Re: #385: HTTP2 Upgrade / Negotiation

Hi Takeshi,

On Wed, Oct 24, 2012 at 02:12:44PM +0900, Takeshi Yoshino wrote:
> > > Do you mean the Chromium HTTP upgrade experiment agl referred to in
> > >
> >
> > No, IIRC there was also some broader experimentation using ads; I'll dig
> > around a bit more.
> >
> >
> Maybe this work which
> resulted in introduction of masking to the WebSocket protocol.

Yes I think this is the work Mark was talking about, but warning,
there were some issues with this test due to the inability to analyse
intermediaries behaviour. While very useful, we must keep in mind that
several of the issues detected were assigned a possible cause that
justified masking (eg: connection resets supposed to be caused by
interpretation of payload contents as requests), while a number of
other causes were much more likely true for a number of us implemntors
(eg: classic close upon response receipt). I don't have all the details
in mind anymore, this was discussed to such great lengths on the hybi
list :-)

Today I'd much rather trust real-world tests from websocket clients
running in browsers !


Received on Thursday, 25 October 2012 07:04:35 UTC