You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fixes land in one that don't necessarily make it to the other.
It might be nice if we try to merge this into one codebase that just treats OpenGL and GLES as some sort of GL that has different sets of extensions.
A side benefit of this would be restoring GLES1 support; we couldn't do this before because logical presentation required a render target, but that requirement is gone now. It would just be a sort of GL that doesn't have many extensions at all.
The renderer would listen for "opengl" "opengles2" and "opengles1" as renderer names, and use that name to decide how to set up the GL context.
In practice, this might be a terrible idea, but it could be worth exploring.
The text was updated successfully, but these errors were encountered:
It seems worth exploring. You might have functions that are heavily bifurcated by available extensions and symbols, but still might be cleaner than the current setup.
A possible approach would be to have two renderers, one for newer OpenGL with shaders (including GLES2), and one for older OpenGL with fixed function APIs (including GLES1). It may also be relatively easy to share code between the two, such as for creating and uploading for textures.
Possibly dumb thought that came to mind:
It might be nice if we try to merge this into one codebase that just treats OpenGL and GLES as some sort of GL that has different sets of extensions.
A side benefit of this would be restoring GLES1 support; we couldn't do this before because logical presentation required a render target, but that requirement is gone now. It would just be a sort of GL that doesn't have many extensions at all.
The renderer would listen for "opengl" "opengles2" and "opengles1" as renderer names, and use that name to decide how to set up the GL context.
In practice, this might be a terrible idea, but it could be worth exploring.
The text was updated successfully, but these errors were encountered: