1 |
Current status of the Malete engine. |
2 |
|
3 |
|
4 |
The |
5 |
> DownLoad 0.9.9.e |
6 |
version is considered the first release candidate for the proposed 1.0 release. |
7 |
|
8 |
|
9 |
* what 1.0 contains |
10 |
|
11 |
- malete server |
12 |
communicating over TCP or UNIX sockets as described in the |
13 |
> Protocol |
14 |
- multi process server option |
15 |
(on systems with working fork and fcntl, i.e. not on windows) |
16 |
- search expressions |
17 |
and a single '?' to retrieve record contents as described in the |
18 |
> Query query spec |
19 |
- sample client interface |
20 |
in PHP and possibly other languages |
21 |
- conversion tools |
22 |
> CmdLine from and to |
23 |
traditional CDS/ISIS, IIF (ISO 2709) and plaintext |
24 |
- stable small file formats as |
25 |
> FileFormats specified |
26 |
(32 bit file access using 8 byte pointers) |
27 |
- collation support |
28 |
is working as specified in |
29 |
> CharSet character set support |
30 |
|
31 |
* what 1.0 does not contain |
32 |
|
33 |
(at least not officially) |
34 |
- protocol features marked as optional extensions |
35 |
- record filtering |
36 |
since that is rarely used and can be done at application level |
37 |
- large files and key compression |
38 |
- any promise on the C API |
39 |
(it IS going to change) |
40 |
- the Malete object model |
41 |
definitively needs some more tweaking before being ready for prime time |
42 |
|
43 |
* what is still TO DO |
44 |
|
45 |
- de/serializing of records to/from files accessed in a stream like fashion |
46 |
needs to be debugged, unified and heavily tested. |
47 |
The concept of a sink probably needs to be straightened out. |
48 |
- the single process server should do proper session switching. |
49 |
Every session should have a separate input buffer. |
50 |
On the other hand, we may also dump single process completely. |
51 |
- both single process and multi process server need heavy testing |
52 |
|
53 |
|
54 |
--- |
55 |
$Id: Status.txt,v 1.3 2004/11/12 11:18:23 kripke Exp $ |