Re: the nvidia/osmesa dilema

[Juan Pablo Carbajal]

On Wed, May 4, 2016 at 5:49 PM, Mike Miller <address@hidden> wrote:
> On Wed, May 04, 2016 at 10:43:34 -0500, Daniel J Sebald wrote:
>> On 05/04/2016 05:36 AM, Juan Pablo Carbajal wrote:
>> > hi all,
>> >
>> > We still have this issue with building the documentation and running
>> > tests (https://savannah.gnu.org/bugs/?44478) that casues segfaults
>> > when the user is running with nvidia drivers.
>> >
>> > The documentation issue, I assume, comes form the attempt of printing
>> > images to files without rendering on the screen. Would it be possible
>> > to give a flag on the make file to disable off-screen printing? Of
>> > course, if possible i would try to do it and submit the patch.
>>
>> How about rather than just disable off-screen printing with a flag, make it
>> a reconfigurability flag, e.g., --with-gnuplot-offscreen.  Or something that
>> will make Octave use a different graphics toolkit.  I tried a couple nights
>> ago to place an .octaverc file that changes the graphics toolkit in the
>> various build directories where octave-run is launched...didn't work.
>
> If you take a look at the functions that print the figures for the
> manual, there is already code in there to handle the case when Octave is
> built without OSMesa support.
>
> If you have the proprietary nvidia drivers and you know that OSMesa
> doesn't work, just try building Octave with "--without-OSMesa" and see
> what happens.
>
Ok, but what exactly you mean by OSMesa doesn't work. Everything works
except off-screen printing with __osmesa_print__
Wouldn't this disable other features of the graphics support that work ok?
I am compiling right now with this flag. Lets see what comes out.

> I think patches would be welcome to make the configure script
> automatically disable OSMesa if we can detect this situation somehow.
>
>> That, or fix the issue with nVidia off-screen printing.
>
> True.
>
> --
> mike