EGL_MESA_screen_surface version 5
Brian Paul
brian.paul at tungstengraphics.com
Thu Apr 7 08:52:43 PDT 2005
Adam Jackson wrote:
> On Wednesday 06 April 2005 18:12, Brian Paul wrote:
>
>
>>Dependencies
>>
>> EGL 1.1 or later.
>
>
> Do we really need 1.1 for this over 1.0?
>
>
>> 5. Should the EGL_PHYSICAL_SIZE_EGL query be kept? The size information
>> isn't always reliable (consider video projectors) but can still be
>> used to determine the pixel aspect ratio.
>>
>> Arguments for:
>>
>> X supports a similar query with DisplayWidthMM(), DisplayHeightMM()
>> If this information can be easily queried with EDID, why not
>> make it available to the user?
>>
>> Arguments against:
>>
>> Historically, these values aren't always accurate. Also, they're
>> not always applicable to the display device.
>>
>> Other options:
>>
>> Perhaps just a pixel aspect ratio should be supported. [M. Danzer]
>>
>> Postpone for a future extension, if needed.
>
>
> I agree with doing this as a layered extension (EGL_MESA_extended_screen_info
> perhaps).
>
>
>> 8. Should eglQueryModeStringMESA() be included? This function returns
>> a human-readable string which corresponds to an EGLMode.
>>
>> Arguments for:
>>
>> A mode name such as "HDTV-720P" might mean more to users than
>> "1280x720 at 60Hz" if the later were generated via code.
>
>
> I would push this into the same extension as EDID stuff. I would prefer to
> keep EGL_MESA_screen_surface algorithmic rather than heuristic.
>
>
>> 10. What should be the default size for screen surfaces?
>>
>> For Pbuffer surfaces the default width and height are zero.
>
>
> This should be clarified to say that we'll do the same for screen surfaces.
>
>
>> Attempting to delete a screen surface which is currently being
>> displayed will result in the error XXX??? being generated.
>
>
> EGL_BAD_ACCESS sounds most appropriate.
Thanks, I've incorporated your comments in the next version. I'll
send it out in a day or two.
-Brian
More information about the dri-egl
mailing list