Re: What are the semantics of a client sent GOAWAY?

On Fri, May 24, 2019 at 11:28 AM Martin Thomson <mt@lowentropy.net> wrote:

>
> My view is that we could add a signal to the protocol that allowed a
> client to block further pushes.  However, to call it GOAWAY would be
> misleading.  GOAWAY exists to facilitate graceful shutdown, but that signal
> couldn't be graceful in the same way that a GOAWAY frame from the server
> is.  I don't think that we can fix h2 now, but we can avoid repeating its
> mistakes.
>

On the HTTP/3 ticket that Alan linked [1], I advocated for a signal like
this. That could be a specialised extension frame for server push (e.g.
STOP_PUSHING), which contains an identifier for the maximum value that is
accepted. With the identifier representing a stream ID in H2, or a push ID
in H3. For push, a signal for what-was-or-maybe-will-be processed is
marginally useful for the safety reasons Martin stated.

Lucas

[1] - https://github.com/quicwg/base-drafts/issues/2632