On Wed, Sep 06, 2006 at 07:01:25PM +0100, Keir Fraser wrote:
> On 6/9/06 18:22, "Muli Ben-Yehuda" <muli@xxxxxxxxxx> wrote:
>
> > - 'noecho' mode, where we discard stuff that we read back from a given
> > fd after we've just written it there ourselves. Without this you get
> > every console command echo'd back to you from the pty
> > immediately. For what it's worth this happens without the socket
> > support as well, except there we put stdout into raw mode so that
> > the terminal does the 'echo cancelation'. I didn't find a way to
> > have the pty not echo everything back at me and I didn't want to
> > rely on the remote tty being in 'raw' mode, so I implemented it
> > myself.
>
> Why is this a problem? In cases where an echo is really not wanted, like in
> a gdb control session, the remote end would not be echoing in the first
> place.
Hmm, are you saying it's not the pty itself that echo's stuff back at
us, it's whatever is on the other side? (the console daemon or the
domain?).
In any case, I have to admit it hasn't bitten us in practice when
using gdb; either gdb is smart enough to handle this on its own or I'm
missing something. However, supporting --noecho is still nicer from an
aesthetic point of view: compare a console session with and without
--noecho and using telnet on the remote side - without it you get
everything you write echo'd back at you.
Cheers,
Muli
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
Home