Yes. But I think having a return value would be more intuitive.
Hmm, I think it duplicates the information and actually changes the logic a bit here: if neither current_state.rect nor rects.visible are fullscreen, the memcmp was supposed to return 0 before [^1]. ie: both rects are similar fullscreen-wise, both not fullscreen.
In practice it doesn't make a difference as because of data->is_fullscreen, one of the two rects has to be fullscreen, so I don't really mind but still IMO unnecessary.
[^1]: Ignoring the issue this MR is fixing.