[slime-devel] Re: What happens when net connection broken

[Roman Budzianowski]

On Jul 29, 2008, at 4:14 AM, Helmut Eller wrote:

> * Roman Budzianowski [2008-07-29 00:24+0200] writes:
>
>> What is the currently implemented behavior of the remote stream in
>> swank when the net connection breaks?
>
> In theory, all swank threads should be killed and all swank streams
> should be closed.  It's not easy to handle errors across multiple
> threads, so I don't know if the current implementation actually does
> what it should.  We had many bugs in this area, and I would be  
> surprised
> if we fixed them all.
>
>> I started a long running function and at some point the network went
>> down. The remote lisp is still running and I managed to
>> reconnect through tty because it was running in "screen". I am trying
>> to restart the swank server. I am just wondering
>> if swank thread is still running or is blocked on output. Is there a
>> way to reconnect, like in 'screen. This is in allegro CL.
>
> You could try swank::restart-server and after that M-x slime-connect.

If all swank threads are killed, that means that whatever calculation  
was running is killed
too, right? So, reconnecting is pointless.

I was hoping for a functionality like "screen", where even in case of  
a loss of connection, the
virtual tty stays up and one can reconnect to the still functioning tty.

In other words, if I make a function call in repl, which may run for  
10 hours, I don't want to
lose my calculation in the case of network outage.


Roman

>
>
> Helmut.
>
> _______________________________________________
> slime-devel site list
> slime-devel at common-lisp.net
> http://common-lisp.net/mailman/listinfo/slime-devel