Re: Windows Crashes in regards to PIXELFORMAT.. setup / ATI/AMD driver

Re: Windows Crashes in regards to PIXELFORMAT.. setup / ATI/AMD driver

While searching for a cause of bug 532, I made the magic '5' Windows GDI/WGL
functions 'switchable' - see:
  http://jogamp.org/git/?p=jogl.git;a=commit;h=c49d29784986b1945343b9a90b5e0c9f3d95d937

Whoever has a problem (crash, etc) w/ setting up the PIXELFORMAT.. (create context etc)
may try enabling the 'wgl' variants as described in the git commit.

It would be great if others (working machine w/ JOGL) try this out as well,
since we may switch over to the 'wgl' functions following the GDI/WGL documentation
(dynamically loaded OpenGL DLL). The switch would allows us to also load a non std.
OpenGL DLL (MCD, ..).

~Sven

Hi

I think that this feature might be useful to work around various problems on Intel and Nvidia graphics cards which worked fine with JOGL 1.1.1a but not with JOGL 2.0 (look at peq's case mentionned in another thread). However, it would fine to use it as a fallback solution when attempting to create a new context on Windows rather than asking the programmer to choose whether this flag should be set to true or false.

On Monday, January 09, 2012 10:05:57 AM gouessej [via jogamp] wrote:
>
> Hi
>
> I think that this feature might be useful to work around various problems on
> Intel and Nvidia graphics cards which worked fine with JOGL 1.1.1a but not
> with JOGL 2.0 (look at peq's case mentionned in another thread).

Sorry, but JOGL 1.1.1a used the current default, the GDI variant.

Well, I guess most old ATI driver/card problems are b/c of the new PFD / ARB attributes.

> However, it
> would fine to use it as a fallback solution when attempting to create a new
> context on Windows rather than asking the programmer to choose whether this
> flag should be set to true or false.

IMHO we should validate whether the 'wgl' variant helps anybody here.
It was just a 'desperate' move for the Intel HD 3000 bug 532
but before trashing this switch - I thought allow others to try if it helps w/ their problems.

~Sven

Ok. It means the root cause of some regressions in context creation on a few hardware comes from something else, sorry.