ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/BDB/Changes
Revision: 1.50
Committed: Tue Jul 29 03:33:16 2008 UTC (15 years, 9 months ago) by root
Branch: MAIN
CVS Tags: rel-1_71
Changes since 1.49: +3 -0 lines
Log Message:
1.71

File Contents

# Content
1 Revision history for perl module BDB.
2
3 1.71 Tue Jul 29 05:30:24 CEST 2008
4 - add const char * to typemap for possible 5.6 support.
5 - untested support for db4.3.
6 - berkelydb needs seriously larger stacksizes on x86 at least
7 then what the default xthread.h provides.
8
9 1.7 Thu Jul 17 14:09:25 CEST 2008
10 - use a totally different mechanism to pass callbacks to
11 requests (see manpage).
12 - optional arguments before the callback can now be left out.
13 - use less memory per request, probably be a bit faster, too.
14 - fix a memory leak (callbacks were never freed).
15 - add a basic testsuite.
16 - add (untested) support for cdsgroup_begin and db_exists.
17 - make BDB::VERSION compatible to UNIVERSAL::VERSION.
18
19 1.6 Tue Jul 8 10:34:58 CEST 2008
20 - add new BDB::VERSION function for easy version checks.
21 - work around broken bsd headers once more.
22 - support for berkeley db 4.7 added.
23 - add support for log_set_config and it's constants,
24 and set_intermediate_dir_mode.
25 - add db_env_dbrename and db_env_dbremove.
26
27 1.5 Sat May 10 22:22:42 CEST 2008
28 - port to solaris perls configured for old posix
29 (analysed by Jost Krieger).
30 - do our own stack memory management for threads - linux
31 allocates outrageous amounts of VM (not so bad unless you
32 use mlockall...), which severely limits the number of threads
33 on 32-bit arches: stack size is the larger of PTHREAD_STACK_MIN
34 and 4096 * sizeof (long) (usually this is 16..64k).
35 - support AnyEvent::BDB.
36
37 1.45 Sun Mar 30 10:01:00 CEST 2008
38 - work around win32 perl's broken filehandles
39 (this cannot work in all cases, so watch out).
40
41 1.44 Sun Mar 30 08:28:11 CEST 2008
42 - encode filenames as utf-8 (for bdb/win32)
43 or binary (everywhere else).
44 - patch $! to support stringification of BDB errno values
45 (utter hack, this should be pluggable :).
46 - fix prototypes of (void) functions.
47
48 1.43 Sun Jan 13 11:52:49 CET 2008
49 - differentiate between nuked and undef objects, so as to
50 not throw an exception when destroying closed handles.
51
52 1.42 Tue Dec 25 15:22:50 CET 2007
53 - do not automatically call set_thread_count.
54
55 1.41 Sat Dec 22 08:32:41 CET 2007
56 - add cursor->set_priority.
57 - minor optimisations and checks.
58 - port to microsoft pseudo-C again.
59
60 1.4 Wed Dec 12 02:20:39 CET 2007
61 - new globals BDB::VERSION and ..VERSION_STRING.
62 - add db_upgrade.
63 - do not die in DESTROY after explicit db_close and similar
64 operations.
65
66 1.3 Fri Dec 7 14:36:25 CET 2007
67 - db_put and c_put leaked memory like hell (reported by Danny).
68 - de-IO::AIO-ify the documentation.
69 - port to microsofts pseudo-C from hell.
70 - added mutex_set_increment, mutex_set_tas_spins
71 and mutex_set_max, mutex_set_align.
72
73 1.2 Tue Dec 4 11:12:47 CET 2007
74 - after fork, invest some work to make sure that the poll_fileno
75 stays stable (by dup2'ing the new fd over the old one), to make
76 it easier for programs/libs that don't control their forking
77 behaviour and cannot recreate their watchers.
78 - added missing DB::RDONLY (reported by Danny).
79
80 1.1 Sun Sep 16 20:15:55 CEST 2007
81 - flag and check TXN_DEADLOCK flag for easier transaction
82 programming (see db_txn_finish).
83 - new custom db_txn_finish request.
84 - improved docs.
85
86 1.0 Mon Aug 13 14:02:42 CEST 2007
87 - some updates for db 4.6.
88 - vastly improved documentation, i.e. there is some now.
89 - this module has been tested a lot in the Crossfire TRT server.
90
91 0.5 Sun Jul 8 16:57:03 CEST 2007
92 - port to db 4.4.20 (probably any 4.4 version).
93 - added bdb error constants.
94 - add set_errfile/set_msgfile/set_verbose.
95 - shared code between IO::AIO and BDB.
96 - upgrade to xthread.h from IO::AIO for win32 compatibility.
97 - port to win32 (there is an unexplainable hang in
98 db_sequence_get...).
99
100 0.1 Mon Mar 5 20:44:18 CET 2007
101 - cloned from IO::AIO.
102