[vbox-dev] vrde with vnc
klaus.espenlaub at oracle.com
Wed Mar 7 06:28:28 PST 2012
On 08.11.2011 22:22, Christophe Devriese wrote:
> I tried getting this integrated. It's not working. But here's something
> you can build as an extpack, and it contains docs. If anybody really
> knows how to move forward with this (maybe have an "unofficial extpacks"
> ftp location or something, where oracle doesn't claim ownership or
> quality or anything ???)
It's in the VBox tree now... finally found some time. A few words of
documentation are in the manual, too.
I hope that the deb/rpm package build logic is now also working, this
was quite a bit of work to figure out how it should be done properly.
For those builds the VNC "extpack" is included in the normal package, as
anything else makes little sense.
> Additionally, it would be *great* to have a way to get custom QT
> interfaces and configuration file formats into extpacks. There was some
> work on this but, and it was implemented for the UDPTunnel stuff, but ...
This shouldn't need custom hacking - the listen address/port works, and
the VNCPassword stuff shouldn't be too hard either, if we get to adding
generic VRDE property editing support in the GUI.
Of course a few changes are necessary, to get over the "RDP" centric
> Anyway, here's the file. It shows how to build and register the extpack,
> and how to get the extpack to operate (I'm afraid you will be using
> python commands to enable this for a specific VM instance).
> Also, I'm having trouble finding the time to work on this, and I don't
> really have a place to put this online. If anybody's willing to do
> something like this ? I'd rather not put it on github or something like
> that, as it'll just get fragmented and we'll end up with 10 different
> versions, all with weird bugs.
I always planned to give it a home in the VBox source tree. No promises
if we'll make binaries available or if we leave this to others. The most
likely variant is the latter.
> Anyway, let me know what you think. Good luck with it.
The contribution was a great starting point. Needed some cleanup, and I
couldn't resist implementing some optional features, just to illustrate
Oh, and I had a go at the keyboard handling (Caps/Scroll/Num Lock not
handled, AltGr not handled, and keypad support was a commented out
theory). It's of course still unusable (or at least bad) on non-US
keyboards, because in general reverse mapping from X keysyms to
scancodes is impossible. For simple keyboard layouts it's doable, but
even there knowledge of the keyboard map is required, and this isn't
there at all.
Thank you very much for the effort you all put into this, and if you (or
anyone else) has improvements, just let us know. Now that the base code
is in the tree it's a lot easier to integrate improvements.
> On Mon, Nov 7, 2011 at 8:41 AM, Klaus Espenlaub
> <klaus.espenlaub at oracle.com <mailto:klaus.espenlaub at oracle.com>> wrote:
> On 05.11.2011 00:45, Geoff Nordli wrote:
> On Friday, November 04, 2011 06:24:17 PM Perry Halbert wrote:
> I guess I am confused. You can compile VBox with the VNC
> (VBOX_WITH_VNC :=1) already.
> Not sure about Windows though. Is that what this is about?
> On 11/04/2011 05:41 PM, Geoff Nordli wrote:
> Howard Su was talking about VNC and VRDE back in February.
> Has there been any more work done with this?
> This hasn't been fully integrated yet. I agree that this would
> improve the VNC user experience significantly.
> Hi Perry.
> Being able to control it via the API is the main thing I am
> looking for.
> There is a chance of having this in 4.1, if an engineer finds enough
> time and the change including the cleanup is low risk. In any case,
> it will first go into trunk.
More information about the vbox-dev