Also, the current PDF permissions model is comically obtuse IMO. Acrobat doesn't even expose the permissions bits directly, rather each permission exposed in Adobe's UI represents various permutations of those permissions bits rather than a 1:1 mapping. It has to, because the design of that bitfield is nuts.
The semantics of the bits are described in ways that are very open to interpretation, and can be used to specify combinations of permissions that just don't make sense (example: allow content extraction for copy/paste, but don't allow content extraction for accessibility purposes, like screen readers for the blind). I suspect there are no two PDF parsers that handle that bitfield in the same way, and that it may not even be possible to implement support for it without baking in contradictions and unexpected behaviors (from a user's perspective at least). I'd be surprised if even Adobe's implementation handled it in its entirety.
The semantics of the bits are described in ways that are very open to interpretation, and can be used to specify combinations of permissions that just don't make sense (example: allow content extraction for copy/paste, but don't allow content extraction for accessibility purposes, like screen readers for the blind). I suspect there are no two PDF parsers that handle that bitfield in the same way, and that it may not even be possible to implement support for it without baking in contradictions and unexpected behaviors (from a user's perspective at least). I'd be surprised if even Adobe's implementation handled it in its entirety.