not be so recognized, causing infinite looping
1583        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1584        with the subject "ab", where knowledge that the repeated group can match
1585        nothing is needed in order to break the loop.
1586    
1587    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1588        exec(), but without supplying a callout function, matching went wrong.
1589    
1590    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1591        leak if the size of the offset vector was greater than 30. When the vector
1592        is smaller, the saved offsets during recursion go onto a local stack
1593        vector, but for larger vectors malloc() is used. It was failing to free
1594        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1595        error, in fact).
1596    
1597    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1598        heapframe that is used only when UTF-8 support is enabled. This caused no
1599        problem, but was untidy.
1600    
1601    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1602        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1603        included within another project.
1604    
1605    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1606        slightly modified by me:
1607    
1608          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1609              not building pcregrep.
1610    
1611          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1612              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1613    
1614    25. Forward references, both numeric and by name, in patterns that made use of
1615        duplicate group numbers, could behave incorrectly or give incorrect errors,
1616        because when scanning forward to find the reference group, PCRE was not
1617        taking into account the duplicate group numbers. A pattern such as
1618        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1619    
1620    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1621        the feature of a custom pointer more persuasive (as requested by a user).
1622    
1623    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1624        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1625        pre-defined.
1626    
1627    28. Added support for (*UTF8) at the start of a pattern.
1628    
1629    29. Arrange for flags added by the "release type" setting in CMake to be shown
1630        in the configuration summary.
1631    
1632    
1633    Version 7.8 05-Sep-08
1634    ---------------------
1635    
1636    1.  Replaced UCP searching code with optimized version as implemented for Ad
1637        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1638        stage table and inline lookup instead of a function, giving speed ups of 2
1639        to 5 times on some simple patterns that I tested. Permission was given to
1640        distribute the MultiStage2.py script that generates the tables (it's not in
1641        the tarball, but is in the Subversion repository).
1642    
1643    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1644        scripts.
1645    
1646    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1647        a group with a zero qualifier. The result of the study could be incorrect,
1648        or the function might crash, depending on the pattern.
1649    
1650    4.  Caseless matching was not working for non-ASCII characters in back
1651        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1652        It now works when Unicode Property Support is available.
1653    
1654    5.  In pcretest, an escape such as \x{de} in the data was always generating
1655        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1656        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1657        truncation.
1658    
1659    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1660    
1661    7.  Added two (int) casts to pcregrep when printing the difference of two
1662        pointers, in case they are 64-bit values.
1663    
1664    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1665        test 2 if it fails.
1666    
1667    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1668        and a #define of that name to empty if it is not externally set. This is to
1669        allow users of MSVC to set it if necessary.
1670    
1671    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1672        the convenience functions in the pcre_get.c source file.
1673    
1674    11. An option change at the start of a pattern that had top-level alternatives
1675        could cause overwriting and/or a crash. This command provoked a crash in
1676        some environments:
1677    
1678          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1679    
1680        This potential security problem was recorded as CVE-2008-2371.
1681    
1682    12. For a pattern where the match had to start at the beginning or immediately
1683        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1684        pcre_dfa_exec() could read past the end of the passed subject if there was
1685        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1686        pcretest so that it places the subject at the end of its malloc-ed buffer.
1687    
1688    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1689        exec() might read past the end of the data buffer in UTF-8 mode.
1690    
1691    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1692        the data contained the byte 0x85 as part of a UTF-8 character within its
1693        first line. This applied both to normal and DFA matching.
1694    
1695    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1696        /^[^d]*?$/8 failed to match "abc".
1697    
1698    16. Added a missing copyright notice to pcrecpp_internal.h.
1699    
1700    17. Make it more clear in the documentation that values returned from
1701        pcre_exec() in ovector are byte offsets, not character counts.
1702    
1703    18. Tidied a few places to stop certain compilers from issuing warnings.
1704    
1705    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1706        supplied by Stefan Weber. I made a further small update for 7.8 because
1707        there is a change of source arrangements: the pcre_searchfuncs.c module is
1708        replaced by pcre_ucd.c.
1709    
1710    
1711    Version 7.7 07-May-08
1712    ---------------------
1713    
1714    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1715        a string to a long long, pretend we don't even have a long long." This is
1716        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1717    
1718    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1719        pre-7.6 versions, which defined a global no_arg variable instead of putting
1720        it in the RE class. (See also #8 below.)
1721    
1722    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1723        Lopes.
1724    
1725    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1726    
1727        (1) The include/exclude patterns were being applied to the whole pathnames
1728            of files, instead of just to the final components.
1729    
1730        (2) If there was more than one level of directory, the subdirectories were
1731            skipped unless they satisfied the include/exclude conditions. This is
1732            inconsistent with GNU grep (and could even be seen as contrary to the
1733            pcregrep specification - which I improved to make it absolutely clear).
1734            The action now is always to scan all levels of directory, and just
1735            apply the include/exclude patterns to regular files.
1736    
1737    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1738        --exclude_dir in the tests to avoid scanning .svn directories.
1739    
1740    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1741        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1742        doesn't support NULs in patterns.
1743    
1744    7.  Added some missing "const"s to declarations of static tables in
1745        pcre_compile.c and pcre_dfa_exec.c.
1746    
1747    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1748        caused by fix #2  above. (Subsequently also a second patch to fix the
1749        first patch. And a third patch - this was a messy problem.)
1750    
1751    9.  Applied Craig's patch to remove the use of push_back().
1752    
1753    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1754        matching function regexec().
1755    
1756    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1757        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1758        references. PCRE supports relative numbers with this syntax (I don't think
1759        Oniguruma does).
1760    
1761    12. Previously, a group with a zero repeat such as (...){0} was completely
1762        omitted from the compiled regex. However, this means that if the group
1763        was called as a subroutine from elsewhere in the pattern, things went wrong
1764        (an internal error was given). Such groups are now left in the compiled
1765        pattern, with a new opcode that causes them to be skipped at execution
1766        time.
1767    
1768    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1769        to the way PCRE behaves:
1770    
1771        (a) A lone ] character is dis-allowed (Perl treats it as data).
1772    
1773        (b) A back reference to an unmatched subpattern matches an empty string
1774            (Perl fails the current match path).
1775    
1776        (c) A data ] in a character class must be notated as \] because if the
1777            first data character in a class is ], it defines an empty class. (In
1778            Perl it is not possible to have an empty class.) The empty class []
1779            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1780            The negative empty class [^] matches any one character, independently
1781            of the DOTALL setting.
1782    
1783    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1784        non-existent subpattern following a character class starting with ']' and
1785        containing () gave an internal compiling error instead of "reference to
1786        non-existent subpattern". Fortunately, when the pattern did exist, the
1787        compiled code was correct. (When scanning forwards to check for the
1788        existencd of the subpattern, it was treating the data ']' as terminating
1789        the class, so got the count wrong. When actually compiling, the reference
1790        was subsequently set up correctly.)
1791    
1792    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1793        it was being rejected as not supported by pcre_dfa_exec(), even though
1794        other assertions are supported. I have made pcre_dfa_exec() support
1795        (*FAIL).
1796    
1797    16. The implementation of 13c above involved the invention of a new opcode,
1798        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1799        cannot be changed at match time, I realized I could make a small
1800        improvement to matching performance by compiling OP_ALLANY instead of
1801        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1802        on the OP_ANY path.
1803    
1804    17. Compiling pcretest on Windows with readline support failed without the
1805        following two fixes: (1) Make the unistd.h include conditional on
1806        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1807    
1808    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1809        ncurses library to be included for pcretest when ReadLine support is
1810        requested, but also to allow for it to be overridden. This patch came from
1811        Daniel Bergström.
1812    
1813    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1814        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1815        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1816        spotting this.
1817    
1818    
1819    Version 7.6 28-Jan-08
1820    ---------------------
1821    
1822    1.  A character class containing a very large number of characters with
1823        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1824        overflow.
1825    
1826    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1827        HAVE_LONG_LONG is not defined.
1828    
1829    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1830        bring them up to date and include new features. This patch includes:
1831    
1832        - Fixed PH's badly added libz and libbz2 support.
1833        - Fixed a problem with static linking.
1834        - Added pcredemo. [But later removed - see 7 below.]
1835        - Fixed dftables problem and added an option.
1836        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1837            HAVE_LONG_LONG.
1838        - Added readline support for pcretest.
1839        - Added an listing of the option settings after cmake has run.
1840    
1841    4.  A user submitted a patch to Makefile that makes it easy to create
1842        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1843        Makefile.am that cause it to include this special target, without
1844        affecting anything else. Note that the same mingw target plus all
1845        the other distribution libraries and programs are now supported
1846        when configuring with CMake (see 6 below) instead of with
1847        Configure/Make.
1848    
1849    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1850        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1851        exported in the Windows port". It has not yet been confirmed that the patch
1852        solves the problem, but it does no harm.
1853    
1854    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1855        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1856        with CMake, and also correct the comment about stack recursion.
1857    
1858    7.  Remove the automatic building of pcredemo from the ./configure system and
1859        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1860        of a program that users should build themselves after PCRE is installed, so
1861        building it automatically is not really right. What is more, it gave
1862        trouble in some build environments.
1863    
1864    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1865    
1866    
1867    Version 7.5 10-Jan-08
1868    ---------------------
1869    
1870    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1871        values in parens when parsing an RE using the C++ wrapper."
1872    
1873    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1874        Characters greater than 255 were excluded from the class instead of being
1875        included.
1876    
1877    3.  The same bug as (2) above applied to negated POSIX classes such as
1878        [:^space:].
1879    
1880    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1881        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1882        I have changed it.
1883    
1884    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1885        first named subpattern) and a construct such as (?&a) would reference the
1886        first named subpattern whose name started with "a" (in other words, the
1887        length check was missing). Both these problems are fixed. "Subpattern name
1888        expected" is now given for (?&) (a zero-length name), and this patch also
1889        makes it give the same error for \k'' (previously it complained that that
1890        was a reference to a non-existent subpattern).
1891    
1892    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1893        this is right because (?- can be followed by option settings as well as by
1894        digits. I have, however, made the messages clearer.
1895    
1896    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1897        than the number used in the conditional) now cause a compile-time error.
1898        This is actually not compatible with Perl, which accepts such patterns, but
1899        treats the conditional as always being FALSE (as PCRE used to), but it
1900        seems to me that giving a diagnostic is better.
1901    
1902    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1903        and messages.
1904    
1905    9.  Fix two occurrences of "backslash" in comments that should have been
1906        "backspace".
1907    
1908    10. Remove two redundant lines of code that can never be obeyed (their function
1909        was moved elsewhere).
1910    
1911    11. The program that makes PCRE's Unicode character property table had a bug
1912        which caused it to generate incorrect table entries for sequences of
1913        characters that have the same character type, but are in different scripts.
1914        It amalgamated them into a single range, with the script of the first of
1915        them. In other words, some characters were in the wrong script. There were
1916        thirteen such cases, affecting characters in the following ranges:
1917    
1918          U+002b0 - U+002c1
1919          U+0060c - U+0060d
1920          U+0061e - U+00612
1921          U+0064b - U+0065e
1922          U+0074d - U+0076d
1923          U+01800 - U+01805
1924          U+01d00 - U+01d77
1925          U+01d9b - U+01dbf
1926          U+0200b - U+0200f
1927          U+030fc - U+030fe
1928          U+03260 - U+0327f
1929          U+0fb46 - U+0fbb1
1930          U+10450 - U+1049d
1931    
1932    12. The -o option (show only the matching part of a line) for pcregrep was not
1933        compatible with GNU grep in that, if there was more than one match in a
1934        line, it showed only the first of them. It now behaves in the same way as
1935        GNU grep.
1936    
1937    13. If the -o and -v options were combined for pcregrep, it printed a blank
1938        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1939        does the same. The return code can be used to tell if there were any
1940        non-matching lines.
1941    
1942    14. Added --file-offsets and --line-offsets to pcregrep.
1943    
1944    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1945        infinitely looping recursion. The bug was that positive lookaheads were not
1946        being skipped when checking for a possible empty match (negative lookaheads
1947        and both kinds of lookbehind were skipped).
1948    
1949    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1950        inclusion of <windows.h> to before rather than after the definition of
1951        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1952    
1953    17. Specifying a possessive quantifier with a specific limit for a Unicode
1954        character property caused pcre_compile() to compile bad code, which led at
1955        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1956        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1957        caused the error; without that there was no problem.
1958    
1959    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1960    
1961    19. Added --enable-pcretest-libreadline.
1962    
1963    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1964        RE::GlobalReplace(). As a result, the number of replacements returned was
1965        double what it should be. I removed one of the increments, but Craig sent a
1966        later patch that removed the other one (the right fix) and added unit tests
1967        that check the return values (which was not done before).
1968    
1969    21. Several CMake things:
1970    
1971        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1972            the names libpcre and libpcreposix, not just pcre and pcreposix.
1973    
1974        (2) The above change means that pcretest and pcregrep are now correctly
1975            linked with the newly-built libraries, not previously installed ones.
1976    
1977        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1978    
1979    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1980        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1981        UTF-8 newline character). The key issue is that the pattern starts .*;
1982        this means that the match must be either at the beginning, or after a
1983        newline. The bug was in the code for advancing after a failed match and
1984        checking that the new position followed a newline. It was not taking
1985        account of UTF-8 characters correctly.
1986    
1987    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1988        character classes. PCRE was not treating the sequence [:...:] as a
1989        character class unless the ... were all letters. Perl, however, seems to
1990        allow any characters between [: and :], though of course it rejects as
1991        unknown any "names" that contain non-letters, because all the known class
1992        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1993        for example, whereas PCRE did not - it did not recognize a POSIX character
1994        class. This seemed a bit dangerous, so the code has been changed to be
1995        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1996        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1997        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1998        Perl does, and where it didn't before.
1999    
2000    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2001        Windows environments %n is disabled by default.
2002    
2003    
2004    Version 7.4 21-Sep-07
2005    ---------------------
2006    
2007    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2008        means that a class such as [\s] counted as "explicit reference to CR or
2009        LF". That isn't really right - the whole point of the change was to try to
2010        help when there was an actual mention of one of the two characters. So now
2011        the change happens only if \r or \n (or a literal CR or LF) character is
2012        encountered.
2013    
2014    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2015        of both had grown to the point where there were only 3 bits left.
2016        Fortunately, there was spare space in the data structure, and so I have
2017        moved the internal flags into a new 16-bit field to free up more option
2018        bits.
2019    
2020    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2021        but did not set the internal JCHANGED flag - either of these is enough to
2022        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2023        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2024        start sets both bits.
2025    
2026    4.  Added options (at build time, compile time, exec time) to change \R from
2027        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2028    
2029    5.  doc/pcresyntax.html was missing from the distribution.
2030    
2031    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2032        compatibility, even though it is no longer used.
2033    
2034    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2035        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2036        windows.h file is present (where different names are used). [This was
2037        reversed later after testing - see 16 below.]
2038    
2039    8.  Changed all #include <config.h> to #include "config.h". There were also
2040        some further <pcre.h> cases that I changed to "pcre.h".
2041    
2042    9.  When pcregrep was used with the --colour option, it missed the line ending
2043        sequence off the lines that it output.
2044    
2045    10. It was pointed out to me that arrays of string pointers cause lots of
2046        relocations when a shared library is dynamically loaded. A technique of
2047        using a single long string with a table of offsets can drastically reduce
2048        these. I have refactored PCRE in four places to do this. The result is
2049        dramatic:
2050    
2051          Originally:                          290
2052          After changing UCP table:            187
2053          After changing error message table:   43
2054          After changing table of "verbs"       36
2055          After changing table of Posix names   22
2056    
2057        Thanks to the folks working on Gregex for glib for this insight.
2058    
2059    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2060        unicode-properties was also set.
2061    
2062    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2063    
2064    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2065        checked only for CRLF.
2066    
2067    14. Added casts to pcretest.c to avoid compiler warnings.
2068    
2069    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2070    
2071    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2072        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2073        entirely. This removes changes made in 7 above.
2074    
2075    17. The CMake files have been updated, and there is now more information about
2076        building with CMake in the NON-UNIX-USE document.
2077    
2078    
2079    Version 7.3 28-Aug-07
2080    ---------------------
2081    
2082     1. In the rejigging of the build system that eventually resulted in 7.1, the
2083        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2084        brackets there is not right, since it causes compilers to look for an
2085        installed pcre.h, not the version that is in the source that is being
2086        compiled (which of course may be different). I have changed it back to:
2087    
2088          #include "pcre.h"
2089    
2090        I have a vague recollection that the change was concerned with compiling in
2091        different directories, but in the new build system, that is taken care of
2092        by the VPATH setting the Makefile.
2093    
2094     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2095        when the subject happened to end in the byte 0x85 (e.g. if the last
2096        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2097        characters but of course it shouldn't be taken as a newline when it is part
2098        of another character. The bug was that, for an unlimited repeat of . in
2099        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2100        characters when looking for a newline.
2101    
2102     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2103    
2104     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2105        in debug output.
2106    
2107     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2108        long printing in the pcrecpp unittest when running under MinGW.
2109    
2110     6. ESC_K was left out of the EBCDIC table.
2111    
2112     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2113        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2114        limit also applies to "virtual nesting" when a pattern is recursive, and in
2115        this case 1000 isn't so big. I have been able to remove this limit at the
2116        expense of backing off one optimization in certain circumstances. Normally,
2117        when pcre_exec() would call its internal match() function recursively and
2118        immediately return the result unconditionally, it uses a "tail recursion"
2119        feature to save stack. However, when a subpattern that can match an empty
2120        string has an unlimited repetition quantifier, it no longer makes this
2121        optimization. That gives it a stack frame in which to save the data for
2122        checking that an empty string has been matched. Previously this was taken
2123        from the 1000-entry workspace that had been reserved. So now there is no
2124        explicit limit, but more stack is used.
2125    
2126     8. Applied Daniel's patches to solve problems with the import/export magic
2127        syntax that is required for Windows, and which was going wrong for the
2128        pcreposix and pcrecpp parts of the library. These were overlooked when this
2129        problem was solved for the main library.
2130    
2131     9. There were some crude static tests to avoid integer overflow when computing
2132        the size of patterns that contain repeated groups with explicit upper
2133        limits. As the maximum quantifier is 65535, the maximum group length was
2134        set at 30,000 so that the product of these two numbers did not overflow a
2135        32-bit integer. However, it turns out that people want to use groups that
2136        are longer than 30,000 bytes (though not repeat them that many times).
2137        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2138        made it possible to implement the integer overflow checks in a much more
2139        dynamic way, which I have now done. The artificial limitation on group
2140        length has been removed - we now have only the limit on the total length of
2141        the compiled pattern, which depends on the LINK_SIZE setting.
2142    
2143    10. Fixed a bug in the documentation for get/copy named substring when
2144        duplicate names are permitted. If none of the named substrings are set, the
2145        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2146        empty string.
2147    
2148    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2149        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2150        because the ] is interpreted as the first data character and the
2151        terminating ] is not found. PCRE has been made compatible with Perl in this
2152        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2153        cause memory overwriting.
2154    
2155    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2156        string has been matched (to stop an infinite loop). It was not recognizing
2157        a conditional subpattern that could match an empty string if that
2158        subpattern was within another subpattern. For example, it looped when
2159        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2160        condition was not nested. This bug has been fixed.
2161    
2162    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2163        past the start of the subject in the presence of bytes with the top bit
2164        set, for example "\x8aBCD".
2165    
2166    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2167        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2168    
2169    14. Optimized (?!) to (*FAIL).
2170    
2171    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2172        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2173        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2174        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2175        does: it's just the validity check that is more restrictive.
2176    
2177    16. Inserted checks for integer overflows during escape sequence (backslash)
2178        processing, and also fixed erroneous offset values for syntax errors during
2179        backslash processing.
2180    
2181    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2182        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2183    
2184    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2185        caused an overrun.
2186    
2187    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2188        something other than just ASCII characters) inside a group that had an
2189        unlimited repeat caused a loop at compile time (while checking to see
2190        whether the group could match an empty string).
2191    
2192    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2193        [\P{Any}] did so. (Error in the code for printing property names.)
2194    
2195    21. An orphan \E inside a character class could cause a crash.
2196    
2197    22. A repeated capturing bracket such as (A)? could cause a wild memory
2198        reference during compilation.
2199    
2200    23. There are several functions in pcre_compile() that scan along a compiled
2201        expression for various reasons (e.g. to see if it's fixed length for look
2202        behind). There were bugs in these functions when a repeated \p or \P was
2203        present in the pattern. These operators have additional parameters compared
2204        with \d, etc, and these were not being taken into account when moving along
2205        the compiled data. Specifically:
2206    
2207        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2208            length.
2209    
2210        (b) An item such as \pL+ within a repeated group could cause crashes or
2211            loops.
2212    
2213        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2214            "reference to non-existent subpattern" error.
2215    
2216        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2217    
2218    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2219        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2220    
2221    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2222    
2223    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2224        character were causing crashes (broken optimization).
2225    
2226    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2227        \p or \P) caused a compile-time loop.
2228    
2229    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2230        break. For example, the unstudied pattern [\r\n]A does not match the string
2231        "\r\nA" because change 7.0/46 below moves the current point on by two
2232        characters after failing to match at the start. However, the pattern \nA
2233        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2234        the same is true. There doesn't seem any very clean way out of this, but
2235        what I have chosen to do makes the common cases work: PCRE now takes note
2236        of whether there can be an explicit match for \r or \n anywhere in the
2237        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2238        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2239        pattern has explicit CR or LF references.
2240    
2241    29. Added (*CR) etc for changing newline setting at start of pattern.
2242    
2243    
2244    Version 7.2 19-Jun-07
2245    ---------------------
2246    
2247     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2248        which is apparently normally available under Windows.
2249    
2250     2. Re-jig the pcregrep tests with different newline settings in an attempt
2251        to make them independent of the local environment's newline setting.
2252    
2253     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2254    
2255     4. Some of the "internals" tests were previously cut out when the link size
2256        was not 2, because the output contained actual offsets. The recent new
2257        "Z" feature of pcretest means that these can be cut out, making the tests
2258        usable with all link sizes.
2259    
2260     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2261        stack recursion. This gives a massive performance boost under BSD, but just
2262        a small improvement under Linux. However, it saves one field in the frame
2263        in all cases.
2264    
2265     6. Added more features from the forthcoming Perl 5.10:
2266    
2267        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2268            recursion call. It refers to the nth most recently opened parentheses.
2269    
2270        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2271            to be opened parentheses.
2272    
2273        (c) Conditions that refer to capturing parentheses can be specified
2274            relatively, for example, (?(-2)... or (?(+3)...
2275    
2276        (d) \K resets the start of the current match so that everything before
2277            is not part of it.
2278    
2279        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2280    
2281        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2282            reference syntax.
2283    
2284        (g) (?| introduces a group in which the numbering of parentheses in each
2285            alternative starts with the same number.
2286    
2287        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2288    
2289     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2290        PCRE_INFO_JCHANGED.
2291    
2292     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2293        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2294        for detecting groups that can match an empty string.
2295    
2296     9. A pattern with a very large number of alternatives (more than several
2297        hundred) was running out of internal workspace during the pre-compile
2298        phase, where pcre_compile() figures out how much memory will be needed. A
2299        bit of new cunning has reduced the workspace needed for groups with
2300        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2301        workspace instead of running out of the 4096 that are available.
2302    
2303    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2304    
2305    11. Applied patch from Google to remove an optimization that didn't quite work.
2306        The report of the bug said:
2307    
2308          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2309          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2310          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2311    
2312    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2313        it matched the wrong number of bytes.
2314    
2315    
2316    Version 7.1 24-Apr-07
2317    ---------------------
2318    
2319     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2320        that is more "standard", making use of automake and other Autotools. There
2321        is some re-arrangement of the files and adjustment of comments consequent
2322        on this.
2323    
2324     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2325        for recursive directory scanning broke on some systems because the files
2326        are not scanned in any specific order and on different systems the order
2327        was different. A call to "sort" has been inserted into RunGrepTest for the
2328        approprate test as a short-term fix. In the longer term there may be an
2329        alternative.
2330    
2331     3. I had an email from Eric Raymond about problems translating some of PCRE's
2332        man pages to HTML (despite the fact that I distribute HTML pages, some
2333        people do their own conversions for various reasons). The problems
2334        concerned the use of low-level troff macros .br and .in. I have therefore
2335        removed all such uses from the man pages (some were redundant, some could
2336        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2337        HTML has been updated to handle .nf/.fi and to complain if it encounters
2338        .br or .in.
2339    
2340     4. Updated comments in configure.ac that get placed in config.h.in and also
2341        arranged for config.h to be included in the distribution, with the name
2342        config.h.generic, for the benefit of those who have to compile without
2343        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2344    
2345     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2346        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2347        makevp.bat for the current PCRE, using the additional files
2348        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2349    
2350     6. A Windows user reported a minor discrepancy with test 2, which turned out
2351        to be caused by a trailing space on an input line that had got lost in his
2352        copy. The trailing space was an accident, so I've just removed it.
2353    
2354     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2355        that is needed.
2356    
2357     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2358        as "const" (a) because they are and (b) because it helps the PHP
2359        maintainers who have recently made a script to detect big data structures
2360        in the php code that should be moved to the .rodata section. I remembered
2361        to update Builducptable as well, so it won't revert if ucptable.h is ever
2362        re-created.
2363    
2364     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2365        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2366        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2367        support is not required. This saves 1.5-2K of code, which is important in
2368        some applications.
2369    
2370        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2371        so as not to refer to the tables, even though these functions will never be
2372        called when UTF-8 support is disabled. Otherwise there are problems with a
2373        shared library.
2374    
2375    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2376    
2377        (a) It was defining its arguments as char * instead of void *.
2378    
2379        (b) It was assuming that all moves were upwards in memory; this was true
2380            a long time ago when I wrote it, but is no longer the case.
2381    
2382        The emulated memove() is provided for those environments that have neither
2383        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2384        is clearly not the case, as these two bugs were recently reported.
2385    
2386    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2387        and Detrail to create the HTML documentation, the .txt form of the man
2388        pages, and it removes trailing spaces from listed files. It also creates
2389        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2390        case, it wraps all the #defines with #ifndefs. This script should be run
2391        before "make dist".
2392    
2393    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2394        with Unicode property support.
2395    
2396        (a) For a maximizing quantifier, if the two different cases of the
2397            character were of different lengths in their UTF-8 codings (there are
2398            some cases like this - I found 11), and the matching function had to
2399            back up over a mixture of the two cases, it incorrectly assumed they
2400            were both the same length.
2401    
2402        (b) When PCRE was configured to use the heap rather than the stack for
2403            recursion during matching, it was not correctly preserving the data for
2404            the other case of a UTF-8 character when checking ahead for a match
2405            while processing a minimizing repeat. If the check also involved
2406            matching a wide character, but failed, corruption could cause an
2407            erroneous result when trying to check for a repeat of the original
2408            character.
2409    
2410    13. Some tidying changes to the testing mechanism:
2411    
2412        (a) The RunTest script now detects the internal link size and whether there
2413            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2414            values substituted by "configure". (The RunGrepTest script already did
2415            this for UTF-8.) The configure.ac script no longer substitutes the
2416            relevant variables.
2417    
2418        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2419            with length and offset values. This means that the output is different
2420            for different internal link sizes. Test 2 is skipped for link sizes
2421            other than 2 because of this, bypassing the problem. Unfortunately,
2422            there was also a test in test 3 (the locale tests) that used /B and
2423            failed for link sizes other than 2. Rather than cut the whole test out,
2424            I have added a new /Z option to pcretest that replaces the length and
2425            offset values with spaces. This is now used to make test 3 independent
2426            of link size. (Test 2 will be tidied up later.)
2427    
2428    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2429        segmentation fault instead of returning the appropriate error message.
2430    
2431    15. In multiline mode when the newline sequence was set to "any", the pattern
2432        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2433        This doesn't seem right; it now treats the CRLF combination as the line
2434        ending, and so does not match in that case. It's only a pattern such as ^$
2435        that would hit this one: something like ^ABC$ would have failed after \r
2436        and then tried again after \r\n.
2437    
2438    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2439        in an attempt to make files that differ only in their line terminators
2440        compare equal. This works on Linux.
2441    
2442    17. Under certain error circumstances pcregrep might try to free random memory
2443        as it exited. This is now fixed, thanks to valgrind.
2444    
2445    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2446        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2447        was because its rules for how to advance for /g after matching an empty
2448        string at the end of a line did not allow for this case. They now check for
2449        it specially.
2450    
2451    20. pcretest is supposed to handle patterns and data of any length, by
2452        extending its buffers when necessary. It was getting this wrong when the
2453        buffer for a data line had to be extended.
2454    
2455    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2456        CRLF as a newline sequence.
2457    
2458    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2459        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2460        I have nevertheless tidied it up.
2461    
2462    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2463    
2464    24. Added a man page for pcre-config.
2465    
2466    
2467    Version 7.0 19-Dec-06
2468    ---------------------
2469    
2470     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2471        moving to gcc 4.1.1.
2472    
2473     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2474        sys/time.h, which is documented as needed for this function. It doesn't
2475        seem to matter on Linux, but it showed up on some releases of OS X.
2476    
2477     3. It seems that there are systems where bytes whose values are greater than
2478        127 match isprint() in the "C" locale. The "C" locale should be the
2479        default when a C program starts up. In most systems, only ASCII printing
2480        characters match isprint(). This difference caused the output from pcretest
2481        to vary, making some of the tests fail. I have changed pcretest so that:
2482    
2483        (a) When it is outputting text in the compiled version of a pattern, bytes
2484            other than 32-126 are always shown as hex escapes.
2485    
2486        (b) When it is outputting text that is a matched part of a subject string,
2487            it does the same, unless a different locale has been set for the match
2488            (using the /L modifier). In this case, it uses isprint() to decide.
2489    
2490     4. Fixed a major bug that caused incorrect computation of the amount of memory
2491        required for a compiled pattern when options that changed within the
2492        pattern affected the logic of the preliminary scan that determines the
2493        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2494        that the computed length was too small. The symptoms of this bug were
2495        either the PCRE error "internal error: code overflow" from pcre_compile(),
2496        or a glibc crash with a message such as "pcretest: free(): invalid next
2497        size (fast)". Examples of patterns that provoked this bug (shown in
2498        pcretest format) are:
2499    
2500          /(?-x: )/x
2501          /(?x)(?-x: \s*#\s*)/
2502          /((?i)[\x{c0}])/8
2503          /(?i:[\x{c0}])/8
2504    
2505        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2506        is now done differently.
2507    
2508     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2509        wrapper classes; (b) implement a new function in the C++ scanner that is
2510        more efficient than the old way of doing things because it avoids levels of
2511        recursion in the regex matching; (c) add a paragraph to the documentation
2512        for the FullMatch() function.
2513    
2514     6. The escape sequence \n was being treated as whatever was defined as
2515        "newline". Not only was this contrary to the documentation, which states
2516        that \n is character 10 (hex 0A), but it also went horribly wrong when
2517        "newline" was defined as CRLF. This has been fixed.
2518    
2519     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2520        was being set to -1 for the "end of line" case (supposedly a value that no
2521        character can have). Though this value is never used (the check for end of
2522        line is "zero bytes in current character"), it caused compiler complaints.
2523        I've changed it to 0xffffffff.
2524    
2525     8. In pcre_version.c, the version string was being built by a sequence of
2526        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2527        string (as it is for production releases) called a macro with an empty
2528        argument. The C standard says the result of this is undefined. The gcc
2529        compiler treats it as an empty string (which was what was wanted) but it is
2530        reported that Visual C gives an error. The source has been hacked around to
2531        avoid this problem.
2532    
2533     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2534        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2535        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2536        of them did).
2537    
2538    10. Originally, pcretest opened its input and output without "b"; then I was
2539        told that "b" was needed in some environments, so it was added for release
2540        5.0 to both the input and output. (It makes no difference on Unix-like
2541        systems.) Later I was told that it is wrong for the input on Windows. I've
2542        now abstracted the modes into two macros, to make it easier to fiddle with
2543        them, and removed "b" from the input mode under Windows.
2544    
2545    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2546    
2547    12. Added -help and --help to pcretest as an official way of being reminded
2548        of the options.
2549    
2550    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2551        and pcrecpp.cc because they annoy compilers at high warning levels.
2552    
2553    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2554    
2555    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2556        scripts are not C programs :-) and which was not noticed because it works
2557        on Linux.
2558    
2559    16. pcretest is supposed to handle any length of pattern and data line (as one
2560        line or as a continued sequence of lines) by extending its input buffer if
2561        necessary. This feature was broken for very long pattern lines, leading to
2562        a string of junk being passed to pcre_compile() if the pattern was longer
2563        than about 50K.
2564    
2565    17. I have done a major re-factoring of the way pcre_compile() computes the
2566        amount of memory needed for a compiled pattern. Previously, there was code
2567        that made a preliminary scan of the pattern in order to do this. That was
2568        OK when PCRE was new, but as the facilities have expanded, it has become
2569        harder and harder to keep it in step with the real compile phase, and there
2570        have been a number of bugs (see for example, 4 above). I have now found a
2571        cunning way of running the real compile function in a "fake" mode that
2572        enables it to compute how much memory it would need, while actually only
2573        ever using a few hundred bytes of working memory and without too many
2574        tests of the mode. This should make future maintenance and development
2575        easier. A side effect of this work is that the limit of 200 on the nesting
2576        depth of parentheses has been removed (though this was never a serious
2577        limitation, I suspect). However, there is a downside: pcre_compile() now
2578        runs more slowly than before (30% or more, depending on the pattern). I
2579        hope this isn't a big issue. There is no effect on runtime performance.
2580    
2581    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2582        newline (only possible for the last line of a file) and it was a
2583        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2584    
2585    19. Added additional timing features to pcretest. (1) The -tm option now times
2586        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2587        separate command line item, by a number that specifies the number of
2588        repeats to use when timing. The default is 50000; this gives better
2589        precision, but takes uncomfortably long for very large patterns.
2590    
2591    20. Extended pcre_study() to be more clever in cases where a branch of a
2592        subpattern has no definite first character. For example, (a*|b*)[cd] would
2593        previously give no result from pcre_study(). Now it recognizes that the
2594        first character must be a, b, c, or d.
2595    
2596    21. There was an incorrect error "recursive call could loop indefinitely" if
2597        a subpattern (or the entire pattern) that was being tested for matching an
2598        empty string contained only one non-empty item after a nested subpattern.
2599        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2600        incorrectly, because the \d was being skipped in the check.
2601    
2602    22. The pcretest program now has a new pattern option /B and a command line
2603        option -b, which is equivalent to adding /B to every pattern. This causes
2604        it to show the compiled bytecode, without the additional information that
2605        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2606        is the same as /B/I).
2607    
2608    23. A new optimization is now able automatically to treat some sequences such
2609        as a*b as a*+b. More specifically, if something simple (such as a character
2610        or a simple class like \d) has an unlimited quantifier, and is followed by
2611        something that cannot possibly match the quantified thing, the quantifier
2612        is automatically "possessified".
2613    
2614    24. A recursive reference to a subpattern whose number was greater than 39
2615        went wrong under certain circumstances in UTF-8 mode. This bug could also
2616        have affected the operation of pcre_study().
2617    
2618    25. Realized that a little bit of performance could be had by replacing
2619        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2620    
2621    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2622    
2623    27. Possessive quantifiers such as a++ were previously implemented by turning
2624        them into atomic groups such as ($>a+). Now they have their own opcodes,
2625        which improves performance. This includes the automatically created ones
2626        from 23 above.
2627    
2628    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2629        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2630        the first matched character to be a colon. This applied both to named and
2631        numbered groups.
2632    
2633    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2634    
2635    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2636        building PCRE on Windows easier, so I have included it in the distribution.
2637    
2638    31. There is now a check in pcretest against a ridiculously large number being
2639        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2640        loop, the loop is abandoned.
2641    
2642    32. Forward references to subpatterns in conditions such as (?(2)...) where
2643        subpattern 2 is defined later cause pcre_compile() to search forwards in
2644        the pattern for the relevant set of parentheses. This search went wrong
2645        when there were unescaped parentheses in a character class, parentheses
2646        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2647    
2648    33. "Subroutine" calls and backreferences were previously restricted to
2649        referencing subpatterns earlier in the regex. This restriction has now
2650        been removed.
2651    
2652    34. Added a number of extra features that are going to be in Perl 5.10. On the
2653        whole, these are just syntactic alternatives for features that PCRE had
2654        previously implemented using the Python syntax or my own invention. The
2655        other formats are all retained for compatibility.
2656    
2657        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2658            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2659            also .NET compatible.
2660    
2661        (b) A recursion or subroutine call to a named group can now be defined as
2662            (?&name) as well as (?P>name).
2663    
2664        (c) A backreference to a named group can now be defined as \k<name> or
2665            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2666            5.10, are also .NET compatible.
2667    
2668        (d) A conditional reference to a named group can now use the syntax
2669            (?(<name>) or (?('name') as well as (?(name).
2670    
2671        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2672            groups (named and numbered) that are never evaluated inline, but can be
2673            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2674            is always false. There may be only one alternative in such a group.
2675    
2676        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2677            as the simple (?(R). The condition is true only if the most recent
2678            recursion is that of the given number or name. It does not search out
2679            through the entire recursion stack.
2680    
2681        (g) The escape \gN or \g{N} has been added, where N is a positive or
2682            negative number, specifying an absolute or relative reference.
2683    
2684    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2685        some "unreachable code" warnings.
2686    
2687    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2688        things, this adds five new scripts.
2689    
2690    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2691        There were also incompatibilities regarding the handling of \Q..\E inside
2692        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2693        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2694    
2695    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2696        matches an empty string, and forcibly breaks the loop. There were bugs in
2697        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2698        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2699        separate and independent bugs (that affected different cases) have been
2700        fixed.
2701    
2702    39. Refactored the code to abolish the use of different opcodes for small
2703        capturing bracket numbers. This is a tidy that I avoided doing when I
2704        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2705        The new approach is not only tidier, it makes it possible to reduce the
2706        memory needed to fix the previous bug (38).
2707    
2708    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2709        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2710        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2711        mode.
2712    
2713    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2714        report.
2715    
2716    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2717        copy construction and assignment in the C++ wrapper.
2718    
2719    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2720        couple of bugs that could have given wrong results in the "--newline=crlf"
2721        case.
2722    
2723    44. Added a number of casts and did some reorganization of signed/unsigned int
2724        variables following suggestions from Dair Grant. Also renamed the variable
2725        "this" as "item" because it is a C++ keyword.
2726    
2727    45. Arranged for dftables to add
2728    
2729          #include "pcre_internal.h"
2730    
2731        to pcre_chartables.c because without it, gcc 4.x may remove the array
2732        definition from the final binary if PCRE is built into a static library and
2733        dead code stripping is activated.
2734    
2735    46. For an unanchored pattern, if a match attempt fails at the start of a
2736        newline sequence, and the newline setting is CRLF or ANY, and the next two
2737        characters are CRLF, advance by two characters instead of one.
2738    
2739    
2740    Version 6.7 04-Jul-06
2741    ---------------------
2742    
2743     1. In order to handle tests when input lines are enormously long, pcretest has
2744        been re-factored so that it automatically extends its buffers when
2745        necessary. The code is crude, but this _is_ just a test program. The
2746        default size has been increased from 32K to 50K.
2747    
2748     2. The code in pcre_study() was using the value of the re argument before
2749        testing it for NULL. (Of course, in any sensible call of the function, it
2750        won't be NULL.)
2751    
2752     3. The memmove() emulation function in pcre_internal.h, which is used on
2753        systems that lack both memmove() and bcopy() - that is, hardly ever -
2754        was missing a "static" storage class specifier.
2755    
2756     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2757        containing an extended class (one that cannot be represented by a bitmap
2758        because it contains high-valued characters or Unicode property items, e.g.
2759        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2760        pattern, but PCRE should not loop if you do not (it no longer does).
2761        [Detail: two cases were found: (a) a repeated subpattern containing an
2762        extended class; (b) a recursive reference to a subpattern that followed a
2763        previous extended class. It wasn't skipping over the extended class
2764        correctly when UTF-8 mode was not set.]
2765    
2766     5. A negated single-character class was not being recognized as fixed-length
2767        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2768        compile error "lookbehind assertion is not fixed length".
2769    
2770     6. The RunPerlTest auxiliary script was showing an unexpected difference
2771        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2772        write a Perl script that can interpret lines of an input file either as
2773        byte characters or as UTF-8, which is what "perltest" was being required to
2774        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2775        can't do is switch easily at run time between having the "use utf8;" pragma
2776        or not. In the end, I fudged it by using the RunPerlTest script to insert
2777        "use utf8;" explicitly for the UTF-8 tests.
2778    
2779     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2780        the end of the subject string, contrary to the documentation and to what
2781        Perl does. This was true of both matching functions. Now it matches only at
2782        the start of the subject and immediately after *internal* newlines.
2783    
2784     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2785        a pointer to an int instead of a pointer to an unsigned long int. This
2786        caused problems on 64-bit systems.
2787    
2788     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2789        instance of the 'standard' template library not being so standard".
2790    
2791    10. There was no check on the number of named subpatterns nor the maximum
2792        length of a subpattern name. The product of these values is used to compute
2793        the size of the memory block for a compiled pattern. By supplying a very
2794        long subpattern name and a large number of named subpatterns, the size
2795        computation could be caused to overflow. This is now prevented by limiting
2796        the length of names to 32 characters, and the number of named subpatterns
2797        to 10,000.
2798    
2799    11. Subpatterns that are repeated with specific counts have to be replicated in
2800        the compiled pattern. The size of memory for this was computed from the
2801        length of the subpattern and the repeat count. The latter is limited to
2802        65535, but there was no limit on the former, meaning that integer overflow
2803        could in principle occur. The compiled length of a repeated subpattern is
2804        now limited to 30,000 bytes in order to prevent this.
2805    
2806    12. Added the optional facility to have named substrings with the same name.
2807    
2808    13. Added the ability to use a named substring as a condition, using the
2809        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2810        are numbers (not recommended). Forward references are permitted.
2811    
2812    14. Added forward references in named backreferences (if you see what I mean).
2813    
2814    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2815        pattern could run off the end of the subject. For example, the pattern
2816        "(?s)(.{1,5})"8 did this with the subject "ab".
2817    
2818    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2819        PCRE_CASELESS was set when matching characters that were quantified with ?
2820        or *.
2821    
2822    17. A character class other than a single negated character that had a minimum
2823        but no maximum quantifier - for example [ab]{6,} - was not handled
2824        correctly by pce_dfa_exec(). It would match only one character.
2825    
2826    18. A valid (though odd) pattern that looked like a POSIX character
2827        class but used an invalid character after [ (for example [[,abc,]]) caused
2828        pcre_compile() to give the error "Failed: internal error: code overflow" or
2829        in some cases to crash with a glibc free() error. This could even happen if
2830        the pattern terminated after [[ but there just happened to be a sequence of
2831        letters, a binary zero, and a closing ] in the memory that followed.
2832    
2833    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2834        over the years. Originally (before any Unicode support), just the bottom 8
2835        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2836        output from "man perlunicode" includes this:
2837    
2838          The regular expression compiler produces polymorphic opcodes.  That
2839          is, the pattern adapts to the data and automatically switches to
2840          the Unicode character scheme when presented with Unicode data--or
2841          instead uses a traditional byte scheme when presented with byte
2842          data.
2843    
2844        Sadly, a wide octal escape does not cause a switch, and in a string with
2845        no other multibyte characters, these octal escapes are treated as before.
2846        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2847        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2848        Unicode string.
2849    
2850        I have not perpetrated such confusion in PCRE. Up till now, it took just
2851        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2852        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2853        translate to the appropriate multibyte character.
2854    
2855    29. Applied some refactoring to reduce the number of warnings from Microsoft
2856        and Borland compilers. This has included removing the fudge introduced
2857        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2858        a warning about an unused variable.
2859    
2860    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2861        characters since release 4.0, because Perl (from release 5.004) does not.
2862        [Or at least, is documented not to: some releases seem to be in conflict
2863        with the documentation.] However, when a pattern was studied with
2864        pcre_study() and all its branches started with \s, PCRE still included VT
2865        as a possible starting character. Of course, this did no harm; it just
2866        caused an unnecessary match attempt.
2867    
2868    22. Removed a now-redundant internal flag bit that recorded the fact that case
2869        dependency changed within the pattern. This was once needed for "required
2870        byte" processing, but is no longer used. This recovers a now-scarce options
2871        bit. Also moved the least significant internal flag bit to the most-
2872        significant bit of the word, which was not previously used (hangover from
2873        the days when it was an int rather than a uint) to free up another bit for
2874        the future.
2875    
2876    23. Added support for CRLF line endings as well as CR and LF. As well as the
2877        default being selectable at build time, it can now be changed at runtime
2878        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2879        specify that it is scanning data with non-default line endings.
2880    
2881    24. Changed the definition of CXXLINK to make it agree with the definition of
2882        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2883    
2884    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2885        recursions. This makes a big different to stack usage for some patterns.
2886    
2887    26. If a subpattern containing a named recursion or subroutine reference such
2888        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2889        the space required for the compiled pattern went wrong and gave too small a
2890        value. Depending on the environment, this could lead to "Failed: internal
2891        error: code overflow at offset 49" or "glibc detected double free or
2892        corruption" errors.
2893    
2894    27. Applied patches from Google (a) to support the new newline modes and (b) to
2895        advance over multibyte UTF-8 characters in GlobalReplace.
2896    
2897    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2898        difference for some implementation of PCRE in some Windows version.
2899    
2900    29. Added some extra testing facilities to pcretest:
2901    
2902        \q<number>   in a data line sets the "match limit" value
2903        \Q<number>   in a data line sets the "match recursion limt" value
2904        -S <number>  sets the stack size, where <number> is in megabytes
2905    
2906        The -S option isn't available for Windows.
2907    
2908    
2909    Version 6.6 06-Feb-06
2910    ---------------------
2911    
2912     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2913        in pcreposix.h. I have copied the definition from pcre.h.
2914    
2915     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2916        because pcre.h is no longer a built file.
2917    
2918     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2919        not normally included in the compiled code.
2920    
2921    
2922    Version 6.5 01-Feb-06
2923    ---------------------
2924    
2925     1. When using the partial match feature with pcre_dfa_exec(), it was not
2926        anchoring the second and subsequent partial matches at the new starting
2927        point. This could lead to incorrect results. For example, with the pattern
2928        /1234/, partially matching against "123" and then "a4" gave a match.
2929    
2930     2. Changes to pcregrep:
2931    
2932        (a) All non-match returns from pcre_exec() were being treated as failures
2933            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2934            error message is output. Some extra information is given for the
2935            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2936            probably the only errors that are likely to be caused by users (by
2937            specifying a regex that has nested indefinite repeats, for instance).
2938            If there are more than 20 of these errors, pcregrep is abandoned.
2939    
2940        (b) A binary zero was treated as data while matching, but terminated the
2941            output line if it was written out. This has been fixed: binary zeroes
2942            are now no different to any other data bytes.
2943    
2944        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2945            used to set a locale for matching. The --locale=xxxx long option has
2946            been added (no short equivalent) to specify a locale explicitly on the
2947            pcregrep command, overriding the environment variables.
2948    
2949        (d) When -B was used with -n, some line numbers in the output were one less
2950            than they should have been.
2951    
2952        (e) Added the -o (--only-matching) option.
2953    
2954        (f) If -A or -C was used with -c (count only), some lines of context were
2955            accidentally printed for the final match.
2956    
2957        (g) Added the -H (--with-filename) option.
2958    
2959        (h) The combination of options -rh failed to suppress file names for files
2960            that were found from directory arguments.
2961    
2962        (i) Added the -D (--devices) and -d (--directories) options.
2963    
2964        (j) Added the -F (--fixed-strings) option.
2965    
2966        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2967    
2968        (l) Added the --colo(u)r option.
2969    
2970        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2971            is not present by default.
2972    
2973     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2974        items such as (?R) or (?1), when the recursion could match a number of
2975        alternatives. If it matched one of the alternatives, but subsequently,
2976        outside the recursion, there was a failure, the code tried to back up into
2977        the recursion. However, because of the way PCRE is implemented, this is not
2978        possible, and the result was an incorrect result from the match.
2979    
2980        In order to prevent this happening, the specification of recursion has
2981        been changed so that all such subpatterns are automatically treated as
2982        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2983    
2984     4. I had overlooked the fact that, in some locales, there are characters for
2985        which isalpha() is true but neither isupper() nor islower() are true. In
2986        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2987        and ordfeminine) are like this. This affected the treatment of \w and \W
2988        when they appeared in character classes, but not when they appeared outside
2989        a character class. The bit map for "word" characters is now created
2990        separately from the results of isalnum() instead of just taking it from the
2991        upper, lower, and digit maps. (Plus the underscore character, of course.)
2992    
2993     5. The above bug also affected the handling of POSIX character classes such as
2994        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2995        permanent tables. Instead, the bit maps for such a class were previously
2996        created as the appropriate unions of the upper, lower, and digit bitmaps.
2997        Now they are created by subtraction from the [[:word:]] class, which has
2998        its own bitmap.
2999    
3000     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3001        It is created by subtracting the vertical space characters (\x09, \x0a,
3002        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3003        subtraction was done in the overall bitmap for a character class, meaning
3004        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3005        be recognized. This bug has been fixed.
3006    
3007     7. Patches from the folks at Google:
3008    
3009          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3010          real life, but is still worth protecting against".
3011    
3012          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3013          regular expressions".
3014    
3015          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3016          have it.
3017    
3018          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3019          "configure" and the latter not, in order to fix a problem somebody had
3020          with compiling the Arg class on HP-UX.
3021    
3022          (e) Improve the error-handling of the C++ wrapper a little bit.
3023    
3024          (f) New tests for checking recursion limiting.
3025    
3026     8. The pcre_memmove() function, which is used only if the environment does not
3027        have a standard memmove() function (and is therefore rarely compiled),
3028        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3029        returning a result (though PCRE never actually uses the result).
3030    
3031     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3032        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3033        returned instead of calling malloc() with an overflowing number that would
3034        most likely cause subsequent chaos.
3035    
3036    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3037    
3038    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3039        with this option is matched, the nmatch and pmatch options of regexec() are
3040        ignored.
3041    
3042    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3043        provided in case anyone wants to the the POSIX interface with UTF-8
3044        strings.
3045    
3046    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3047        C++ linking (needed for some HP-UX environments).
3048    
3049    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3050        (unused parameter) and in the pcre_printint() function (omitted "default"
3051        switch label when the default is to do nothing).
3052    
3053    15. Added some code to make it possible, when PCRE is compiled as a C++
3054        library, to replace subject pointers for pcre_exec() with a smart pointer
3055        class, thus making it possible to process discontinuous strings.
3056    
3057    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3058        much the same function. They were added by different people who were trying
3059        to make PCRE easy to compile on non-Unix systems. It has been suggested
3060        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3061        for compiling on Windows systems. I have therefore replaced it with
3062        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3063        defaults to "extern" for C or "extern C" for C++, which works fine on
3064        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3065        SCOPE with something explicit in config.h. In addition:
3066    
3067        (a) pcreposix.h still had just "extern" instead of either of these macros;
3068            I have replaced it with PCRE_DATA_SCOPE.
3069    
3070        (b) Functions such as _pcre_xclass(), which are internal to the library,
3071            but external in the C sense, all had PCRE_EXPORT in their definitions.
3072            This is apparently wrong for the Windows case, so I have removed it.
3073            (It makes no difference on Unix-like systems.)
3074    
3075    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3076        of recursive calls to match(). This is different to MATCH_LIMIT because
3077        that limits the total number of calls to match(), not all of which increase
3078        the depth of recursion. Limiting the recursion depth limits the amount of
3079        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3080        when PCRE is compiled, and changed at run time. A patch from Google adds
3081        this functionality to the C++ interface.
3082    
3083    18. Changes to the handling of Unicode character properties:
3084    
3085        (a) Updated the table to Unicode 4.1.0.
3086    
3087        (b) Recognize characters that are not in the table as "Cn" (undefined).
3088    
3089        (c) I revised the way the table is implemented to a much improved format
3090            which includes recognition of ranges. It now supports the ranges that
3091            are defined in UnicodeData.txt, and it also amalgamates other
3092            characters into ranges. This has reduced the number of entries in the
3093            table from around 16,000 to around 3,000, thus reducing its size
3094            considerably. I realized I did not need to use a tree structure after
3095            all - a binary chop search is just as efficient. Having reduced the
3096            number of entries, I extended their size from 6 bytes to 8 bytes to
3097            allow for more data.
3098    
3099        (d) Added support for Unicode script names via properties such as \p{Han}.
3100    
3101    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3102        matching that character.
3103    
3104    20. When matching a repeated Unicode property with a minimum greater than zero,
3105        (for example \pL{2,}), PCRE could look past the end of the subject if it
3106        reached it while seeking the minimum number of characters. This could
3107        happen only if some of the characters were more than one byte long, because
3108        there is a check for at least the minimum number of bytes.
3109    
3110    21. Refactored the implementation of \p and \P so as to be more general, to
3111        allow for more different types of property in future. This has changed the
3112        compiled form incompatibly. Anybody with saved compiled patterns that use
3113        \p or \P will have to recompile them.
3114    
3115    22. Added "Any" and "L&" to the supported property types.
3116    
3117    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3118        but give a compile time error if the value is greater than 0xff.
3119    
3120    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3121        accidentally not being installed or uninstalled.
3122    
3123    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3124        made were to insert the current release number. This seemed silly, because
3125        it made things harder for people building PCRE on systems that don't run
3126        "configure". I have turned pcre.h into a distributed file, no longer built
3127        by "configure", with the version identification directly included. There is
3128        no longer a pcre.h.in file.
3129    
3130        However, this change necessitated a change to the pcre-config script as
3131        well. It is built from pcre-config.in, and one of the substitutions was the
3132        release number. I have updated configure.ac so that ./configure now finds
3133        the release number by grepping pcre.h.
3134    
3135    26. Added the ability to run the tests under valgrind.
3136    
3137    
3138    Version 6.4 05-Sep-05
3139    ---------------------
3140    
3141     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3142        "--" to be printed when multiple files were scanned, even when none of the
3143        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3144        consider it to be a bug, and have restored the previous behaviour.
3145    
3146     2. A couple of code tidies to get rid of compiler warnings.
3147    
3148     3. The pcretest program used to cheat by referring to symbols in the library
3149        whose names begin with _pcre_. These are internal symbols that are not
3150        really supposed to be visible externally, and in some environments it is
3151        possible to suppress them. The cheating is now confined to including
3152        certain files from the library's source, which is a bit cleaner.
3153    
3154     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3155        file's purpose clearer.
3156    
3157     5. Reorganized pcre_ucp_findchar().
3158    
3159    
3160    Version 6.3 15-Aug-05
3161    ---------------------
3162    
3163     1. The file libpcre.pc.in did not have general read permission in the tarball.
3164    
3165     2. There were some problems when building without C++ support:
3166    
3167        (a) If C++ support was not built, "make install" and "make test" still
3168            tried to test it.
3169    
3170        (b) There were problems when the value of CXX was explicitly set. Some
3171            changes have been made to try to fix these, and ...
3172    
3173        (c) --disable-cpp can now be used to explicitly disable C++ support.
3174    
3175        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3176            backslash in a target when C++ was disabled. This confuses some
3177            versions of "make", apparently. Using an intermediate variable solves
3178            this. (Same for CPP_LOBJ.)
3179    
3180     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3181        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3182        necessary on certain architectures.
3183    
3184     4. Added a setting of -export-symbols-regex to the link command to remove
3185        those symbols that are exported in the C sense, but actually are local
3186        within the library, and not documented. Their names all begin with
3187        "_pcre_". This is not a perfect job, because (a) we have to except some
3188        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3189        available (and never for static libraries). I have made a note to try to
3190        find a way round (a) in the future.
3191    
3192    
3193    Version 6.2 01-Aug-05
3194    ---------------------
3195    
3196     1. There was no test for integer overflow of quantifier values. A construction
3197        such as {1111111111111111} would give undefined results. What is worse, if
3198        a minimum quantifier for a parenthesized subpattern overflowed and became
3199        negative, the calculation of the memory size went wrong. This could have
3200        led to memory overwriting.
3201    
3202     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3203    
3204     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3205        operating environments where this matters.
3206    
3207     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3208        PCRE options from within the C++ wrapper.
3209    
3210     5. Named capturing subpatterns were not being correctly counted when a pattern
3211        was compiled. This caused two problems: (a) If there were more than 100
3212        such subpatterns, the calculation of the memory needed for the whole
3213        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3214        back references of the form \12, where the number was greater than 9, were
3215        not recognized as back references, even though there were sufficient
3216        previous subpatterns.
3217    
3218     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3219        versions of gcc, e.g. 2.95.4.
3220    
3221    
3222    Version 6.1 21-Jun-05
3223    ---------------------
3224    
3225     1. There was one reference to the variable "posix" in pcretest.c that was not
3226        surrounded by "#if !defined NOPOSIX".
3227    
3228     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3229        the cross-check on the old pcre_info() function, for the benefit of the
3230        cut-down version of PCRE that is currently imported into Exim.
3231    
3232     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3233        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3234        patterns that start (?i) but I don't think that matters. The use of (?i) is
3235        just an example; this all applies to the other options as well.
3236    
3237     4. Since libtool seems to echo the compile commands it is issuing, the output
3238        from "make" can be reduced a bit by putting "@" in front of each libtool
3239        compile command.
3240    
3241     5. Patch from the folks at Google for configure.in to be a bit more thorough
3242        in checking for a suitable C++ installation before trying to compile the
3243        C++ stuff. This should fix a reported problem when a compiler was present,
3244        but no suitable headers.
3245    
3246     6. The man pages all had just "PCRE" as their title. I have changed them to
3247        be the relevant file name. I have also arranged that these names are
3248        retained in the file doc/pcre.txt, which is a concatenation in text format
3249        of all the man pages except the little individual ones for each function.
3250    
3251     7. The NON-UNIX-USE file had not been updated for the different set of source
3252        files that come with release 6. I also added a few comments about the C++
3253        wrapper.
3254    
3255    
3256    Version 6.0 07-Jun-05
3257    ---------------------
3258    
3259     1. Some minor internal re-organization to help with my DFA experiments.
3260    
3261     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3262        didn't matter for the library itself when fully configured, but did matter
3263        when compiling without UCP support, or within Exim, where the ucp files are
3264        not imported.
3265    
3266     3. Refactoring of the library code to split up the various functions into
3267        different source modules. The addition of the new DFA matching code (see
3268        below) to a single monolithic source would have made it really too
3269        unwieldy, quite apart from causing all the code to be include in a
3270        statically linked application, when only some functions are used. This is
3271        relevant even without the DFA addition now that patterns can be compiled in
3272        one application and matched in another.
3273    
3274        The downside of splitting up is that there have to be some external
3275        functions and data tables that are used internally in different modules of
3276        the library but which are not part of the API. These have all had their
3277        names changed to start with "_pcre_" so that they are unlikely to clash
3278        with other external names.
3279    
3280     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3281        a different (DFA) algorithm. Although it is slower than the original
3282        function, it does have some advantages for certain types of matching
3283        problem.
3284    
3285     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3286        including restarting after a partial match.
3287    
3288     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3289        defined when compiling for Windows was sent to me. I have put it into the
3290        code, though I have no means of testing or verifying it.
3291    
3292     7. Added the pcre_refcount() auxiliary function.
3293    
3294     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3295        match before or at the first newline in the subject string. In pcretest,
3296        the /f option on a pattern can be used to set this.
3297    
3298     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3299        would behave wrongly. This has been present in PCRE since release 4.0.
3300    
3301    10. A number of changes to the pcregrep command:
3302    
3303        (a) Refactored how -x works; insert ^(...)$ instead of setting
3304            PCRE_ANCHORED and checking the length, in preparation for adding
3305            something similar for -w.
3306    
3307        (b) Added the -w (match as a word) option.
3308    
3309        (c) Refactored the way lines are read and buffered so as to have more
3310            than one at a time available.
3311    
3312        (d) Implemented a pcregrep test script.
3313    
3314        (e) Added the -M (multiline match) option. This allows patterns to match
3315            over several lines of the subject. The buffering ensures that at least
3316            8K, or the rest of the document (whichever is the shorter) is available
3317            for matching (and similarly the previous 8K for lookbehind assertions).
3318    
3319        (f) Changed the --help output so that it now says
3320    
3321              -w, --word-regex(p)
3322    
3323            instead of two lines, one with "regex" and the other with "regexp"
3324            because that confused at least one person since the short forms are the
3325            same. (This required a bit of code, as the output is generated
3326            automatically from a table. It wasn't just a text change.)
3327    
3328        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3329            option but starts with a hyphen. Could be a pattern or a path name
3330            starting with a hyphen, for instance.
3331    
3332        (h) "-" can be given as a file name to represent stdin.
3333    
3334        (i) When file names are being printed, "(standard input)" is used for
3335            the standard input, for compatibility with GNU grep. Previously
3336            "<stdin>" was used.
3337    
3338        (j) The option --label=xxx can be used to supply a name to be used for
3339            stdin when file names are being printed. There is no short form.
3340    
3341        (k) Re-factored the options decoding logic because we are going to add
3342            two more options that take data. Such options can now be given in four
3343            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3344    
3345        (l) Added the -A, -B, and -C options for requesting that lines of context
3346            around matches be printed.
3347    
3348        (m) Added the -L option to print the names of files that do not contain
3349            any matching lines, that is, the complement of -l.
3350    
3351        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3352            continue to scan other files.
3353    
3354        (o) The -s option was incorrectly implemented. For compatibility with other
3355            greps, it now suppresses the error message for a non-existent or non-
3356            accessible file (but not the return code). There is a new option called
3357            -q that suppresses the output of matching lines, which was what -s was
3358            previously doing.
3359    
3360        (p) Added --include and --exclude options to specify files for inclusion
3361            and exclusion when recursing.
3362    
3363    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3364        Hopefully, it now does.
3365    
3366    12. Missing cast in pcre_study().
3367    
3368    13. Added an "uninstall" target to the makefile.
3369    
3370    14. Replaced "extern" in the function prototypes in Makefile.in with
3371        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3372        world, but is set differently for Windows.
3373    
3374    15. Added a second compiling function called pcre_compile2(). The only
3375        difference is that it has an extra argument, which is a pointer to an
3376        integer error code. When there is a compile-time failure, this is set
3377        non-zero, in addition to the error test pointer being set to point to an
3378        error message. The new argument may be NULL if no error number is required
3379        (but then you may as well call pcre_compile(), which is now just a
3380        wrapper). This facility is provided because some applications need a
3381        numeric error indication, but it has also enabled me to tidy up the way
3382        compile-time errors are handled in the POSIX wrapper.
3383    
3384    16. Added VPATH=.libs to the makefile; this should help when building with one
3385        prefix path and installing with another. (Or so I'm told by someone who
3386        knows more about this stuff than I do.)
3387    
3388    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3389        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3390        match everything, including newlines. This is not POSIX-compatible, but
3391        somebody wanted the feature. From pcretest it can be activated by using
3392        both the P and the s flags.
3393    
3394    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3395    
3396    19. libpcre.pc was being incorrectly installed as executable.
3397    
3398    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3399        it now also looks for '\r' so that it will work unmodified on Windows.
3400    
3401    21. Added Google's contributed C++ wrapper to the distribution.
3402    
3403    22. Added some untidy missing memory free() calls in pcretest, to keep
3404        Electric Fence happy when testing.
3405    
3406    
3407    
3408    Version 5.0 13-Sep-04
3409    ---------------------
3410    
3411     1. Internal change: literal characters are no longer packed up into items
3412        containing multiple characters in a single byte-string. Each character
3413        is now matched using a separate opcode. However, there may be more than one
3414        byte in the character in UTF-8 mode.
3415    
3416     2. The pcre_callout_block structure has two new fields: pattern_position and
3417        next_item_length. These contain the offset in the pattern to the next match
3418        item, and its length, respectively.
3419    
3420     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3421        insertion of callouts before each pattern item. Added the /C option to
3422        pcretest to make use of this.
3423    
3424     4. On the advice of a Windows user, the lines
3425    
3426          #if defined(_WIN32) || defined(WIN32)
3427          _setmode( _fileno( stdout ), 0x8000 );
3428          #endif  /* defined(_WIN32) || defined(WIN32) */
3429    
3430        have been added to the source of pcretest. This apparently does useful
3431        magic in relation to line terminators.
3432    
3433     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3434        for the benefit of those environments where the "b" makes a difference.
3435    
3436     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3437        to know about it. I have put a hack into configure.in that adds in code
3438        to set GCC=yes if CC=icc. This seems to end up at a point in the
3439        generated configure script that is early enough to affect the setting of
3440        compiler options, which is what is needed, but I have no means of testing
3441        whether it really works. (The user who reported this had patched the
3442        generated configure script, which of course I cannot do.)
3443    
3444        LATER: After change 22 below (new libtool files), the configure script
3445        seems to know about icc (and also ecc). Therefore, I have commented out
3446        this hack in configure.in.
3447    
3448     7. Added support for pkg-config (2 patches were sent in).
3449    
3450     8. Negated POSIX character classes that used a combination of internal tables
3451        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3452        [[:^ascii]]. Typically, they would match almost any characters. The other
3453        POSIX classes were not broken in this way.
3454    
3455     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3456        to find the match, as PCRE was deluded into thinking that the match had to
3457        start at the start point or following a newline. The same bug applied to
3458        patterns with negative forward assertions or any backward assertions
3459        preceding ".*" at the start, unless the pattern required a fixed first
3460        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3461    
3462    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3463        starting at the last subject character, bytes beyond the end of the subject
3464        string were read.
3465    
3466    11. Renamed the variable "class" as "classbits" to make life easier for C++
3467        users. (Previously there was a macro definition, but it apparently wasn't
3468        enough.)
3469    
3470    12. Added the new field "tables" to the extra data so that tables can be passed
3471        in at exec time, or the internal tables can be re-selected. This allows
3472        a compiled regex to be saved and re-used at a later time by a different
3473        program that might have everything at different addresses.
3474    
3475    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3476        -R library as well as a -L library.
3477    
3478    14. The debugging options of pcretest (-d on the command line or D on a
3479        pattern) showed incorrect output for anything following an extended class
3480        that contained multibyte characters and which was followed by a quantifier.
3481    
3482    15. Added optional support for general category Unicode character properties
3483        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3484        support. It adds about 90K to the size of the library. The meanings of the
3485        inbuilt class escapes such as \d and \s have NOT been changed.
3486    
3487    16. Updated pcredemo.c to include calls to free() to release the memory for the
3488        compiled pattern.
3489    
3490    17. The generated file chartables.c was being created in the source directory
3491        instead of in the building directory. This caused the build to fail if the
3492        source directory was different from the building directory, and was
3493        read-only.
3494    
3495    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3496        file. No doubt somebody will tell me if they don't make sense... Also added
3497        Dan Mooney's comments about building on OpenVMS.
3498    
3499    19. Added support for partial matching via the PCRE_PARTIAL option for
3500        pcre_exec() and the \P data escape in pcretest.
3501    
3502    20. Extended pcretest with 3 new pattern features:
3503    
3504        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3505              write the compiled pattern to the file whose name is "rest-of-line".
3506              This is a straight binary dump of the data, with the saved pointer to
3507              the character tables forced to be NULL. The study data, if any, is
3508              written too. After writing, pcretest reads a new pattern.
3509    
3510        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3511              compiled pattern from the given file. There must not be any
3512              occurrences of "<" in the file name (pretty unlikely); if there are,
3513              pcretest will instead treat the initial "<" as a pattern delimiter.
3514              After reading in the pattern, pcretest goes on to read data lines as
3515              usual.
3516    
3517        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3518              and 16-bit fields in a compiled pattern, to simulate a pattern that
3519              was compiled on a host of opposite endianness.
3520    
3521    21. The pcre-exec() function can now cope with patterns that were compiled on
3522        hosts of opposite endianness, with this restriction:
3523    
3524          As for any compiled expression that is saved and used later, the tables
3525          pointer field cannot be preserved; the extra_data field in the arguments
3526          to pcre_exec() should be used to pass in a tables address if a value
3527          other than the default internal tables were used at compile time.
3528    
3529    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3530        now diagnosed as an error. Previously, most of the time, a negative number
3531        would have been treated as zero, but if in addition "ovector" was passed as
3532        NULL, a crash could occur.
3533    
3534    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3535        new versions from the libtool 1.5 distribution (the last one is a copy of
3536        a file called libtool.m4). This seems to have fixed the need to patch
3537        "configure" to support Darwin 1.3 (which I used to do). However, I still
3538        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3539        workstation).
3540    
3541    24. Changed the PCRE licence to be the more standard "BSD" licence.
3542    
3543    
3544    Version 4.5 01-Dec-03
3545    ---------------------
3546    
3547     1. There has been some re-arrangement of the code for the match() function so
3548        that it can be compiled in a version that does not call itself recursively.
3549        Instead, it keeps those local variables that need separate instances for
3550        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3551        needs to "recurse". Keeping track of where control must go is done by means
3552        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3553        hide most of the details from the main code, and operates only if
3554        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3555        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3556        operating.
3557    
3558        To make it easier for callers to provide specially tailored get/free
3559        functions for this usage, two new functions, pcre_stack_malloc, and
3560        pcre_stack_free, are used. They are always called in strict stacking order,
3561        and the size of block requested is always the same.
3562    
3563        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3564        PCRE has been compiled to use the stack or the heap for recursion. The
3565        -C option of pcretest uses this to show which version is compiled.
3566    
3567        A new data escape \S, is added to pcretest; it causes the amounts of store
3568        obtained and freed by both kinds of malloc/free at match time to be added
3569        to the output.
3570    
3571     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3572        what's available on my current Linux desktop machine.
3573    
3574     3. When matching a UTF-8 string, the test for a valid string at the start has
3575        been extended. If start_offset is not zero, PCRE now checks that it points
3576        to a byte that is the start of a UTF-8 character. If not, it returns
3577        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3578        this is necessary because there may be backward assertions in the pattern.
3579        When matching the same subject several times, it may save resources to use
3580        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3581    
3582     4. The code for checking the validity of UTF-8 strings has been tightened so
3583        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3584        containing "overlong sequences".
3585    
3586     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3587        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3588        should have been "&", but it just so happened that all the cases this let
3589        through by mistake were picked up later in the function.
3590    
3591     6. I had used a variable called "isblank" - this is a C99 function, causing
3592        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3593    
3594     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3595        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3596        so the version is shown; (c) stop "make test" echoing "./RunTest".
3597    
3598     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3599    
3600     9. The prototype for memmove() for systems that don't have it was using
3601        size_t, but the inclusion of the header that defines size_t was later. I've
3602        moved the #includes for the C headers earlier to avoid this.
3603    
3604    10. Added some adjustments to the code to make it easier to compiler on certain
3605        special systems:
3606    
3607          (a) Some "const" qualifiers were missing.
3608          (b) Added the macro EXPORT before all exported functions; by default this
3609              is defined to be empty.
3610          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3611              that it reads its output file name as an argument instead of writing
3612              to the standard output and assuming this can be redirected.
3613    
3614    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3615        class containing characters with values greater than 255, PCRE compilation
3616        went into a loop.
3617    
3618    12. A recursive reference to a subpattern that was within another subpattern
3619        that had a minimum quantifier of zero caused PCRE to crash. For example,
3620        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3621        recursion. If the recursively-called subpattern itself had a zero repeat,
3622        that was OK.
3623    
3624    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3625        buffer into which it was copied (for escape processing) was still set at
3626        1024, so long lines caused crashes.
3627    
3628    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3629        "internal error: code overflow...". This applied to any character class
3630        that was followed by a possessive quantifier.
3631    
3632    15. Modified the Makefile to add libpcre.la as a prerequisite for
3633        libpcreposix.la because I was told this is needed for a parallel build to
3634        work.
3635    
3636    16. If a pattern that contained .* following optional items at the start was
3637        studied, the wrong optimizing data was generated, leading to matching
3638        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3639        matching string must start with a or b or c. The correct conclusion for
3640        this pattern is that a match can start with any character.
3641    
3642    
3643    Version 4.4 13-Aug-03
3644    ---------------------
3645    
3646     1. In UTF-8 mode, a character class containing characters with values between
3647        127 and 255 was not handled correctly if the compiled pattern was studied.
3648        In fixing this, I have also improved the studying algorithm for such
3649        classes (slightly).
3650    
3651     2. Three internal functions had redundant arguments passed to them. Removal
3652        might give a very teeny performance improvement.
3653    
3654     3. Documentation bug: the value of the capture_top field in a callout is *one
3655        more than* the number of the hightest numbered captured substring.
3656    
3657     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3658        in incorrectly linking with a previously installed version. They now link
3659        explicitly with libpcre.la.
3660    
3661     5. configure.in no longer needs to recognize Cygwin specially.
3662    
3663     6. A problem in pcre.in for Windows platforms is fixed.
3664    
3665     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3666        pcretest, it used to include the size of the study block as part of its
3667        output. Unfortunately, the structure contains a field that has a different
3668        size on different hardware architectures. This meant that the tests that
3669        showed this size failed. As the block is currently always of a fixed size,
3670        this information isn't actually particularly useful in pcretest output, so
3671        I have just removed it.
3672    
3673     8. Three pre-processor statements accidentally did not start in column 1.
3674        Sadly, there are *still* compilers around that complain, even though
3675        standard C has not required this for well over a decade. Sigh.
3676    
3677     9. In pcretest, the code for checking callouts passed small integers in the
3678        callout_data field, which is a void * field. However, some picky compilers
3679        complained about the casts involved for this on 64-bit systems. Now
3680        pcretest passes the address of the small integer instead, which should get
3681        rid of the warnings.
3682    
3683    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3684        both compile and run time, and gives an error if an invalid UTF-8 sequence
3685        is found. There is a option for disabling this check in cases where the
3686        string is known to be correct and/or the maximum performance is wanted.
3687    
3688    11. In response to a bug report, I changed one line in Makefile.in from
3689    
3690            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3691        to
3692            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3693    
3694        to look similar to other lines, but I have no way of telling whether this
3695        is the right thing to do, as I do not use Windows. No doubt I'll get told
3696        if it's wrong...
3697    
3698    
3699    Version 4.3 21-May-03
3700    ---------------------
3701    
3702    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3703       Makefile.
3704    
3705    2. Some refactoring to improve the quality of the code:
3706    
3707       (i)   The utf8_table... variables are now declared "const".
3708    
3709       (ii)  The code for \cx, which used the "case flipping" table to upper case
3710             lower case letters, now just substracts 32. This is ASCII-specific,
3711             but the whole concept of \cx is ASCII-specific, so it seems
3712             reasonable.
3713    
3714       (iii) PCRE was using its character types table to recognize decimal and
3715             hexadecimal digits in the pattern. This is silly, because it handles
3716             only 0-9, a-f, and A-F, but the character types table is locale-
3717             specific, which means strange things might happen. A private
3718             table is now used for this - though it costs 256 bytes, a table is
3719             much faster than multiple explicit tests. Of course, the standard
3720             character types table is still used for matching digits in subject
3721             strings against \d.
3722    
3723       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3724             ending in _t are). So I've renamed it as ESC_tee.
3725    
3726    3. The first argument for regexec() in the POSIX wrapper should have been
3727       defined as "const".
3728    
3729    4. Changed pcretest to use malloc() for its buffers so that they can be
3730       Electric Fenced for debugging.
3731    
3732    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3733       to read one or more bytes before the start of the subject string. Often this
3734       had no effect on PCRE's behaviour, but in some circumstances it could
3735       provoke a segmentation fault.
3736    
3737    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3738       to try to read one or more bytes before the start of the subject string.
3739    
3740    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3741       UTF-8 support could misbehave in various ways if the subject string
3742       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3743       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3744       back over UTF-8 characters.)
3745    
3746    
3747    Version 4.2 14-Apr-03
3748    ---------------------
3749    
3750    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3751    
3752    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3753         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3754         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3755         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3756         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3757           and BUILD_EXEEXT
3758         Note: automatic setting of the BUILD variables is not yet working
3759         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3760           compile-time but not at link-time
3761         [LINK]: use for linking executables only
3762         make different versions for Windows and non-Windows
3763         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3764           libraries
3765         [LINK_FOR_BUILD]: new variable
3766         [OBJEXT]: use throughout
3767         [EXEEXT]: use throughout
3768         <winshared>: new target
3769         <wininstall>: new target
3770         <dftables.o>: use native compiler
3771         <dftables>: use native linker
3772         <install>: handle Windows platform correctly
3773         <clean>: ditto
3774         <check>: ditto
3775         copy DLL to top builddir before testing
3776    
3777       As part of these changes, -no-undefined was removed again. This was reported
3778       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3779       in any case.
3780    
3781    3. Some tidies to get rid of compiler warnings:
3782    
3783       . In the match_data structure, match_limit was an unsigned long int, whereas
3784         match_call_count was an int. I've made them both unsigned long ints.
3785    
3786       . In pcretest the fact that a const uschar * doesn't automatically cast to
3787         a void * provoked a warning.
3788