[Ovirt-devel] oVirt console (again!)

[Hugh O. Brock]

On Thu, Nov 06, 2008 at 05:01:18PM +0000, Richard W.M. Jones wrote:
> Previous discussion:
> https://www.redhat.com/archives/ovirt-devel/2008-August/msg00004.html
> 
> Back to the oVirt console issue.
> 
> So far we've look at:
> 
>  (1) A Gtk-based VNC browser plugin.  Unfortunately we couldn't
>      make this stable enough for real production use.
> 
>  (2) Launching an external program such as virt-viewer from a browser
>      plugin.  This works, but it's a security issue, and we can't use
>      a Gtk dialog to get around the warning issue because of (1).
> 
>  (3) Running virt-viewer or viagre as separate, standalone programs.
>      This works, but requires the user to type in some very long
>      and complicated command line by hand, and there are unresolved
>      authentication problems.
> 
> We are now in a state where we have a working virt-viewer binary for
> Windows.  Vinagre is not far off.  And we know pretty well how to
> write C / Gtk programs which cross-compile and run under Windows
> reliably.
> 
> On the call today we discussed:
> 
>  (4) Write a custom C/Gtk/Gtk-VNC Windows program which contacts the
>      oVirt WUI to do authentication, get available consoles, and
>      launch a Gtk-VNC widget with the appropriate tunnelling
>      (openssh.exe based?).
> 
> Thoughts?
> 

Well, we could certainly contact the REST API on the server securely
to get that information.

Would the Windows program be launched by our browser plugin, or be an
entirely separate program?

--H