Thought I'd give a short heads up on what we we're currently focusing on with regards the code. As said earlier one of the top things we want to do is integrate the ssh code in the client and stop using the ugly bash scripts we're currently using.
Our original thought has been using libssh and writing our own ssh server and client. However while we weigh the benefits of having our own in-code server and client for now we've decided to stick to using the OpenSSH client and server but rerouting stdin,stdout and stderr so our client has perfect control over the process.
So this is what we're each looking into right now...
Jeetu - Develop the generic routines to pipe and reroute stdin,stdout and stderr for a program (could be sshd,ssh,fuse,etc) and send/receive/process data on these pipes.
Erle - Client code review, figure what parts of the code we will eventually like to improve. Higher level understanding of how the entire client will function, specifically but not restricted to the new SSH code.
Thoughts we need to ponder:
-
Do we want sshd to startup by default when client runs? Currently our sshd script kicks in when a user requests a software should be run.
-
What error conditions for the client and server need to be caught and handled by our code.
-
How are the keys going to be generated? How are these stored?
- Should eBrainPool run as it's own user id?