- From: James M Snell <jasnell@gmail.com>
- Date: Mon, 12 Aug 2013 16:01:04 -0700
- To: Jeff Pinner <jpinner@twitter.com>
- Cc: Martin Thomson <martin.thomson@gmail.com>, HTTP Working Group <ietf-http-wg@w3.org>
How so? Intermediary receives a PUSH_PROMISE with END_STREAM, if it chooses not to forward that on, it sends an empty DATA frame with END_STREAM, none of the semantics are changed really. That said, I have no strong opinion about it one way or the other. On Mon, Aug 12, 2013 at 3:58 PM, Jeff Pinner <jpinner@twitter.com> wrote: > PUSH_PROMISE is hop-by-hop -- if you end the stream with a push promise flag > you have to tell intermediates to forward the flag on some other frame, and > that may or may not break the semantics of the layered application. > > > On Mon, Aug 12, 2013 at 3:54 PM, James M Snell <jasnell@gmail.com> wrote: >> >> I see no reason at all why PUSH_PROMISE can't have END_STREAM. >> >> On Mon, Aug 12, 2013 at 3:47 PM, Martin Thomson >> <martin.thomson@gmail.com> wrote: >> > https://github.com/http2/http2-spec/issues/228 >> > >> > PUSH_PROMISE does not have an END_STREAM flag. >> > >> > However, a sender might emit a PUSH_PROMISE followed by a CONTINUATION >> > with the END_STREAM set. As specified, this ends the stream. >> > >> > I think that we need to either prohibit this, or we add END_STREAM to >> > PUSH_PROMISE. I don't care which. Bike shed... go! >> > >> >
Received on Monday, 12 August 2013 23:01:51 UTC