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

Diff of /code/trunk/ChangeLog

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

revision 98 by ph10, Mon Mar 5 15:55:02 2007 UTC revision 1250 by zherczeg, Wed Feb 20 09:59:29 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 05-Mar-07  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    9.  Optimizing fast_forward_start_bits in JIT.
38    
39    10. Adding experimental support for callouts in JIT, and fixing some
40        issues revealed during this work. Namely:
41    
42        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43    
44        (b) Minimum length was not checked before the matching is started.
45    
46    11. The value of capture_last that is passed to callouts was incorrect in some
47        cases when there was a capture on one path that was subsequently abandoned
48        after a backtrack. Also, the capture_last value is now reset after a
49        recursion, since all captures are also reset in this case.
50    
51    12. The interpreter no longer returns the "too many substrings" error in the
52        case when an overflowing capture is in a branch that is subsequently
53        abandoned after a backtrack.
54    
55    13. In the pathological case when an offset vector of size 2 is used, pcretest
56        now prints out the matched string after a yield of 0 or 1.
57    
58    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59        Only supported by the JIT compiler at the moment.
60    
61    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62    
63    
64    Version 8.32 30-November-2012
65    -----------------------------
66    
67    1.  Improved JIT compiler optimizations for first character search and single
68        character iterators.
69    
70    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
71        Patch by Daniel Richard G.
72    
73    3.  Single character iterator optimizations in the JIT compiler.
74    
75    4.  Improved JIT compiler optimizations for character ranges.
76    
77    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
78        Reported by Giuseppe D'Angelo.
79    
80    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
81        of a line, was being set incorrectly in cases where .* appeared inside
82        atomic brackets at the start of a pattern, or where there was a subsequent
83        *PRUNE or *SKIP.
84    
85    7.  Improved instruction cache flush for POWER/PowerPC.
86        Patch by Daniel Richard G.
87    
88    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
89        grep:
90    
91        (a) There is now no limit to the number of patterns to be matched.
92    
93        (b) An error is given if a pattern is too long.
94    
95        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
96            are now supported.
97    
98        (d) --exclude-from and --include-from (multiple use) have been added.
99    
100        (e) Exclusions and inclusions now apply to all files and directories, not
101            just to those obtained from scanning a directory recursively.
102    
103        (f) Multiple uses of -f and --file-list are now supported.
104    
105        (g) In a Windows environment, the default for -d has been changed from
106            "read" (the GNU grep default) to "skip", because otherwise the presence
107            of a directory in the file list provokes an error.
108    
109        (h) The documentation has been revised and clarified in places.
110    
111    9.  Improve the matching speed of capturing brackets.
112    
113    10. Changed the meaning of \X so that it now matches a Unicode extended
114        grapheme cluster.
115    
116    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
117        out POSIX threads when JIT support is configured.
118    
119    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
120    
121    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
122        structure could go wrong in environments where size_t is not the same size
123        as int.
124    
125    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
126        set.
127    
128    15. The EBCDIC support had decayed; later updates to the code had included
129        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
130        general tidy up of EBCDIC-related issues, and the documentation was also
131        not quite right. There is now a test that can be run on ASCII systems to
132        check some of the EBCDIC-related things (but is it not a full test).
133    
134    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
135        in a small tidy to the code.
136    
137    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
138    
139    18. If the --only-matching (-o) option in pcregrep is specified multiple
140        times, each one causes appropriate output. For example, -o1 -o2 outputs the
141        substrings matched by the 1st and 2nd capturing parentheses. A separating
142        string can be specified by --om-separator (default empty).
143    
144    19. Improving the first n character searches.
145    
146    20. Turn case lists for horizontal and vertical white space into macros so that
147        they are defined only once.
148    
149    21. This set of changes together give more compatible Unicode case-folding
150        behaviour for characters that have more than one other case when UCP
151        support is available.
152    
153        (a) The Unicode property table now has offsets into a new table of sets of
154            three or more characters that are case-equivalent. The MultiStage2.py
155            script that generates these tables (the pcre_ucd.c file) now scans
156            CaseFolding.txt instead of UnicodeData.txt for character case
157            information.
158    
159        (b) The code for adding characters or ranges of characters to a character
160            class has been abstracted into a generalized function that also handles
161            case-independence. In UTF-mode with UCP support, this uses the new data
162            to handle characters with more than one other case.
163    
164        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
165            whose other case is greater than 256 are now correctly matched
166            caselessly. Previously, the high codepoint matched the low one, but not
167            vice versa.
168    
169        (d) The processing of \h, \H, \v, and \ in character classes now makes use
170            of the new class addition function, using character lists defined as
171            macros alongside the case definitions of 20 above.
172    
173        (e) Caseless back references now work with characters that have more than
174            one other case.
175    
176        (f) General caseless matching of characters with more than one other case
177            is supported.
178    
179    22. Unicode character properties were updated from Unicode 6.2.0
180    
181    23. Improved CMake support under Windows. Patch by Daniel Richard G.
182    
183    24. Add support for 32-bit character strings, and UTF-32
184    
185    25. Major JIT compiler update (code refactoring and bugfixing).
186        Experimental Sparc 32 support is added.
187    
188    26. Applied a modified version of Daniel Richard G's patch to create
189        pcre.h.generic and config.h.generic by "make" instead of in the
190        PrepareRelease script.
191    
192    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
193        pcre_compile.c when checking for a zero character.
194    
195    28. Introducing a native interface for JIT. Through this interface, the compiled
196        machine code can be directly executed. The purpose of this interface is to
197        provide fast pattern matching, so several sanity checks are not performed.
198        However, feature tests are still performed. The new interface provides
199        1.4x speedup compared to the old one.
200    
201    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
202        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
203        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
204    
205    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
206        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
207        the "old" RFC 2279). Instead, it ended up passing a negative length to
208        pcre_exec().
209    
210    31. Add support for GCC's visibility feature to hide internal functions.
211    
212    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
213        "unknown -C option" after outputting 0 or 1.
214    
215    33. There is now support for generating a code coverage report for the test
216        suite in environments where gcc is the compiler and lcov is installed. This
217        is mainly for the benefit of the developers.
218    
219    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
220        unaddressable using valgrind annotations, allowing valgrind to detect
221        invalid memory accesses. This is mainly for the benefit of the developers.
222    
223    25. (*UTF) can now be used to start a pattern in any of the three libraries.
224    
225    26. Give configure error if --enable-cpp but no C++ compiler found.
226    
227    
228    Version 8.31 06-July-2012
229    -------------------------
230    
231    1.  Fixing a wrong JIT test case and some compiler warnings.
232    
233    2.  Removed a bashism from the RunTest script.
234    
235    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
236        to unsigned type, result still unsigned" that was given by an MS compiler
237        on encountering the code "-sizeof(xxx)".
238    
239    4.  Partial matching support is added to the JIT compiler.
240    
241    5.  Fixed several bugs concerned with partial matching of items that consist
242        of more than one character:
243    
244        (a) /^(..)\1/ did not partially match "aba" because checking references was
245            done on an "all or nothing" basis. This also applied to repeated
246            references.
247    
248        (b) \R did not give a hard partial match if \r was found at the end of the
249            subject.
250    
251        (c) \X did not give a hard partial match after matching one or more
252            characters at the end of the subject.
253    
254        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
255            a partial match for the string "\r".
256    
257        (e) When newline was set to CRLF, the metacharacter "." did not recognize
258            a partial match for a CR character at the end of the subject string.
259    
260    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
261        running pcretest, the text "(JIT)" added to the output whenever JIT is
262        actually used to run the match.
263    
264    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
265        or /S+[+] with a digit between 1 and 7.
266    
267    8.  OP_NOT now supports any UTF character not just single-byte ones.
268    
269    9.  (*MARK) control verb is now supported by the JIT compiler.
270    
271    10. The command "./RunTest list" lists the available tests without actually
272        running any of them. (Because I keep forgetting what they all are.)
273    
274    11. Add PCRE_INFO_MAXLOOKBEHIND.
275    
276    12. Applied a (slightly modified) user-supplied patch that improves performance
277        when the heap is used for recursion (compiled with --disable-stack-for-
278        recursion). Instead of malloc and free for each heap frame each time a
279        logical recursion happens, frames are retained on a chain and re-used where
280        possible. This sometimes gives as much as 30% improvement.
281    
282    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
283        call.
284    
285    14. As documented, (*COMMIT) is now confined to within a positive assertion.
286    
287    15. It is now possible to link pcretest with libedit as an alternative to
288        libreadline.
289    
290    16. (*COMMIT) control verb is now supported by the JIT compiler.
291    
292    17. The Unicode data tables have been updated to Unicode 6.1.0.
293    
294    18. Added --file-list option to pcregrep.
295    
296    19. Added binary file support to pcregrep, including the -a, --binary-files,
297        -I, and --text options.
298    
299    20. The madvise function is renamed for posix_madvise for QNX compatibility
300        reasons. Fixed by Giuseppe D'Angelo.
301    
302    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
303    
304    22. Changed the option for creating symbolic links for 16-bit man pages from
305        -s to -sf so that re-installing does not cause issues.
306    
307    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
308    
309    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
310        matches in certain environments (the workspace was not being correctly
311        retained). Also added to pcre_dfa_exec() a simple plausibility check on
312        some of the workspace data at the beginning of a restart.
313    
314    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
315        was not doing so when it should - probably a typo introduced by SVN 528
316        (change 8.10/14).
317    
318    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
319        \w+ when the character tables indicated that \x{c4} was a word character.
320        There were several related cases, all because the tests for doing a table
321        lookup were testing for characters less than 127 instead of 255.
322    
323    27. If a pattern contains capturing parentheses that are not used in a match,
324        their slots in the ovector are set to -1. For those that are higher than
325        any matched groups, this happens at the end of processing. In the case when
326        there were back references that the ovector was too small to contain
327        (causing temporary malloc'd memory to be used during matching), and the
328        highest capturing number was not used, memory off the end of the ovector
329        was incorrectly being set to -1. (It was using the size of the temporary
330        memory instead of the true size.)
331    
332    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
333        ovector size, it uses memory at the end of the block that it has got.
334    
335    29. Check for an overlong MARK name and give an error at compile time. The
336        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
337    
338    30. JIT compiler update.
339    
340    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
341        Rill for the patch.
342    
343    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
344    
345    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
346    
347    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
348        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
349        was enabled.
350    
351    35. Improve JIT code generation for greedy plus quantifier.
352    
353    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
354        1 to "aa" instead of to an empty string. The bug affected repeated groups
355        that could potentially match an empty string.
356    
357    37. Optimizing single character iterators in JIT.
358    
359    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
360        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
361        codepoints that are too big for the mode are faulted, and in a UTF mode,
362        disallowed codepoints are also faulted.
363    
364    39. If PCRE was compiled with UTF support, in three places in the DFA
365        matcher there was code that should only have been obeyed in UTF mode, but
366        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
367        processing when bytes with values greater than 127 were present. In 16-bit
368        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
369        both cases the values are those that cannot be the first data item in a UTF
370        character. The three items that might have provoked this were recursions,
371        possessively repeated groups, and atomic groups.
372    
373    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
374        and pcregrep, because some OS require shared objects to be explicitly
375        passed to ld, causing the link step to fail if they are not.
376    
377    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
378        mode, patterns that started with \h* or \R* might be incorrectly matched.
379    
380    
381    Version 8.30 04-February-2012
382    -----------------------------
383    
384    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
385        name is defined in ctype.h.
386    
387    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
388        only in quite long subpatterns.
389    
390    3.  Removed the function pcre_info(), which has been obsolete and deprecated
391        since it was replaced by pcre_fullinfo() in February 2000.
392    
393    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
394        match a (*MARK), and the match failed at the start of the subject, a
395        reference to memory before the start of the subject could occur. This bug
396        was introduced by fix 17 of release 8.21.
397    
398    5.  A reference to an unset group with zero minimum repetition was giving
399        totally wrong answers (in non-JavaScript-compatibility mode). For example,
400        /(another)?(\1?)test/ matched against "hello world test". This bug was
401        introduced in release 8.13.
402    
403    6.  Add support for 16-bit character strings (a large amount of work involving
404        many changes and refactorings).
405    
406    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
407        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
408        from a file.
409    
410    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
411        rounding is not applied in this particular case).
412    
413    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
414        if they appear, or are escaped, in patterns.
415    
416    10. Get rid of a number of -Wunused-but-set-variable warnings.
417    
418    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
419        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
420        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
421        also returns the mark "x". This bug applied to capturing parentheses,
422        non-capturing parentheses, and atomic parentheses. It also applied to some
423        assertions.
424    
425    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
426        information out of configure.ac instead of relying on pcre.h.generic, which
427        is not stored in the repository.
428    
429    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
430        -lreadline.
431    
432    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
433    
434    15. Applied Graycode's patch to put the top-level frame on the stack rather
435        than the heap when not using the stack for recursion. This gives a
436        performance improvement in many cases when recursion is not deep.
437    
438    16. Experimental code added to "pcretest -C" to output the stack frame size.
439    
440    
441    Version 8.21 12-Dec-2011
442    ------------------------
443    
444    1.  Updating the JIT compiler.
445    
446    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
447        are added as well.
448    
449    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
450        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
451        calling _pcre_jit_exec. Some extra comments are added.
452    
453    4.  (*MARK) settings inside atomic groups that do not contain any capturing
454        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
455        was introduced by change 18 for 8.20.
456    
457    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
458        ECMA-262 standard.
459    
460    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
461        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
462        This bug was probably introduced by change 9 of 8.13.
463    
464    7.  While fixing 6 above, I noticed that a number of other items were being
465        incorrectly rejected as "not fixed length". This arose partly because newer
466        opcodes had not been added to the fixed-length checking code. I have (a)
467        corrected the bug and added tests for these items, and (b) arranged for an
468        error to occur if an unknown opcode is encountered while checking for fixed
469        length instead of just assuming "not fixed length". The items that were
470        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
471        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
472        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
473    
474    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
475        being incorrectly compiled and would have given unpredicatble results.
476    
477    9.  A possessively repeated subpattern with minimum repeat count greater than
478        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
479        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
480        the first (A) could occur when it should not.
481    
482    10. Add a cast and remove a redundant test from the code.
483    
484    11. JIT should use pcre_malloc/pcre_free for allocation.
485    
486    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
487        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
488        is anything other than /usr, -L is still shown).
489    
490    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
491    
492    14. Perl does not support \N without a following name in a [] class; PCRE now
493        also gives an error.
494    
495    15. If a forward reference was repeated with an upper limit of around 2000,
496        it caused the error "internal error: overran compiling workspace". The
497        maximum number of forward references (including repeats) was limited by the
498        internal workspace, and dependent on the LINK_SIZE. The code has been
499        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
500        the default depends on LINK_SIZE. There is a new upper limit (for safety)
501        of around 200,000 forward references. While doing this, I also speeded up
502        the filling in of repeated forward references.
503    
504    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
505        incorrectly expecting the subject to contain another "a" after the start.
506    
507    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
508        in the match, the SKIP should be ignored. This was not happening; instead
509        the SKIP was being treated as NOMATCH. For patterns such as
510        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
511        tested.
512    
513    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
514        now much more compatible with Perl, in particular in cases where the result
515        is a non-match for a non-anchored pattern. For example, if
516        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
517        "m", where previously it did not return a name. A side effect of this
518        change is that for partial matches, the last encountered mark name is
519        returned, as for non matches. A number of tests that were previously not
520        Perl-compatible have been moved into the Perl-compatible test files. The
521        refactoring has had the pleasing side effect of removing one argument from
522        the match() function, thus reducing its stack requirements.
523    
524    19. If the /S+ option was used in pcretest to study a pattern using JIT,
525        subsequent uses of /S (without +) incorrectly behaved like /S+.
526    
527    21. Retrieve executable code size support for the JIT compiler and fixing
528        some warnings.
529    
530    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
531        not work when the shorter character appeared right at the end of the
532        subject string.
533    
534    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
535        systems.
536    
537    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
538        output it when the /M option is used in pcretest.
539    
540    25. The CheckMan script was not being included in the distribution. Also, added
541        an explicit "perl" to run Perl scripts from the PrepareRelease script
542        because this is reportedly needed in Windows.
543    
544    26. If study data was being save in a file and studying had not found a set of
545        "starts with" bytes for the pattern, the data written to the file (though
546        never used) was taken from uninitialized memory and so caused valgrind to
547        complain.
548    
549    27. Updated RunTest.bat as provided by Sheri Pierce.
550    
551    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
552    
553    29. Computation of memory usage for the table of capturing group names was
554        giving an unnecessarily large value.
555    
556    
557    Version 8.20 21-Oct-2011
558    ------------------------
559    
560    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
561        a POSIX class. After further experiments with Perl, which convinced me that
562        Perl has bugs and confusions, a closing square bracket is no longer allowed
563        in a POSIX name. This bug also affected patterns with classes that started
564        with full stops.
565    
566    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
567        captured substring, but while checking the failing first alternative,
568        substring 1 is temporarily captured. If the output vector supplied to
569        pcre_exec() was not big enough for this capture, the yield of the function
570        was still zero ("insufficient space for captured substrings"). This cannot
571        be totally fixed without adding another stack variable, which seems a lot
572        of expense for a edge case. However, I have improved the situation in cases
573        such as /(a)(b)x|abc/ matched against "abc", where the return code
574        indicates that fewer than the maximum number of slots in the ovector have
575        been set.
576    
577    3.  Related to (2) above: when there are more back references in a pattern than
578        slots in the output vector, pcre_exec() uses temporary memory during
579        matching, and copies in the captures as far as possible afterwards. It was
580        using the entire output vector, but this conflicts with the specification
581        that only 2/3 is used for passing back captured substrings. Now it uses
582        only the first 2/3, for compatibility. This is, of course, another edge
583        case.
584    
585    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
586        main code base, and can be used by building with --enable-jit. When this is
587        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
588        runtime --no-jit option is given.
589    
590    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
591        ovector, the return from the function was zero, implying that there were
592        other matches that did not fit. The correct "exactly full" value is now
593        returned.
594    
595    6.  If a subpattern that was called recursively or as a subroutine contained
596        (*PRUNE) or any other control that caused it to give a non-standard return,
597        invalid errors such as "Error -26 (nested recursion at the same subject
598        position)" or even infinite loops could occur.
599    
600    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
601        computing the minimum length on reaching *ACCEPT, and so ended up with the
602        wrong value of 1 rather than 0. Further investigation indicates that
603        computing a minimum subject length in the presence of *ACCEPT is difficult
604        (think back references, subroutine calls), and so I have changed the code
605        so that no minimum is registered for a pattern that contains *ACCEPT.
606    
607    8.  If (*THEN) was present in the first (true) branch of a conditional group,
608        it was not handled as intended. [But see 16 below.]
609    
610    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
611        Sheri Pierce.
612    
613    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
614        the first byte in a match must be "a".
615    
616    11. Change 17 for 8.13 increased the recursion depth for patterns like
617        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
618        pattern contains any instances of (*THEN). If it does not, the old
619        optimizations are restored. It would be nice to do this on a per-group
620        basis, but at the moment that is not feasible.
621    
622    12. In some environments, the output of pcretest -C is CRLF terminated. This
623        broke RunTest's code that checks for the link size. A single white space
624        character after the value is now allowed for.
625    
626    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
627        For "fr", it uses the Windows-specific input and output files.
628    
629    14. If (*THEN) appeared in a group that was called recursively or as a
630        subroutine, it did not work as intended. [But see next item.]
631    
632    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
633        pattern fragments (but not containing any | characters). If A and B are
634        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
635        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
636        D. In other words, Perl considers parentheses that do not contain any |
637        characters to be part of a surrounding alternative, whereas PCRE was
638        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
639        differently. PCRE now behaves in the same way as Perl, except in the case
640        of subroutine/recursion calls such as (?1) which have in any case always
641        been different (but PCRE had them first :-).
642    
643    16. Related to 15 above: Perl does not treat the | in a conditional group as
644        creating alternatives. Such a group is treated in the same way as an
645        ordinary group without any | characters when processing (*THEN). PCRE has
646        been changed to match Perl's behaviour.
647    
648    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
649        RunGrepTest script failed.
650    
651    18. Change 22 for version 13 caused atomic groups to use more stack. This is
652        inevitable for groups that contain captures, but it can lead to a lot of
653        stack use in large patterns. The old behaviour has been restored for atomic
654        groups that do not contain any capturing parentheses.
655    
656    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
657        suppress the check for a minimum subject length at run time. (If it was
658        given to pcre_exec() or pcre_dfa_exec() it did work.)
659    
660    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
661        fail to work when decoding hex characters in data strings in EBCDIC
662        environments.
663    
664    21. It appears that in at least one Mac OS environment, the isxdigit() function
665        is implemented as a macro that evaluates to its argument more than once,
666        contravening the C 90 Standard (I haven't checked a later standard). There
667        was an instance in pcretest which caused it to go wrong when processing
668        \x{...} escapes in subject strings. The has been rewritten to avoid using
669        things like p++ in the argument of isxdigit().
670    
671    
672    Version 8.13 16-Aug-2011
673    ------------------------
674    
675    1.  The Unicode data tables have been updated to Unicode 6.0.0.
676    
677    2.  Two minor typos in pcre_internal.h have been fixed.
678    
679    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
680        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
681        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
682    
683    4.  There were a number of related bugs in the code for matching backrefences
684        caselessly in UTF-8 mode when codes for the characters concerned were
685        different numbers of bytes. For example, U+023A and U+2C65 are an upper
686        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
687        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
688        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
689        2-byte code at the end of the subject (it thought there wasn't enough data
690        left).
691    
692    5.  Comprehensive information about what went wrong is now returned by
693        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
694        as the output vector has at least 2 elements. The offset of the start of
695        the failing character and a reason code are placed in the vector.
696    
697    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
698        now returned is for the first byte of the failing character, instead of the
699        last byte inspected. This is an incompatible change, but I hope it is small
700        enough not to be a problem. It makes the returned offset consistent with
701        pcre_exec() and pcre_dfa_exec().
702    
703    7.  pcretest now gives a text phrase as well as the error number when
704        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
705        failure, the offset and reason code are output.
706    
707    8.  When \R was used with a maximizing quantifier it failed to skip backwards
708        over a \r\n pair if the subsequent match failed. Instead, it just skipped
709        back over a single character (\n). This seems wrong (because it treated the
710        two characters as a single entity when going forwards), conflicts with the
711        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
712        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
713        has been changed.
714    
715    9.  Some internal refactoring has changed the processing so that the handling
716        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
717        time (the PCRE_DOTALL option was changed this way some time ago: version
718        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
719        which was always a bit of a fudge. It also means that there is one less
720        argument for the match() function, which reduces its stack requirements
721        slightly. This change also fixes an incompatibility with Perl: the pattern
722        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
723    
724    10. More internal refactoring has drastically reduced the number of recursive
725        calls to match() for possessively repeated groups such as (abc)++ when
726        using pcre_exec().
727    
728    11. While implementing 10, a number of bugs in the handling of groups were
729        discovered and fixed:
730    
731        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
732        (a|)*(?1) gave a compile-time internal error.
733        ((a|)+)+  did not notice that the outer group could match an empty string.
734        (^a|^)+   was not marked as anchored.
735        (.*a|.*)+ was not marked as matching at start or after a newline.
736    
737    12. Yet more internal refactoring has removed another argument from the match()
738        function. Special calls to this function are now indicated by setting a
739        value in a variable in the "match data" data block.
740    
741    13. Be more explicit in pcre_study() instead of relying on "default" for
742        opcodes that mean there is no starting character; this means that when new
743        ones are added and accidentally left out of pcre_study(), testing should
744        pick them up.
745    
746    14. The -s option of pcretest has been documented for ages as being an old
747        synonym of -m (show memory usage). I have changed it to mean "force study
748        for every regex", that is, assume /S for every regex. This is similar to -i
749        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
750        using it. It makes it easier to run collections of tests with and without
751        study enabled, and thereby test pcre_study() more easily. All the standard
752        tests are now run with and without -s (but some patterns can be marked as
753        "never study" - see 20 below).
754    
755    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
756        restoration of the capturing data to the outer values was not happening
757        correctly.
758    
759    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
760        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
761        pattern had matched an empty string, and so incorrectly returned a no
762        match.
763    
764    17. There was optimizing code for the last branch of non-capturing parentheses,
765        and also for the obeyed branch of a conditional subexpression, which used
766        tail recursion to cut down on stack usage. Unfortunately, now that there is
767        the possibility of (*THEN) occurring in these branches, tail recursion is
768        no longer possible because the return has to be checked for (*THEN). These
769        two optimizations have therefore been removed. [But see 8.20/11 above.]
770    
771    18. If a pattern containing \R was studied, it was assumed that \R always
772        matched two bytes, thus causing the minimum subject length to be
773        incorrectly computed because \R can also match just one byte.
774    
775    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
776        was incorrectly computed.
777    
778    20. If /S is present twice on a test pattern in pcretest input, it now
779        *disables* studying, thereby overriding the use of -s on the command line
780        (see 14 above). This is necessary for one or two tests to keep the output
781        identical in both cases.
782    
783    21. When (*ACCEPT) was used in an assertion that matched an empty string and
784        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
785    
786    22. When an atomic group that contained a capturing parenthesis was
787        successfully matched, but the branch in which it appeared failed, the
788        capturing was not being forgotten if a higher numbered group was later
789        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
790        group 1 to "a", when in fact it should be unset. This applied to multi-
791        branched capturing and non-capturing groups, repeated or not, and also to
792        positive assertions (capturing in negative assertions does not happen
793        in PCRE) and also to nested atomic groups.
794    
795    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
796        subject after a captured substring, to make it easier to tell which of a
797        number of identical substrings has been captured.
798    
799    24. The way atomic groups are processed by pcre_exec() has been changed so that
800        if they are repeated, backtracking one repetition now resets captured
801        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
802        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
803        "aaa". Previously, it would have been "a". As part of this code
804        refactoring, the way recursive calls are handled has also been changed.
805    
806    25. If an assertion condition captured any substrings, they were not passed
807        back unless some other capturing happened later. For example, if
808        (?(?=(a))a) was matched against "a", no capturing was returned.
809    
810    26. When studying a pattern that contained subroutine calls or assertions,
811        the code for finding the minimum length of a possible match was handling
812        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
813        group 1 called group 2 while simultaneously a separate group 2 called group
814        1). A stack overflow occurred in this case. I have fixed this by limiting
815        the recursion depth to 10.
816    
817    27. Updated RunTest.bat in the distribution to the version supplied by Tom
818        Fortmann. This supports explicit test numbers on the command line, and has
819        argument validation and error reporting.
820    
821    28. An instance of \X with an unlimited repeat could fail if at any point the
822        first character it looked at was a mark character.
823    
824    29. Some minor code refactoring concerning Unicode properties and scripts
825        should reduce the stack requirement of match() slightly.
826    
827    30. Added the '=' option to pcretest to check the setting of unused capturing
828        slots at the end of the pattern, which are documented as being -1, but are
829        not included in the return count.
830    
831    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
832        compiled something random. Now it gives a compile-time error (as does
833        Perl).
834    
835    32. A *MARK encountered during the processing of a positive assertion is now
836        recorded and passed back (compatible with Perl).
837    
838    33. If --only-matching or --colour was set on a pcregrep call whose pattern
839        had alternative anchored branches, the search for a second match in a line
840        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
841        matched the line "0102" twice. The same bug affected patterns that started
842        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
843        twice.
844    
845    34. Previously, PCRE did not allow quantification of assertions. However, Perl
846        does, and because of capturing effects, quantifying parenthesized
847        assertions may at times be useful. Quantifiers are now allowed for
848        parenthesized assertions.
849    
850    35. A minor code tidy in pcre_compile() when checking options for \R usage.
851    
852    36. \g was being checked for fancy things in a character class, when it should
853        just be a literal "g".
854    
855    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
856        appearance of a nested POSIX class supersedes an apparent external class.
857        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
858        unescaped square brackets may also appear as part of class names. For
859        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
860        more like Perl. (But see 8.20/1 above.)
861    
862    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
863        was because it thought it was \N{name}, which is not supported).
864    
865    39. Add minix to OS list not supporting the -S option in pcretest.
866    
867    40. PCRE tries to detect cases of infinite recursion at compile time, but it
868        cannot analyze patterns in sufficient detail to catch mutual recursions
869        such as ((?1))((?2)). There is now a runtime test that gives an error if a
870        subgroup is called recursively as a subpattern for a second time at the
871        same position in the subject string. In previous releases this might have
872        been caught by the recursion limit, or it might have run out of stack.
873    
874    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
875        happen only once. PCRE was, however incorrectly giving a compile time error
876        "recursive call could loop indefinitely" because it cannot analyze the
877        pattern in sufficient detail. The compile time test no longer happens when
878        PCRE is compiling a conditional subpattern, but actual runaway loops are
879        now caught at runtime (see 40 above).
880    
881    42. It seems that Perl allows any characters other than a closing parenthesis
882        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
883        has been changed to be the same.
884    
885    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
886        as not to get warnings when autogen.sh is called. Also changed
887        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
888    
889    44. To help people who use pcregrep to scan files containing exceedingly long
890        lines, the following changes have been made:
891    
892        (a) The default value of the buffer size parameter has been increased from
893            8K to 20K. (The actual buffer used is three times this size.)
894    
895        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
896            PCRE is built.
897    
898        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
899            to be set at run time.
900    
901        (d) Numerical values in pcregrep options can be followed by K or M, for
902            example --buffer-size=50K.
903    
904        (e) If a line being scanned overflows pcregrep's buffer, an error is now
905            given and the return code is set to 2.
906    
907    45. Add a pointer to the latest mark to the callout data block.
908    
909    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
910        partial match of an empty string instead of no match. This was specific to
911        the use of ".".
912    
913    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
914        complete match instead of a partial match. This bug was dependent on both
915        the PCRE_UTF8 and PCRE_DOTALL options being set.
916    
917    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
918        starting byte set, because \b was not being ignored.
919    
920    
921    Version 8.12 15-Jan-2011
922    ------------------------
923    
924    1.  Fixed some typos in the markup of the man pages, and wrote a script that
925        checks for such things as part of the documentation building process.
926    
927    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
928        --match-limit and --recursion-limit options (added for 8.11). In
929        particular, this made one of the standard tests fail. (The integer value
930        went into the wrong half of a long int.)
931    
932    3.  If the --colour option was given to pcregrep with -v (invert match), it
933        did strange things, either producing crazy output, or crashing. It should,
934        of course, ignore a request for colour when reporting lines that do not
935        match.
936    
937    4.  Another pcregrep bug caused similar problems if --colour was specified with
938        -M (multiline) and the pattern match finished with a line ending.
939    
940    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
941        matched in multiline mode, the following line was shown as part of the
942        match. This seems wrong, so I have changed it.
943    
944    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
945        the check for further matches in the same line (so they could be coloured)
946        to overrun the end of the current line. If another match was found, it was
947        incorrectly shown (and then shown again when found in the next line).
948    
949    7.  If pcregrep was compiled under Windows, there was a reference to the
950        function pcregrep_exit() before it was defined. I am assuming this was
951        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
952        reported by a user. I've moved the definition above the reference.
953    
954    
955    Version 8.11 10-Dec-2010
956    ------------------------
957    
958    1.  (*THEN) was not working properly if there were untried alternatives prior
959        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
960        backtracked to try for "b" instead of moving to the next alternative branch
961        at the same level (in this case, to look for "c"). The Perl documentation
962        is clear that when (*THEN) is backtracked onto, it goes to the "next
963        alternative in the innermost enclosing group".
964    
965    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
966        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
967        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
968        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
969        (*THEN).
970    
971    3.  If \s appeared in a character class, it removed the VT character from
972        the class, even if it had been included by some previous item, for example
973        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
974        of \s, but is part of the POSIX "space" class.)
975    
976    4.  A partial match never returns an empty string (because you can always
977        match an empty string at the end of the subject); however the checking for
978        an empty string was starting at the "start of match" point. This has been
979        changed to the "earliest inspected character" point, because the returned
980        data for a partial match starts at this character. This means that, for
981        example, /(?<=abc)def/ gives a partial match for the subject "abc"
982        (previously it gave "no match").
983    
984    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
985        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
986        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
987        has an implication that the given string is incomplete (because a partial
988        match is preferred over a full match). For this reason, these items now
989        give a partial match in this situation. [Aside: previously, the one case
990        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
991        match rather than a full match, which was wrong by the old rules, but is
992        now correct.]
993    
994    6.  There was a bug in the handling of #-introduced comments, recognized when
995        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
996        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
997        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
998        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
999        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1000        places in pcre_compile().
1001    
1002    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1003        comments when looking ahead for named forward references to subpatterns,
1004        the only newline sequence it recognized was NL. It now handles newlines
1005        according to the set newline convention.
1006    
1007    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1008        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1009        cater for a lack of strerror(). These oversights have been fixed.
1010    
1011    9.  Added --match-limit and --recursion-limit to pcregrep.
1012    
1013    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1014    
1015    11. When the -o option was used, pcregrep was setting a return code of 1, even
1016        when matches were found, and --line-buffered was not being honoured.
1017    
1018    12. Added an optional parentheses number to the -o and --only-matching options
1019        of pcregrep.
1020    
1021    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1022        can match an empty string. The code to do it in pcretest and pcredemo
1023        needed fixing:
1024    
1025        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1026            only one byte after an empty string match just before CRLF (this case
1027            just got forgotten; "any" and "anycrlf" were OK).
1028    
1029        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1030            mode when an empty string match preceded an ASCII character followed by
1031            a non-ASCII character. (The code for advancing by one character rather
1032            than one byte was nonsense.)
1033    
1034        (c) The pcredemo.c sample program did not have any code at all to handle
1035            the cases when CRLF is a valid newline sequence.
1036    
1037    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1038        as a starting offset was within the subject string. There is now a new
1039        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1040        negative or greater than the length of the string. In order to test this,
1041        pcretest is extended to allow the setting of negative starting offsets.
1042    
1043    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1044        starting offset points to the beginning of a UTF-8 character was
1045        unnecessarily clumsy. I tidied it up.
1046    
1047    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1048        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1049    
1050    17. Nobody had reported that the --include_dir option, which was added in
1051        release 7.7 should have been called --include-dir (hyphen, not underscore)
1052        for compatibility with GNU grep. I have changed it to --include-dir, but
1053        left --include_dir as an undocumented synonym, and the same for
1054        --exclude-dir, though that is not available in GNU grep, at least as of
1055        release 2.5.4.
1056    
1057    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1058        characters from a string of bytes) have been redefined so as not to use
1059        loops, in order to improve performance in some environments. At the same
1060        time, I abstracted some of the common code into auxiliary macros to save
1061        repetition (this should not affect the compiled code).
1062    
1063    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1064        compile-time error is now given if \c is not followed by an ASCII
1065        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1066        different, and any byte value is allowed.)
1067    
1068    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1069        START_OPTIMIZE option, which is now allowed at compile time - but just
1070        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1071        to pcregrep and other applications that have no direct access to PCRE
1072        options. The new /Y option in pcretest sets this option when calling
1073        pcre_compile().
1074    
1075    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1076        back references. Groups containing recursive back references were forced to
1077        be atomic by that change, but in the case of named groups, the amount of
1078        memory required was incorrectly computed, leading to "Failed: internal
1079        error: code overflow". This has been fixed.
1080    
1081    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1082        pcretest.c, to avoid build problems in some Borland environments.
1083    
1084    
1085    Version 8.10 25-Jun-2010
1086    ------------------------
1087    
1088    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1089        THEN.
1090    
1091    2.  (*ACCEPT) was not working when inside an atomic group.
1092    
1093    3.  Inside a character class, \B is treated as a literal by default, but
1094        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1095        causes the error). The code is unchanged, but I tidied the documentation.
1096    
1097    4.  Inside a character class, PCRE always treated \R and \X as literals,
1098        whereas Perl faults them if its -w option is set. I have changed PCRE so
1099        that it faults them when PCRE_EXTRA is set.
1100    
1101    5.  Added support for \N, which always matches any character other than
1102        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1103    
1104    6.  When compiling pcregrep with newer versions of gcc which may have
1105        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1106        declared with attribute warn_unused_result" were given. Just casting the
1107        result to (void) does not stop the warnings; a more elaborate fudge is
1108        needed. I've used a macro to implement this.
1109    
1110    7.  Minor change to pcretest.c to avoid a compiler warning.
1111    
1112    8.  Added four artifical Unicode properties to help with an option to make
1113        \s etc use properties (see next item). The new properties are: Xan
1114        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1115    
1116    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1117        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1118        this option. Modified pcretest to add /W to test this facility. Added
1119        REG_UCP to make it available via the POSIX interface.
1120    
1121    10. Added --line-buffered to pcregrep.
1122    
1123    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1124        studied, and the match started with a letter with a code point greater than
1125        127 whose first byte was different to the first byte of the other case of
1126        the letter, the other case of this starting letter was not recognized
1127        (#976).
1128    
1129    12. If a pattern that was studied started with a repeated Unicode property
1130        test, for example, \p{Nd}+, there was the theoretical possibility of
1131        setting up an incorrect bitmap of starting bytes, but fortunately it could
1132        not have actually happened in practice until change 8 above was made (it
1133        added property types that matched character-matching opcodes).
1134    
1135    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1136        possible starting bytes for non-anchored patterns.
1137    
1138    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1139        \R, and also a number of cases that involve Unicode properties, both
1140        explicit and implicit when PCRE_UCP is set.
1141    
1142    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1143        input, it could crash or give wrong results if characters with values
1144        greater than 0xc0 were present in the subject string. (Detail: it assumed
1145        UTF-8 input when processing these items.)
1146    
1147    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1148        size_t is 64-bit (#991).
1149    
1150    17. Added a check for running out of memory when PCRE is compiled with
1151        --disable-stack-for-recursion (#990).
1152    
1153    18. If the last data line in a file for pcretest does not have a newline on
1154        the end, a newline was missing in the output.
1155    
1156    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1157        less than 128) in its various bitmaps. However, there is a facility for
1158        generating tables according to the current locale when PCRE is compiled. It
1159        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1160        characters, are recognized by isspace() and therefore were getting set in
1161        these tables, and indeed these tables seem to approximate to ISO 8859. This
1162        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1163        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1164        which of course cannot start UTF-8 characters. I have changed the code so
1165        that only real ASCII characters (less than 128) and the correct starting
1166        bytes for UTF-8 encodings are set for characters greater than 127 when in
1167        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1168        altogether.)
1169    
1170    20. Added the /T option to pcretest so as to be able to run tests with non-
1171        standard character tables, thus making it possible to include the tests
1172        used for 19 above in the standard set of tests.
1173    
1174    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1175        reference to a subpattern the other side of a comment that contains an
1176        opening parenthesis caused either an internal compiling error, or a
1177        reference to the wrong subpattern.
1178    
1179    
1180    Version 8.02 19-Mar-2010
1181    ------------------------
1182    
1183    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1184    
1185    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1186        configured.
1187    
1188    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1189        original author of that file, following a query about its status.
1190    
1191    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1192        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1193    
1194    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1195        quantifier applied to a forward-referencing subroutine call, could compile
1196        incorrect code or give the error "internal error: previously-checked
1197        referenced subpattern not found".
1198    
1199    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1200        variables to point to external functions. For these systems, therefore,
1201        pcre_malloc etc. are now initialized to local functions that call the
1202        relevant global functions.
1203    
1204    7.  There were two entries missing in the vectors called coptable and poptable
1205        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1206        I've fixed the data, and added a kludgy way of testing at compile time that
1207        the lengths are correct (equal to the number of opcodes).
1208    
1209    8.  Following on from 7, I added a similar kludge to check the length of the
1210        eint vector in pcreposix.c.
1211    
1212    9.  Error texts for pcre_compile() are held as one long string to avoid too
1213        much relocation at load time. To find a text, the string is searched,
1214        counting zeros. There was no check for running off the end of the string,
1215        which could happen if a new error number was added without updating the
1216        string.
1217    
1218    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1219    
1220    11. \K was not working if it appeared in an atomic group or in a group that
1221        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1222        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1223        the assertion is positive, but not if it is negative.
1224    
1225    12. Change 11 fortuitously reduced the size of the stack frame used in the
1226        "match()" function of pcre_exec.c by one pointer. Forthcoming
1227        implementation of support for (*MARK) will need an extra pointer on the
1228        stack; I have reserved it now, so that the stack frame size does not
1229        decrease.
1230    
1231    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1232        item in branch that calls a recursion is a subroutine call - as in the
1233        second branch in the above example - was incorrectly given the compile-
1234        time error "recursive call could loop indefinitely" because pcre_compile()
1235        was not correctly checking the subroutine for matching a non-empty string.
1236    
1237    14. The checks for overrunning compiling workspace could trigger after an
1238        overrun had occurred. This is a "should never occur" error, but it can be
1239        triggered by pathological patterns such as hundreds of nested parentheses.
1240        The checks now trigger 100 bytes before the end of the workspace.
1241    
1242    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1243    
1244    
1245    Version 8.01 19-Jan-2010
1246    ------------------------
1247    
1248    1.  If a pattern contained a conditional subpattern with only one branch (in
1249        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1250        computed the wrong minimum data length (which is of course zero for such
1251        subpatterns). This could cause incorrect "no match" results.
1252    
1253    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1254        the pattern is reset in the first branch, pcre_compile() failed with
1255        "internal error: code overflow at offset...". This happened only when
1256        the reset was to the original external option setting. (An optimization
1257        abstracts leading options settings into an external setting, which was the
1258        cause of this.)
1259    
1260    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1261        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1262        assertion pattern did not match (meaning that the assertion was true), it
1263        was incorrectly treated as false if the SKIP had been reached during the
1264        matching. This also applied to assertions used as conditions.
1265    
1266    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1267        assertion subpattern, including such a pattern used as a condition,
1268        unpredictable results occurred, instead of the error return
1269        PCRE_ERROR_DFA_UITEM.
1270    
1271    5.  The C++ GlobalReplace function was not working like Perl for the special
1272        situation when an empty string is matched. It now does the fancy magic
1273        stuff that is necessary.
1274    
1275    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1276        removed. (These were left over from very, very early versions of PCRE.)
1277    
1278    7.  Some cosmetic changes to the code to make life easier when compiling it
1279        as part of something else:
1280    
1281        (a) Change DEBUG to PCRE_DEBUG.
1282    
1283        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1284            called "current" as "current_branch", to prevent a collision with the
1285            Linux macro when compiled as a kernel module.
1286    
1287        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1288            prevent a collision with the Linux macro when compiled as a kernel
1289            module.
1290    
1291    8.  In pcre_compile() there are some checks for integer overflows that used to
1292        cast potentially large values to (double). This has been changed to that
1293        when building, a check for int64_t is made, and if it is found, it is used
1294        instead, thus avoiding the use of floating point arithmetic. (There is no
1295        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1296        double.
1297    
1298    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1299        2005 (difference between two addresses compared to an unsigned value).
1300    
1301    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1302        custom one, because of the following reported problem in Windows:
1303    
1304          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1305              under Win32.
1306          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1307              therefore missing the function definition.
1308          - The compiler thus generates a "C" signature for the test function.
1309          - The linker fails to find the "C" function.
1310          - PCRE fails to configure if asked to do so against libbz2.
1311    
1312    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1313        messages were output:
1314    
1315          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1316          rerunning libtoolize, to keep the correct libtool macros in-tree.
1317          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1318    
1319        I have done both of these things.
1320    
1321    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1322        most of the time, it *can* run out if it is given a pattern that contains a
1323        runaway infinite recursion. I updated the discussion in the pcrestack man
1324        page.
1325    
1326    13. Now that we have gone to the x.xx style of version numbers, the minor
1327        version may start with zero. Using 08 or 09 is a bad idea because users
1328        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1329        interpreted as invalid octal numbers. I've updated the previous comment in
1330        configure.ac, and also added a check that gives an error if 08 or 09 are
1331        used.
1332    
1333    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1334        causing partial matching to fail when the end of the subject matched \W
1335        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1336    
1337    15. There were some discrepancies between the declarations in pcre_internal.h
1338        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1339        their definitions. The declarations used "const uschar *" and the
1340        definitions used USPTR. Even though USPTR is normally defined as "const
1341        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1342        reported that: "This difference in casting confuses some C++ compilers, for
1343        example, SunCC recognizes above declarations as different functions and
1344        generates broken code for hbpcre." I have changed the declarations to use
1345        USPTR.
1346    
1347    16. GNU libtool is named differently on some systems. The autogen.sh script now
1348        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1349        (FreeBSD).
1350    
1351    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1352        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1353        comment: "Figure out how to create a longlong from a string: strtoll and
1354        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1355        instance, but it only takes 2 args instead of 3!"
1356    
1357    18. A subtle bug concerned with back references has been fixed by a change of
1358        specification, with a corresponding code fix. A pattern such as
1359        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1360        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1361        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1362        same bug. Such groups have to be quantified to be useful, or contained
1363        inside another quantified group. (If there's no repetition, the reference
1364        can never match.) The problem arises because, having left the group and
1365        moved on to the rest of the pattern, a later failure that backtracks into
1366        the group uses the captured value from the final iteration of the group
1367        rather than the correct earlier one. I have fixed this in PCRE by forcing
1368        any group that contains a reference to itself to be an atomic group; that
1369        is, there cannot be any backtracking into it once it has completed. This is
1370        similar to recursive and subroutine calls.
1371    
1372    
1373    Version 8.00 19-Oct-09
1374    ----------------------
1375    
1376    1.  The table for translating pcre_compile() error codes into POSIX error codes
1377        was out-of-date, and there was no check on the pcre_compile() error code
1378        being within the table. This could lead to an OK return being given in
1379        error.
1380    
1381    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1382        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1383        in a Windows environment.
1384    
1385    3.  The pcregrep --count option prints the count for each file even when it is
1386        zero, as does GNU grep. However, pcregrep was also printing all files when
1387        --files-with-matches was added. Now, when both options are given, it prints
1388        counts only for those files that have at least one match. (GNU grep just
1389        prints the file name in this circumstance, but including the count seems
1390        more useful - otherwise, why use --count?) Also ensured that the
1391        combination -clh just lists non-zero counts, with no names.
1392    
1393    4.  The long form of the pcregrep -F option was incorrectly implemented as
1394        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1395        but it seems right to fix it, and I didn't think it was worth preserving
1396        the old behaviour.
1397    
1398    5.  The command line items --regex=pattern and --regexp=pattern were not
1399        recognized by pcregrep, which required --regex pattern or --regexp pattern
1400        (with a space rather than an '='). The man page documented the '=' forms,
1401        which are compatible with GNU grep; these now work.
1402    
1403    6.  No libpcreposix.pc file was created for pkg-config; there was just
1404        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1405    
1406    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1407        when UCP support is not needed, by modifying the Python script that
1408        generates it from Unicode data files. This should not matter if the module
1409        is correctly used as a library, but I received one complaint about 50K of
1410        unwanted data. My guess is that the person linked everything into his
1411        program rather than using a library. Anyway, it does no harm.
1412    
1413    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1414        was a minimum greater than 1 for a wide character in a possessive
1415        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1416        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1417        character. Chaos in the form of incorrect output or a compiling loop could
1418        result.
1419    
1420    9.  The restrictions on what a pattern can contain when partial matching is
1421        requested for pcre_exec() have been removed. All patterns can now be
1422        partially matched by this function. In addition, if there are at least two
1423        slots in the offset vector, the offset of the earliest inspected character
1424        for the match and the offset of the end of the subject are set in them when
1425        PCRE_ERROR_PARTIAL is returned.
1426    
1427    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1428        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1429        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1430        and may be more useful for multi-segment matching.
1431    
1432    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1433        used to be given if ever the end of the subject was reached; now it is
1434        given only if matching could not proceed because another character was
1435        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1436        string "Z", which now yields "no match" instead of "partial match". In the
1437        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1438        final character ended with (*FAIL).
1439    
1440    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1441        if the pattern had a "must contain" character that was already found in the
1442        earlier partial match, unless partial matching was again requested. For
1443        example, with the pattern /dog.(body)?/, the "must contain" character is
1444        "g". If the first part-match was for the string "dog", restarting with
1445        "sbody" failed. This bug has been fixed.
1446    
1447    13. The string returned by pcre_dfa_exec() after a partial match has been
1448        changed so that it starts at the first inspected character rather than the
1449        first character of the match. This makes a difference only if the pattern
1450        starts with a lookbehind assertion or \b or \B (\K is not supported by
1451        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1452        matching functions compatible, and I think it's the right thing to do.
1453    
1454    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1455        so that the demonstration program is easily available in environments where
1456        PCRE has not been installed from source.
1457    
1458    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1459        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1460        library.
1461    
1462    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1463        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1464        is not the first non-POSIX option to be added. Clearly some people find
1465        these options useful.
1466    
1467    17. If a caller to the POSIX matching function regexec() passes a non-zero
1468        value for nmatch with a NULL value for pmatch, the value of
1469        nmatch is forced to zero.
1470    
1471    18. RunGrepTest did not have a test for the availability of the -u option of
1472        the diff command, as RunTest does. It now checks in the same way as
1473        RunTest, and also checks for the -b option.
1474    
1475    19. If an odd number of negated classes containing just a single character
1476        interposed, within parentheses, between a forward reference to a named
1477        subpattern and the definition of the subpattern, compilation crashed with
1478        an internal error, complaining that it could not find the referenced
1479        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1480        [The bug was that it was starting one character too far in when skipping
1481        over the character class, thus treating the ] as data rather than
1482        terminating the class. This meant it could skip too much.]
1483    
1484    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1485        /g option in pcretest when the pattern contains \K, which makes it possible
1486        to have an empty string match not at the start, even when the pattern is
1487        anchored. Updated pcretest and pcredemo to use this option.
1488    
1489    21. If the maximum number of capturing subpatterns in a recursion was greater
1490        than the maximum at the outer level, the higher number was returned, but
1491        with unset values at the outer level. The correct (outer level) value is
1492        now given.
1493    
1494    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1495        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1496        make it do so. The string so far is captured, making this feature
1497        compatible with Perl.
1498    
1499    23. The tests have been re-organized, adding tests 11 and 12, to make it
1500        possible to check the Perl 5.10 features against Perl 5.10.
1501    
1502    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1503        pattern matches a fixed length string. PCRE did not allow this; now it
1504        does. Neither allows recursion.
1505    
1506    25. I finally figured out how to implement a request to provide the minimum
1507        length of subject string that was needed in order to match a given pattern.
1508        (It was back references and recursion that I had previously got hung up
1509        on.) This code has now been added to pcre_study(); it finds a lower bound
1510        to the length of subject needed. It is not necessarily the greatest lower
1511        bound, but using it to avoid searching strings that are too short does give
1512        some useful speed-ups. The value is available to calling programs via
1513        pcre_fullinfo().
1514    
1515    26. While implementing 25, I discovered to my embarrassment that pcretest had
1516        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1517        study optimizations had never been tested with that matching function.
1518        Oops. What is worse, even when it was passed study data, there was a bug in
1519        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1520        were also very few tests of studied patterns with pcre_dfa_exec().
1521    
1522    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1523        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1524        on the other side of the coin, they are no longer allowed to have different
1525        names, because these cannot be distinguished in PCRE, and this has caused
1526        confusion. (This is a difference from Perl.)
1527    
1528    28. When duplicate subpattern names are present (necessarily with different
1529        numbers, as required by 27 above), and a test is made by name in a
1530        conditional pattern, either for a subpattern having been matched, or for
1531        recursion in such a pattern, all the associated numbered subpatterns are
1532        tested, and the overall condition is true if the condition is true for any
1533        one of them. This is the way Perl works, and is also more like the way
1534        testing by number works.
1535    
1536    
1537    Version 7.9 11-Apr-09
1538    ---------------------
1539    
1540    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1541        (pcretest), all targets were linked against these libraries. This included
1542        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1543        libraries. This caused unwanted dependencies to be created. This problem
1544        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1545        pcretest is linked with readline.
1546    
1547    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1548        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1549        moved outside it again, because FALSE and TRUE are already defined in AIX,
1550        but BOOL is not.
1551    
1552    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1553        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1554    
1555    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1556        hyphens) following file names and line numbers when outputting matching
1557        lines. This is not true; no spaces are inserted. I have also clarified the
1558        wording for the --colour (or --color) option.
1559    
1560    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1561        was not coloured; this is different to GNU grep, so I have changed it to be
1562        the same.
1563    
1564    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1565        each matching line was coloured. Now it goes on to look for further matches
1566        of any of the test patterns, which is the same behaviour as GNU grep.
1567    
1568    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1569        doesn't make sense to accept an empty string match in pcregrep, so I have
1570        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1571        seems to be how GNU grep behaves.
1572    
1573    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1574        start or after a newline", because the conditional assertion was not being
1575        correctly handled. The rule now is that both the assertion and what follows
1576        in the first alternative must satisfy the test.
1577    
1578    9.  If auto-callout was enabled in a pattern with a conditional group whose
1579        condition was an assertion, PCRE could crash during matching, both with
1580        pcre_exec() and pcre_dfa_exec().
1581    
1582    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1583        used for matching.
1584    
1585    11. Unicode property support in character classes was not working for
1586        characters (bytes) greater than 127 when not in UTF-8 mode.
1587    
1588    12. Added the -M command line option to pcretest.
1589    
1590    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1591    
1592    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1593    
1594    16. Added comments and documentation about mis-use of no_arg in the C++
1595        wrapper.
1596    
1597    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1598        from Martin Jerabek that uses macro names for all relevant character and
1599        string constants.
1600    
1601    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1602        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1603        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1604        these, but not everybody uses configure.
1605    
1606    19. A conditional group that had only one branch was not being correctly
1607        recognized as an item that could match an empty string. This meant that an
1608        enclosing group might also not be so recognized, causing infinite looping
1609        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1610        with the subject "ab", where knowledge that the repeated group can match
1611        nothing is needed in order to break the loop.
1612    
1613    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1614        exec(), but without supplying a callout function, matching went wrong.
1615    
1616    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1617        leak if the size of the offset vector was greater than 30. When the vector
1618        is smaller, the saved offsets during recursion go onto a local stack
1619        vector, but for larger vectors malloc() is used. It was failing to free
1620        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1621        error, in fact).
1622    
1623    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1624        heapframe that is used only when UTF-8 support is enabled. This caused no
1625        problem, but was untidy.
1626    
1627    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1628        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1629        included within another project.
1630    
1631    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1632        slightly modified by me:
1633    
1634          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1635              not building pcregrep.
1636    
1637          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1638              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1639    
1640    25. Forward references, both numeric and by name, in patterns that made use of
1641        duplicate group numbers, could behave incorrectly or give incorrect errors,
1642        because when scanning forward to find the reference group, PCRE was not
1643        taking into account the duplicate group numbers. A pattern such as
1644        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1645    
1646    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1647        the feature of a custom pointer more persuasive (as requested by a user).
1648    
1649    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1650        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1651        pre-defined.
1652    
1653    28. Added support for (*UTF8) at the start of a pattern.
1654    
1655    29. Arrange for flags added by the "release type" setting in CMake to be shown
1656        in the configuration summary.
1657    
1658    
1659    Version 7.8 05-Sep-08
1660    ---------------------
1661    
1662    1.  Replaced UCP searching code with optimized version as implemented for Ad
1663        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1664        stage table and inline lookup instead of a function, giving speed ups of 2
1665        to 5 times on some simple patterns that I tested. Permission was given to
1666        distribute the MultiStage2.py script that generates the tables (it's not in
1667        the tarball, but is in the Subversion repository).
1668    
1669    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1670        scripts.
1671    
1672    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1673        a group with a zero qualifier. The result of the study could be incorrect,
1674        or the function might crash, depending on the pattern.
1675    
1676    4.  Caseless matching was not working for non-ASCII characters in back
1677        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1678        It now works when Unicode Property Support is available.
1679    
1680    5.  In pcretest, an escape such as \x{de} in the data was always generating
1681        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1682        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1683        truncation.
1684    
1685    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1686    
1687    7.  Added two (int) casts to pcregrep when printing the difference of two
1688        pointers, in case they are 64-bit values.
1689    
1690    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1691        test 2 if it fails.
1692    
1693    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1694        and a #define of that name to empty if it is not externally set. This is to
1695        allow users of MSVC to set it if necessary.
1696    
1697    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1698        the convenience functions in the pcre_get.c source file.
1699    
1700    11. An option change at the start of a pattern that had top-level alternatives
1701        could cause overwriting and/or a crash. This command provoked a crash in
1702        some environments:
1703    
1704          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1705    
1706        This potential security problem was recorded as CVE-2008-2371.
1707    
1708    12. For a pattern where the match had to start at the beginning or immediately
1709        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1710        pcre_dfa_exec() could read past the end of the passed subject if there was
1711        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1712        pcretest so that it places the subject at the end of its malloc-ed buffer.
1713    
1714    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1715        exec() might read past the end of the data buffer in UTF-8 mode.
1716    
1717    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1718        the data contained the byte 0x85 as part of a UTF-8 character within its
1719        first line. This applied both to normal and DFA matching.
1720    
1721    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1722        /^[^d]*?$/8 failed to match "abc".
1723    
1724    16. Added a missing copyright notice to pcrecpp_internal.h.
1725    
1726    17. Make it more clear in the documentation that values returned from
1727        pcre_exec() in ovector are byte offsets, not character counts.
1728    
1729    18. Tidied a few places to stop certain compilers from issuing warnings.
1730    
1731    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1732        supplied by Stefan Weber. I made a further small update for 7.8 because
1733        there is a change of source arrangements: the pcre_searchfuncs.c module is
1734        replaced by pcre_ucd.c.
1735    
1736    
1737    Version 7.7 07-May-08
1738    ---------------------
1739    
1740    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1741        a string to a long long, pretend we don't even have a long long." This is
1742        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1743    
1744    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1745        pre-7.6 versions, which defined a global no_arg variable instead of putting
1746        it in the RE class. (See also #8 below.)
1747    
1748    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1749        Lopes.
1750    
1751    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1752    
1753        (1) The include/exclude patterns were being applied to the whole pathnames
1754            of files, instead of just to the final components.
1755    
1756        (2) If there was more than one level of directory, the subdirectories were
1757            skipped unless they satisfied the include/exclude conditions. This is
1758            inconsistent with GNU grep (and could even be seen as contrary to the
1759            pcregrep specification - which I improved to make it absolutely clear).
1760            The action now is always to scan all levels of directory, and just
1761            apply the include/exclude patterns to regular files.
1762    
1763    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1764        --exclude_dir in the tests to avoid scanning .svn directories.
1765    
1766    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1767        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1768        doesn't support NULs in patterns.
1769    
1770    7.  Added some missing "const"s to declarations of static tables in
1771        pcre_compile.c and pcre_dfa_exec.c.
1772    
1773    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1774        caused by fix #2  above. (Subsequently also a second patch to fix the
1775        first patch. And a third patch - this was a messy problem.)
1776    
1777    9.  Applied Craig's patch to remove the use of push_back().
1778    
1779    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1780        matching function regexec().
1781    
1782    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1783        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1784        references. PCRE supports relative numbers with this syntax (I don't think
1785        Oniguruma does).
1786    
1787    12. Previously, a group with a zero repeat such as (...){0} was completely
1788        omitted from the compiled regex. However, this means that if the group
1789        was called as a subroutine from elsewhere in the pattern, things went wrong
1790        (an internal error was given). Such groups are now left in the compiled
1791        pattern, with a new opcode that causes them to be skipped at execution
1792        time.
1793    
1794    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1795        to the way PCRE behaves:
1796    
1797        (a) A lone ] character is dis-allowed (Perl treats it as data).
1798    
1799        (b) A back reference to an unmatched subpattern matches an empty string
1800            (Perl fails the current match path).
1801    
1802        (c) A data ] in a character class must be notated as \] because if the
1803            first data character in a class is ], it defines an empty class. (In
1804            Perl it is not possible to have an empty class.) The empty class []
1805            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1806            The negative empty class [^] matches any one character, independently
1807            of the DOTALL setting.
1808    
1809    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1810        non-existent subpattern following a character class starting with ']' and
1811        containing () gave an internal compiling error instead of "reference to
1812        non-existent subpattern". Fortunately, when the pattern did exist, the
1813        compiled code was correct. (When scanning forwards to check for the
1814        existencd of the subpattern, it was treating the data ']' as terminating
1815        the class, so got the count wrong. When actually compiling, the reference
1816        was subsequently set up correctly.)
1817    
1818    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1819        it was being rejected as not supported by pcre_dfa_exec(), even though
1820        other assertions are supported. I have made pcre_dfa_exec() support
1821        (*FAIL).
1822    
1823    16. The implementation of 13c above involved the invention of a new opcode,
1824        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1825        cannot be changed at match time, I realized I could make a small
1826        improvement to matching performance by compiling OP_ALLANY instead of
1827        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1828        on the OP_ANY path.
1829    
1830    17. Compiling pcretest on Windows with readline support failed without the
1831        following two fixes: (1) Make the unistd.h include conditional on
1832        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1833    
1834    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1835        ncurses library to be included for pcretest when ReadLine support is
1836        requested, but also to allow for it to be overridden. This patch came from
1837        Daniel Bergström.
1838    
1839    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1840        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1841        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1842        spotting this.
1843    
1844    
1845    Version 7.6 28-Jan-08
1846    ---------------------
1847    
1848    1.  A character class containing a very large number of characters with
1849        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1850        overflow.
1851    
1852    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1853        HAVE_LONG_LONG is not defined.
1854    
1855    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1856        bring them up to date and include new features. This patch includes:
1857    
1858        - Fixed PH's badly added libz and libbz2 support.
1859        - Fixed a problem with static linking.
1860        - Added pcredemo. [But later removed - see 7 below.]
1861        - Fixed dftables problem and added an option.
1862        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1863            HAVE_LONG_LONG.
1864        - Added readline support for pcretest.
1865        - Added an listing of the option settings after cmake has run.
1866    
1867    4.  A user submitted a patch to Makefile that makes it easy to create
1868        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1869        Makefile.am that cause it to include this special target, without
1870        affecting anything else. Note that the same mingw target plus all
1871        the other distribution libraries and programs are now supported
1872        when configuring with CMake (see 6 below) instead of with
1873        Configure/Make.
1874    
1875    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1876        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1877        exported in the Windows port". It has not yet been confirmed that the patch
1878        solves the problem, but it does no harm.
1879    
1880    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1881        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1882        with CMake, and also correct the comment about stack recursion.
1883    
1884    7.  Remove the automatic building of pcredemo from the ./configure system and
1885        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1886        of a program that users should build themselves after PCRE is installed, so
1887        building it automatically is not really right. What is more, it gave
1888        trouble in some build environments.
1889    
1890    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1891    
1892    
1893    Version 7.5 10-Jan-08
1894    ---------------------
1895    
1896    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1897        values in parens when parsing an RE using the C++ wrapper."
1898    
1899    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1900        Characters greater than 255 were excluded from the class instead of being
1901        included.
1902    
1903    3.  The same bug as (2) above applied to negated POSIX classes such as
1904        [:^space:].
1905    
1906    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1907        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1908        I have changed it.
1909    
1910    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1911        first named subpattern) and a construct such as (?&a) would reference the
1912        first named subpattern whose name started with "a" (in other words, the
1913        length check was missing). Both these problems are fixed. "Subpattern name
1914        expected" is now given for (?&) (a zero-length name), and this patch also
1915        makes it give the same error for \k'' (previously it complained that that
1916        was a reference to a non-existent subpattern).
1917    
1918    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1919        this is right because (?- can be followed by option settings as well as by
1920        digits. I have, however, made the messages clearer.
1921    
1922    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1923        than the number used in the conditional) now cause a compile-time error.
1924        This is actually not compatible with Perl, which accepts such patterns, but
1925        treats the conditional as always being FALSE (as PCRE used to), but it
1926        seems to me that giving a diagnostic is better.
1927    
1928    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1929        and messages.
1930    
1931    9.  Fix two occurrences of "backslash" in comments that should have been
1932        "backspace".
1933    
1934    10. Remove two redundant lines of code that can never be obeyed (their function
1935        was moved elsewhere).
1936    
1937    11. The program that makes PCRE's Unicode character property table had a bug
1938        which caused it to generate incorrect table entries for sequences of
1939        characters that have the same character type, but are in different scripts.
1940        It amalgamated them into a single range, with the script of the first of
1941        them. In other words, some characters were in the wrong script. There were
1942        thirteen such cases, affecting characters in the following ranges:
1943    
1944          U+002b0 - U+002c1
1945          U+0060c - U+0060d
1946          U+0061e - U+00612
1947          U+0064b - U+0065e
1948          U+0074d - U+0076d
1949          U+01800 - U+01805
1950          U+01d00 - U+01d77
1951          U+01d9b - U+01dbf
1952          U+0200b - U+0200f
1953          U+030fc - U+030fe
1954          U+03260 - U+0327f
1955          U+0fb46 - U+0fbb1
1956          U+10450 - U+1049d
1957    
1958    12. The -o option (show only the matching part of a line) for pcregrep was not
1959        compatible with GNU grep in that, if there was more than one match in a
1960        line, it showed only the first of them. It now behaves in the same way as
1961        GNU grep.
1962    
1963    13. If the -o and -v options were combined for pcregrep, it printed a blank
1964        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1965        does the same. The return code can be used to tell if there were any
1966        non-matching lines.
1967    
1968    14. Added --file-offsets and --line-offsets to pcregrep.
1969    
1970    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1971        infinitely looping recursion. The bug was that positive lookaheads were not
1972        being skipped when checking for a possible empty match (negative lookaheads
1973        and both kinds of lookbehind were skipped).
1974    
1975    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1976        inclusion of <windows.h> to before rather than after the definition of
1977        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1978    
1979    17. Specifying a possessive quantifier with a specific limit for a Unicode
1980        character property caused pcre_compile() to compile bad code, which led at
1981        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1982        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1983        caused the error; without that there was no problem.
1984    
1985    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1986    
1987    19. Added --enable-pcretest-libreadline.
1988    
1989    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1990        RE::GlobalReplace(). As a result, the number of replacements returned was
1991        double what it should be. I removed one of the increments, but Craig sent a
1992        later patch that removed the other one (the right fix) and added unit tests
1993        that check the return values (which was not done before).
1994    
1995    21. Several CMake things:
1996    
1997        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1998            the names libpcre and libpcreposix, not just pcre and pcreposix.
1999    
2000        (2) The above change means that pcretest and pcregrep are now correctly
2001            linked with the newly-built libraries, not previously installed ones.
2002    
2003        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2004    
2005    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2006        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2007        UTF-8 newline character). The key issue is that the pattern starts .*;
2008        this means that the match must be either at the beginning, or after a
2009        newline. The bug was in the code for advancing after a failed match and
2010        checking that the new position followed a newline. It was not taking
2011        account of UTF-8 characters correctly.
2012    
2013    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2014        character classes. PCRE was not treating the sequence [:...:] as a
2015        character class unless the ... were all letters. Perl, however, seems to
2016        allow any characters between [: and :], though of course it rejects as
2017        unknown any "names" that contain non-letters, because all the known class
2018        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2019        for example, whereas PCRE did not - it did not recognize a POSIX character
2020        class. This seemed a bit dangerous, so the code has been changed to be
2021        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2022        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2023        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2024        Perl does, and where it didn't before.
2025    
2026    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2027        Windows environments %n is disabled by default.
2028    
2029    
2030    Version 7.4 21-Sep-07
2031    ---------------------
2032    
2033    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2034        means that a class such as [\s] counted as "explicit reference to CR or
2035        LF". That isn't really right - the whole point of the change was to try to
2036        help when there was an actual mention of one of the two characters. So now
2037        the change happens only if \r or \n (or a literal CR or LF) character is
2038        encountered.
2039    
2040    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2041        of both had grown to the point where there were only 3 bits left.
2042        Fortunately, there was spare space in the data structure, and so I have
2043        moved the internal flags into a new 16-bit field to free up more option
2044        bits.
2045    
2046    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2047        but did not set the internal JCHANGED flag - either of these is enough to
2048        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2049        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2050        start sets both bits.
2051    
2052    4.  Added options (at build time, compile time, exec time) to change \R from
2053        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2054    
2055    5.  doc/pcresyntax.html was missing from the distribution.
2056    
2057    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2058        compatibility, even though it is no longer used.
2059    
2060    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2061        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2062        windows.h file is present (where different names are used). [This was
2063        reversed later after testing - see 16 below.]
2064    
2065    8.  Changed all #include <config.h> to #include "config.h". There were also
2066        some further <pcre.h> cases that I changed to "pcre.h".
2067    
2068    9.  When pcregrep was used with the --colour option, it missed the line ending
2069        sequence off the lines that it output.
2070    
2071    10. It was pointed out to me that arrays of string pointers cause lots of
2072        relocations when a shared library is dynamically loaded. A technique of
2073        using a single long string with a table of offsets can drastically reduce
2074        these. I have refactored PCRE in four places to do this. The result is
2075        dramatic:
2076    
2077          Originally:                          290
2078          After changing UCP table:            187
2079          After changing error message table:   43
2080          After changing table of "verbs"       36
2081          After changing table of Posix names   22
2082    
2083        Thanks to the folks working on Gregex for glib for this insight.
2084    
2085    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2086        unicode-properties was also set.
2087    
2088    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2089    
2090    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2091        checked only for CRLF.
2092    
2093    14. Added casts to pcretest.c to avoid compiler warnings.
2094    
2095    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2096    
2097    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2098        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2099        entirely. This removes changes made in 7 above.
2100    
2101    17. The CMake files have been updated, and there is now more information about
2102        building with CMake in the NON-UNIX-USE document.
2103    
2104    
2105    Version 7.3 28-Aug-07
2106    ---------------------
2107    
2108     1. In the rejigging of the build system that eventually resulted in 7.1, the
2109        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2110        brackets there is not right, since it causes compilers to look for an
2111        installed pcre.h, not the version that is in the source that is being
2112        compiled (which of course may be different). I have changed it back to:
2113    
2114          #include "pcre.h"
2115    
2116        I have a vague recollection that the change was concerned with compiling in
2117        different directories, but in the new build system, that is taken care of
2118        by the VPATH setting the Makefile.
2119    
2120     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2121        when the subject happened to end in the byte 0x85 (e.g. if the last
2122        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2123        characters but of course it shouldn't be taken as a newline when it is part
2124        of another character. The bug was that, for an unlimited repeat of . in
2125        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2126        characters when looking for a newline.
2127    
2128     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2129    
2130     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2131        in debug output.
2132    
2133     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2134        long printing in the pcrecpp unittest when running under MinGW.
2135    
2136     6. ESC_K was left out of the EBCDIC table.
2137    
2138     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2139        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2140        limit also applies to "virtual nesting" when a pattern is recursive, and in
2141        this case 1000 isn't so big. I have been able to remove this limit at the
2142        expense of backing off one optimization in certain circumstances. Normally,
2143        when pcre_exec() would call its internal match() function recursively and
2144        immediately return the result unconditionally, it uses a "tail recursion"
2145        feature to save stack. However, when a subpattern that can match an empty
2146        string has an unlimited repetition quantifier, it no longer makes this
2147        optimization. That gives it a stack frame in which to save the data for
2148        checking that an empty string has been matched. Previously this was taken
2149        from the 1000-entry workspace that had been reserved. So now there is no
2150        explicit limit, but more stack is used.
2151    
2152     8. Applied Daniel's patches to solve problems with the import/export magic
2153        syntax that is required for Windows, and which was going wrong for the
2154        pcreposix and pcrecpp parts of the library. These were overlooked when this
2155        problem was solved for the main library.
2156    
2157     9. There were some crude static tests to avoid integer overflow when computing
2158        the size of patterns that contain repeated groups with explicit upper
2159        limits. As the maximum quantifier is 65535, the maximum group length was
2160        set at 30,000 so that the product of these two numbers did not overflow a
2161        32-bit integer. However, it turns out that people want to use groups that
2162        are longer than 30,000 bytes (though not repeat them that many times).
2163        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2164        made it possible to implement the integer overflow checks in a much more
2165        dynamic way, which I have now done. The artificial limitation on group
2166        length has been removed - we now have only the limit on the total length of
2167        the compiled pattern, which depends on the LINK_SIZE setting.
2168    
2169    10. Fixed a bug in the documentation for get/copy named substring when
2170        duplicate names are permitted. If none of the named substrings are set, the
2171        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2172        empty string.
2173    
2174    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2175        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2176        because the ] is interpreted as the first data character and the
2177        terminating ] is not found. PCRE has been made compatible with Perl in this
2178        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2179        cause memory overwriting.
2180    
2181    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2182        string has been matched (to stop an infinite loop). It was not recognizing
2183        a conditional subpattern that could match an empty string if that
2184        subpattern was within another subpattern. For example, it looped when
2185        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2186        condition was not nested. This bug has been fixed.
2187    
2188    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2189        past the start of the subject in the presence of bytes with the top bit
2190        set, for example "\x8aBCD".
2191    
2192    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2193        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2194    
2195    14. Optimized (?!) to (*FAIL).
2196    
2197    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2198        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2199        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2200        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2201        does: it's just the validity check that is more restrictive.
2202    
2203    16. Inserted checks for integer overflows during escape sequence (backslash)
2204        processing, and also fixed erroneous offset values for syntax errors during
2205        backslash processing.
2206    
2207    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2208        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2209    
2210    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2211        caused an overrun.
2212    
2213    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2214        something other than just ASCII characters) inside a group that had an
2215        unlimited repeat caused a loop at compile time (while checking to see
2216        whether the group could match an empty string).
2217    
2218    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2219        [\P{Any}] did so. (Error in the code for printing property names.)
2220    
2221    21. An orphan \E inside a character class could cause a crash.
2222    
2223    22. A repeated capturing bracket such as (A)? could cause a wild memory
2224        reference during compilation.
2225    
2226    23. There are several functions in pcre_compile() that scan along a compiled
2227        expression for various reasons (e.g. to see if it's fixed length for look
2228        behind). There were bugs in these functions when a repeated \p or \P was
2229        present in the pattern. These operators have additional parameters compared
2230        with \d, etc, and these were not being taken into account when moving along
2231        the compiled data. Specifically:
2232    
2233        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2234            length.
2235    
2236        (b) An item such as \pL+ within a repeated group could cause crashes or
2237            loops.
2238    
2239        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2240            "reference to non-existent subpattern" error.
2241    
2242        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2243    
2244    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2245        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2246    
2247    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2248    
2249    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2250        character were causing crashes (broken optimization).
2251    
2252    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2253        \p or \P) caused a compile-time loop.
2254    
2255    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2256        break. For example, the unstudied pattern [\r\n]A does not match the string
2257        "\r\nA" because change 7.0/46 below moves the current point on by two
2258        characters after failing to match at the start. However, the pattern \nA
2259        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2260        the same is true. There doesn't seem any very clean way out of this, but
2261        what I have chosen to do makes the common cases work: PCRE now takes note
2262        of whether there can be an explicit match for \r or \n anywhere in the
2263        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2264        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2265        pattern has explicit CR or LF references.
2266    
2267    29. Added (*CR) etc for changing newline setting at start of pattern.
2268    
2269    
2270    Version 7.2 19-Jun-07
2271    ---------------------
2272    
2273     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2274        which is apparently normally available under Windows.
2275    
2276     2. Re-jig the pcregrep tests with different newline settings in an attempt
2277        to make them independent of the local environment's newline setting.
2278    
2279     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2280    
2281     4. Some of the "internals" tests were previously cut out when the link size
2282        was not 2, because the output contained actual offsets. The recent new
2283        "Z" feature of pcretest means that these can be cut out, making the tests
2284        usable with all link sizes.
2285    
2286     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2287        stack recursion. This gives a massive performance boost under BSD, but just
2288        a small improvement under Linux. However, it saves one field in the frame
2289        in all cases.
2290    
2291     6. Added more features from the forthcoming Perl 5.10:
2292    
2293        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2294            recursion call. It refers to the nth most recently opened parentheses.
2295    
2296        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2297            to be opened parentheses.
2298    
2299        (c) Conditions that refer to capturing parentheses can be specified
2300            relatively, for example, (?(-2)... or (?(+3)...
2301    
2302        (d) \K resets the start of the current match so that everything before
2303            is not part of it.
2304    
2305        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2306    
2307        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2308            reference syntax.
2309    
2310        (g) (?| introduces a group in which the numbering of parentheses in each
2311            alternative starts with the same number.
2312    
2313        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2314    
2315     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2316        PCRE_INFO_JCHANGED.
2317    
2318     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2319        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2320        for detecting groups that can match an empty string.
2321    
2322     9. A pattern with a very large number of alternatives (more than several
2323        hundred) was running out of internal workspace during the pre-compile
2324        phase, where pcre_compile() figures out how much memory will be needed. A
2325        bit of new cunning has reduced the workspace needed for groups with
2326        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2327        workspace instead of running out of the 4096 that are available.
2328    
2329    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2330    
2331    11. Applied patch from Google to remove an optimization that didn't quite work.
2332        The report of the bug said:
2333    
2334          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2335          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2336          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2337    
2338    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2339        it matched the wrong number of bytes.
2340    
2341    
2342    Version 7.1 24-Apr-07
2343  ---------------------  ---------------------
2344    
2345   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2346      that is more "standard", making use of automake and other autotools.      that is more "standard", making use of automake and other Autotools. There
2347        is some re-arrangement of the files and adjustment of comments consequent
2348        on this.
2349    
2350     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2351        for recursive directory scanning broke on some systems because the files
2352        are not scanned in any specific order and on different systems the order
2353        was different. A call to "sort" has been inserted into RunGrepTest for the
2354        approprate test as a short-term fix. In the longer term there may be an
2355        alternative.
2356    
2357     3. I had an email from Eric Raymond about problems translating some of PCRE's
2358        man pages to HTML (despite the fact that I distribute HTML pages, some
2359        people do their own conversions for various reasons). The problems
2360        concerned the use of low-level troff macros .br and .in. I have therefore
2361        removed all such uses from the man pages (some were redundant, some could
2362        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2363        HTML has been updated to handle .nf/.fi and to complain if it encounters
2364        .br or .in.
2365    
2366     4. Updated comments in configure.ac that get placed in config.h.in and also
2367        arranged for config.h to be included in the distribution, with the name
2368        config.h.generic, for the benefit of those who have to compile without
2369        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2370    
2371     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2372        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2373        makevp.bat for the current PCRE, using the additional files
2374        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2375    
2376     6. A Windows user reported a minor discrepancy with test 2, which turned out
2377        to be caused by a trailing space on an input line that had got lost in his
2378        copy. The trailing space was an accident, so I've just removed it.
2379    
2380     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2381        that is needed.
2382    
2383     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2384        as "const" (a) because they are and (b) because it helps the PHP
2385        maintainers who have recently made a script to detect big data structures
2386        in the php code that should be moved to the .rodata section. I remembered
2387        to update Builducptable as well, so it won't revert if ucptable.h is ever
2388        re-created.
2389    
2390     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2391        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2392        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2393        support is not required. This saves 1.5-2K of code, which is important in
2394        some applications.
2395    
2396        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2397        so as not to refer to the tables, even though these functions will never be
2398        called when UTF-8 support is disabled. Otherwise there are problems with a
2399        shared library.
2400    
2401    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2402    
2403        (a) It was defining its arguments as char * instead of void *.
2404    
2405        (b) It was assuming that all moves were upwards in memory; this was true
2406            a long time ago when I wrote it, but is no longer the case.
2407    
2408        The emulated memove() is provided for those environments that have neither
2409        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2410        is clearly not the case, as these two bugs were recently reported.
2411    
2412    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2413        and Detrail to create the HTML documentation, the .txt form of the man
2414        pages, and it removes trailing spaces from listed files. It also creates
2415        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2416        case, it wraps all the #defines with #ifndefs. This script should be run
2417        before "make dist".
2418    
2419    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2420        with Unicode property support.
2421    
2422        (a) For a maximizing quantifier, if the two different cases of the
2423            character were of different lengths in their UTF-8 codings (there are
2424            some cases like this - I found 11), and the matching function had to
2425            back up over a mixture of the two cases, it incorrectly assumed they
2426            were both the same length.
2427    
2428        (b) When PCRE was configured to use the heap rather than the stack for
2429            recursion during matching, it was not correctly preserving the data for
2430            the other case of a UTF-8 character when checking ahead for a match
2431            while processing a minimizing repeat. If the check also involved
2432            matching a wide character, but failed, corruption could cause an
2433            erroneous result when trying to check for a repeat of the original
2434            character.
2435    
2436    13. Some tidying changes to the testing mechanism:
2437    
2438        (a) The RunTest script now detects the internal link size and whether there
2439            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2440            values substituted by "configure". (The RunGrepTest script already did
2441            this for UTF-8.) The configure.ac script no longer substitutes the
2442            relevant variables.
2443    
2444        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2445            with length and offset values. This means that the output is different
2446            for different internal link sizes. Test 2 is skipped for link sizes
2447            other than 2 because of this, bypassing the problem. Unfortunately,
2448            there was also a test in test 3 (the locale tests) that used /B and
2449            failed for link sizes other than 2. Rather than cut the whole test out,
2450            I have added a new /Z option to pcretest that replaces the length and
2451            offset values with spaces. This is now used to make test 3 independent
2452            of link size. (Test 2 will be tidied up later.)
2453    
2454    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2455        segmentation fault instead of returning the appropriate error message.
2456    
2457    15. In multiline mode when the newline sequence was set to "any", the pattern
2458        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2459        This doesn't seem right; it now treats the CRLF combination as the line
2460        ending, and so does not match in that case. It's only a pattern such as ^$
2461        that would hit this one: something like ^ABC$ would have failed after \r
2462        and then tried again after \r\n.
2463    
2464    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2465        in an attempt to make files that differ only in their line terminators
2466        compare equal. This works on Linux.
2467    
2468    17. Under certain error circumstances pcregrep might try to free random memory
2469        as it exited. This is now fixed, thanks to valgrind.
2470    
2471    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2472        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2473        was because its rules for how to advance for /g after matching an empty
2474        string at the end of a line did not allow for this case. They now check for
2475        it specially.
2476    
2477    20. pcretest is supposed to handle patterns and data of any length, by
2478        extending its buffers when necessary. It was getting this wrong when the
2479        buffer for a data line had to be extended.
2480    
2481    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2482        CRLF as a newline sequence.
2483    
2484    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2485        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2486        I have nevertheless tidied it up.
2487    
2488    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2489    
2490    24. Added a man page for pcre-config.
2491    
2492    
2493  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06

Legend:
Removed from v.98  
changed lines
  Added in v.1250

  ViewVC Help
Powered by ViewVC 1.1.5