[Spice-devel] Current status of XSpice driver

[John A. Sullivan III]

On Sat, 2011-06-25 at 20:38 +0200, Alon Levy wrote:
> On Sat, Jun 25, 2011 at 06:01:27PM +0200, Sebastian Hesselbarth wrote:
> > Hi,
> > 
> > I'd like to ask about the current and future status of the XSpice driver for
> > Xorg. I am very interested in virtualizing desktops for a bunch of users without
> It is in active development, no stable release yet. It's meant to allow using spice
> clients to connect to an X server directly, never mind if it is running in a vm or not.
> It reuses the X driver, so it has the same features (and lacks the same features).
> 
> > the need of a full-blown virtual machine. I once compiled the driver from git
> > but without luck of connecting a client to it (errors were about server and
> > client disagree about modes AFAIR). 
> > 
> If you used master branch please pull and try again. I have no such problems, so
> I'd like to know why it isn't working for you. I'm using master spice+spice-protocl
> and just standard X clients. I am compiling against X master, so perhaps that's the
> problem. (or maybe I neglected to pull something and I'm actually using an old version
> of one of the protos).
> 
> > Can someone please clarify the requirements for compiling the current XSpice 
> > driver. I really like to help improve the driver with debugging, coding, and
> > suggestions.
> I mainly test it with everything needed built from git. Also, I've just updated
> master (forced), but generally the most uptodate right now is the xspice.vN (N=5 atm).
> 
> Everything means xserver and other dependencies needed at runtime (xkbcomp etc.).
> I haven't documented everything yet, but see README.xspice:
>  http://cgit.freedesktop.org/~alon/xspice/tree/README.xspice
>  <snip>
Hmm . . . looking at that README again, can one have SPICE using
multiple displays? I'm guessing not but I'm thinking ahead to X2Go as a
wrapper.  We would connect from the X2Go client to the server to spin up
the instance on the next available display according to the X2Go
database.  Then that user would connect to say, display :3.  The next
user might spin a new one on :4 and connect to it.  However, I'm
guessing that unlike NX, SPICE is simply looking at the connecting port
and would have not idea how to distinguish between the packets from the
user on :3 and the user on :4.  Is that correct? Thanks - John