Is there a usecase for vectored I/O?
We currently don't provide NIO's interface to readv/writev (aka vectored io, aka scattered reading and gathered writes).
Is there a use case for it?
#1 Updated by hans over 5 years ago
I'm sure there is, but since we haven't needed it yet, I think its low priority. We haven't entirely proven IOCipher yet, so I think its important that we first get one approach working really well. Once that's done, we can say that this whole libsqlfs approach actually works. Then we can implement all of the various other parts of this API knowing that the work won't be thrown away.