Re: push_promise sent after frames that reference promised resources

Me too. When a pushed resource arrives ahead of a browser initiated
resource for the same object, send RST_STREAM and use the pushed resource.

Any reason this clarification shouldn't be incorporated into the spec?
Seems an important detail to the server push implementation that we'd want
to make consistent across all browsers. SPDY proxy vendors doing server
push would appreciate it anyway.



On Sun, Nov 3, 2013 at 9:34 AM, Martin Thomson <>wrote:

> On 3 November 2013 09:29, Peter Lepeska <> wrote:
> > Question: if the browser issues a normal HEADERS frame GET for an object
> > that subsequently arrives via a PUSH_PROMISE stream, will it use the
> pushed
> > copy if that pushed copy arrives before any data frames arrive on the
> stream
> > the browser initiated?
> That's the browser's business.  If it were me, I'd try to reduce the
> possibility the number of bytes wasted, so I'd probably send
> RST_STREAM on one or other stream, probably the one I initiated since
> that is likely to be the later one.

Received on Sunday, 3 November 2013 17:41:42 UTC