You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
My current project uses h2 directly without hyper negotiated by TLS with ALPN. To support an AsyncRead+AsyncWrite transport on top of an HTTP2 stream I had to write an adapter like hyper's H2Upgraded, which was non-trivial due to buffering and flow-control capacity management.
Now I wonder if you would consider moving the h2 support types from hyper to h2 and expose them, so users of h2 can build binary protocols on top of h2 and even can take advantage of BDP support. I don't believe this would bloat h2 with code unrelated to HTTP2 since it does support upgrading to stream-based protocols.
I know the chance of this is slim, but I had to ask. :)
The text was updated successfully, but these errors were encountered:
My current project uses h2 directly without hyper negotiated by TLS with ALPN. To support an
AsyncRead
+AsyncWrite
transport on top of an HTTP2 stream I had to write an adapter like hyper'sH2Upgraded
, which was non-trivial due to buffering and flow-control capacity management.Now I wonder if you would consider moving the h2 support types from hyper to h2 and expose them, so users of h2 can build binary protocols on top of h2 and even can take advantage of BDP support. I don't believe this would bloat h2 with code unrelated to HTTP2 since it does support upgrading to stream-based protocols.
I know the chance of this is slim, but I had to ask. :)
The text was updated successfully, but these errors were encountered: