[RFC] Is that error mapping correct?

[DJ Delorie]

> >	$ cat //foo/bar/baz
> >	cat: //foo/bar/baz: No such device or address
> >
> >seems somewhat strange. Have I missed something, perhaps?
> 
> I don't know.  Is there some UNIX equivalent that we could use for
> testing?  It is nice to differentiate between a non-existent file and a
> non-existent host but I agree that the current error doesn't make much
> sense.

I unplugged my SGI's ethernet cable and tried to access a file on it
from Linux (via nfs) and got EIO - "Input/output error".  I'm not sure
if this is the same situation, though.