[Paraview-developers] IceT? strangeness

[Utkarsh Ayachit]

Ah, yes. This is independent of IceT. This is due to the fact that
that the code that paste's back the server image on client simply uses
a texture and render the image on a quad. With remote desktop, the
OpenGL drivers used are pretty crappy and don' support
non-power-of-two-textures and hence the issues.

I can see if I can fix this issue in time for 3.12.

Utkarsh

On Tue, Aug 16, 2011 at 2:53 PM, Moreland, Kenneth <kmorel at sandia.gov> wrote:
>
> Utkarsh,
> This sounds similar to the remote desktop issues Alan was having.  I remember it having something to do with the remote desktop software, but I can't remember if it had anything to do with remote rendering.  Do you remember what the problem was or how (if) it was resolved.
> -Ken
>    ****      Kenneth Moreland
>     ***      Sandia National Laboratories
> ***********
> *** *** ***  email: kmorel at sandia.gov
> **  ***  **  phone: (505) 844-8919
>     ***      web:   http://www.cs.unm.edu/~kmorel
>
> From: "Biddiscombe, John A." <biddisco at cscs.ch>
> Date: Tue, 16 Aug 2011 14:51:05 +0000
> To: Kenneth Moreland <kmorel at sandia.gov>
> Cc: ParaView Developers <paraview-developers at paraview.org>
> Subject: IceT? strangeness
>
> Ken (or other interested party)
>
>
>
> I have a strange error which I think is IceT related. It started a few months back when some fixes to IceT went into the repo. I didn’t bother about it much as it’s an odd use-case, but today I noticed something I didn’t see before and so was thought I’d ask about it.
>
>
>
> Everything works ok, but when I run paraview client/server in remote desktop mode on windows (home to office, or remote to laptop – client and server on the same remote desktop) – the main paraview window always displays a grey mush instead of the image being generated by the pvserver.
>
> In this screenshot, you can see one pvserver is rendering – the image looks fine – but the image in the pv-gui is grey. Notice how the axes widget/marker overlay is rendered correctly ‘on top’ of the iceT copied image.
>
>
>
> Any idea if this can be fixed (or maybe it’s not worth fixing)? I use remote desktop a lot, so I’m used to working with invisible data now – mostly I only need to know if things run without crashing, but today I made the server window visible (by accident) and noticed the image is actually correct (I always assumed it was grey too), it’s just the copy to gui that goes grey. An openGL pixelBuffer copy issue?
>
>
>
> Anyway. Posted in the interest of discovering something new.
>
>
>
> ttfn
>
>
>
> JB
>
>
>
>