[X2go-dev] VcXsrv
John A. Sullivan III
jsullivan at opensourcedevel.com
Sat Jan 29 21:50:43 CET 2011
On Sat, 2011-01-29 at 21:28 +0100, Mike Gabriel wrote:
> Hi John,
>
> On Sa 29 Jan 2011 21:19:15 CET "John A. Sullivan III" wrote:
>
> > <snip>
> > Just as a reminder, we probably want to move this code to after the
> > session parameters are chosen and allow passing the -fullscreen and
> > -rootless (available area) parameters. Thanks - John
>
> Thanks for the reminder...
>
> In PyHoca-GUI I currently use the Linux XServer or the Windows XServer
> in multi-window mode.
>
> How about the idea to have one global XServer in multi-window mode
> running in the background, that can be used for all
> <width>x<height>ish sessions.
>
> Only if a fullscreen or rootless session is requested then a separate
> XServer instance on a separate port gets started...
>
> Main aspect of this though is performance, esp. on slow machines.
>
> BTW: what would be the equivalent to the Linux x2goclient? Will
> fullscreen sessions than start a new XServer instance on a new tty?
>
> Greets,
> Mike
>
Hi, Mike. Phil and I kicked that around quite a bit. Of course, our
discussion was germane to our environment so you may arrive at a
different conclusion. What you suggest was our initial thought. But
then we realized that the vast majority of users only run a single
session (except, I suppose, for published app users but we only use
desktops). Thus, it seemed like an awful lot of needless bloat to
launch the multi-window X server anyway.
Is there a simple way to detect if there is a current multi-windows X
Server running. If so, we could test for that before launching a new
multi-window session, i.e., we only launch what is requested
(fullscreen, rootless, or multi-window) but, if multi-window is
selected, we check for a running multi-window and, if one exists, we do
not launch another.
The last question is interesting and I really don't know the answer. If
I recall correctly, X2Goclient does not officially support running more
than one simultaneous client. I do it all the time, however - right now
in fact! I do generate errors. Typically the previous running session
will generate an error about failing to connect on port 30001 but
nothing seems to break. Thanks - John
More information about the x2go-dev
mailing list