IIRC, the overhead was fairly small but slightly noticeable. I didn't do a rigorous analysis, though. I do know that if we make that the default we should set an upper limit on number of rotations.
We can revisit this decision with data, obv., and 2.0.0 might be the time to make this change. On Mon, Jan 11, 2021 at 9:20 AM Peter Kronenberg <[email protected]> wrote: > What is the overhead of having ApplyRotation set to true, even if most of > your pages don’t need rotation correction? I’m wondering if there is a use > case for not having it set to true. I assume the overhead of actually > correcting the rotation is higher than just calculating the rotation. >
