Luke Mewburn wrote:

>On Thu, Dec 09, 2004 at 01:17:36AM -0500, Dave wrote:
>  | Here is a compromise/workaround I use for ssh-agent access while
>  | under screen, and it satisfies me (but YMMV):
>  | 
>  |   [...]
>  | 
>  | This approach buys you extra flexibility over the vanilla ssh-agent/
>  | subshell approach: it lets you gain access to your agent from any
>  | shell, even one that had already been opened in a separate window
>  | before the agent was ever started (without too much typing).
>
>Or you could just fire up the agent the first time you connect to
>your system, run a screen, and do all your work in windows on that
>screen, using multi-attach screen (screen -x) to give you multiple
>parallel views into the same "workspace".
>
>I describe this elsewhere in this thread.
>  
>
Yes, this I actually thought of.  However, much of my screen useage is 
actually the classical "log into remote machine, run screen, reconnect 
from different machines which may or may not stay on the network".  So 
while I could do this, I'd have to keep that master screen attached with 
the agent running.  Your other post solves my problem and is so 
beautiful I could cry.. *sniff*.

    Ian