Nathaniel Smith wrote:
> On Tue, Nov 10, 2009 at 11:29 AM, Antoine Martin <[email protected]> 
> wrote:
>> On a related note, I've added some timeouts to the connects:
>> http://xpra.devloop.org.uk/patches/v0.0.8pre/socket-timeout.patch
>>
>> The values may need adjusting but if the local socket is not answering
>> within 5 seconds, chances are it never will.
> 
> I'm curious how important this is in practice... the system already
> has some built in timeouts on these things; do we know better than it
> what the appropriate timeout values should be? Did you run into this,
> or is it a just-in-case kind of thing?
It happens every now and then. The doc says that the default timeout for
sockets is unset, so it will wait forever.

> (This is more relevant for the TCP timeout than the unix domain one --
> I wouldn't want to set the timeout too low and make someone who
> already has congested network's day even *worse*.)
Absolutely. Since it's the exception rather than the rule, maybe 1
minute? If the network is more congested than that, your chances of
using xpra over it are pretty slim!

Antoine

_______________________________________________
Parti-discuss mailing list
[email protected]
http://lists.partiwm.org/cgi-bin/mailman/listinfo/parti-discuss

Reply via email to