Post-Quantum Cryptography #15
Conversation
robin-nitrokey
left a comment
robin-nitrokey
left a comment
There was a problem hiding this comment.
Generally, I would prefer to make Dispatch generic over the message size (maybe with a type alias using the old default of 7609 bytes) so that this crate does not need to be aware of the message size required by specific algorithms. As we have already made the App trait generic over the message size and extracted it into a separate crate, only the runner, trussed-usbip and usbd-ctaphid would be affected by this change. So while it would add some complexity, the scope would be very limited. @sosthene-nitrokey What do you think?
|
I was thinking of the same thing. I like the idea of keeping an alias to avoid having the value show up in so many places. |
|
@robin-nitrokey @sosthene-nitrokey I'm all for making it generic and removing the need for this change, but this sounds like a structural change in |
|
Yeah, we should take that in a separate PR. It should not be a big change. |
The fixed buffer size can be problematic for some use cases, see: - #15 To avoid adding support for different buffer sizes to the crate, we can just make the dispatch implementation generic over the buffer size. For simple cases, a type alias is provided using the old buffer size.
The fixed buffer size can be problematic for some use cases, see: - #15 To avoid adding support for different buffer sizes to the crate, we can just make the dispatch implementation generic over the buffer size. For simple cases, a type alias is provided using the old buffer size.
|
Replaced by: #16 |
3 participants
No description provided.