On Sun Aug 14 15:29:26 2022 +0000, Nikolay Sivov wrote:
It would be valuable to test such error cases.
The question is does it really work like that. This will need a basic test with a couple lines of data, to see that destination stride is respected, and padding is not being written over.