Seems like there's an optimization to not perform a bunch of gl operations when the new layer has the same dimensions as the existing buffer size. But the optimization is never run because we're always comparing points to physical pixels.

https://github.com/flutter/engine/blob/master/shell/platform/darwin/ios/ios_gl_context.mm#L170