On Wed, Nov 29, 2006 at 07:43:17PM +0100, Narcis wrote:
> I'm not supprised, it simply terminates the socket. There is a fix in  
> 1.1.0 that allows you to press Q in game and it will send a CP_QUIT.
> the results are different per server. Some put on the self destruct
> sequence, some simply close the connection.

This might be happening because of race condition ... see if you can
capture more data, such as network packets.  A connection reset can
cause loss of that CP_QUIT in the server buffer before it is read by the
server process.  I'm happy to help with that.

> P.S.  I've reached a state where RSA is more or less in the code, and  
> would like to check it on some servers. Shall i post the generated
> MacTrek_Key file here ?

Sure ... I could add it to inl.real-time.com's key list for you to test
with.

> P.P.S  There is also the issue of which features are regarded as  
> "enhancement" and which may be regarded as "borg" [...]

This doesn't worry me, but I'm not the one to approve the key for
distribution to all servers.  At minimum you should have the features
present, but disable them when the server asks for your key?  The best
solution is to define more FEATURE packets in the protocol, default them
to on, and leave it up to the server owners whether they respond to
demands to turn them off.

-- 
James Cameron    mailto:quozl at us.netrek.org     http://quozl.netrek.org/