To get around the issue with IDA not wanting to replace a db's bytes with entirely different ones, we swap out IDA's internal file pointer for the image.
Perhaps by replacing it with something backed with BytesIO, this could be mitigated.
perhaps also IDAs internal representation could be swapped out with a memory mapped object handled as ACCESS_COPY and then written over during decryption?
To get around the issue with IDA not wanting to replace a db's bytes with entirely different ones, we swap out IDA's internal file pointer for the image.
Perhaps by replacing it with something backed with BytesIO, this could be mitigated.