Skip to main content

in reply to daniel:// stenberg://

I wonder if the Linux kernel server-side QUIC implementation will be available before the OpenSSL one 🤔
in reply to Remi Gacogne

@rgacogne Good point. This kernel side effort seems to have a lot of activity and progress compared to the other one...
in reply to daniel:// stenberg://

I could see this coming a mile away. Quite frankly, this is much more sane than using any userland QUIC library...
(which might be a reason why my involvement in a certain implementation has been ... shall we say lack luster?)
in reply to Richard Levitte

@levitte I agree that it seems like a good idea and certainly will make it simpler to implement things using QUIC.

I suppose it now depends on how willing the kernel people are to accept such a thing and how good the actual patch and related work is (I'm not really able to judge)

in reply to daniel:// stenberg://

@levitte I’m not into linux kernal/networking but I thought one of the advantages of quic/http3 is, that it can be implemented in userland?
in reply to daniel:// stenberg://

This is going to be in Rust, right? We shouldn't be writing new network protocol implementations in C unless we really can't help it.
(In some sense this is a shitpost, but the point is quite serious...)