i think this should return false
i was trying to run this function with
const EGLint pi32ConfigAttribs =
and it would fail matching a config but the function would return true letting me think that everything went fine.
https://bugs.freedesktop.org/show_bug.cgi?id=50870 <-- might be similar to this
i changed the data to this
pi32ConfigAttribs = EGL_SURFACE_TYPE;
pi32ConfigAttribs = EGL_WINDOW_BIT;
pi32ConfigAttribs = EGL_RENDERABLE_TYPE;
pi32ConfigAttribs = EGL_OPENGL_ES2_BIT;
pi32ConfigAttribs = EGL_NONE;
and it worked
From my understanding of the EGL spec, Mesa behaves correctly.
eglChooseConfig should return false only if an error occured. And, if an error occurred, then there is always an accompanying error emitted by eglGetError. Suppose that the user calls eglChooseConfig, passes it a valid attribute list, no internal EGL errors occur, and the EGL implementation determines that no available config matches the requested attribute list. No error has actually occured. Moreover, there is no EGL error code appropriate for this situation. If no error gets emitted, then there is no error.
If you are checking only the return code of eglChooseConfig, then you use the function incorrectly. If the function succeeds, you must inspect the 'num_config' out-param before reading the 'configs' output. If no configs were returned, and no error occurred, then 'num_config' will be 0. However, you must check the return value of eglChooseConfig for failure *before* checking 'num_config', because 'num_config' remains unmodified when the function fails, according to the spec.
Francois, for your first example attribute list, did eglChooseConfg set 'num_config' to 0? If so, then the bug should be closed as NOTABUG.
-- GitLab Migration Automatic Message --
This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.
You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/148.