[SPDY] Add a SpdyStreamType enum and use it for SpdyStream

[SPDY] Add a SpdyStreamType enum and use it for SpdyStream

Use SpdyStreamType to enforce stream and stream delegate semantics
instead of trying to infer it from various parameters and return values.

Remove now-redundant return value of
SpdyStream::Delegate::OnSendRequestHeadersComplete().

Comment expected semantics of SpdyStream::Delegate member
functions and related SpdyStream member functions.

Enforce some restrictions on push streams. Namely, set send_status_
to NO_MORE_DATA_TO_SEND immediately for them, and make sure
its entry point is SpdyStream::OnResponseReceived().

Also clarify how push streams interact with the stream histograms.

BUG=243643
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=202401

[akalin, 2013-05-25 07:40:43]

+rch for review

https://codereview.chromium.org/15701009/diff/1/net/spdy/spdy_stream.cc
File net/spdy/spdy_stream.cc (left):

https://codereview.chromium.org/15701009/diff/1/net/spdy/spdy_stream.cc#oldco...
net/spdy/spdy_stream.cc:723: if (!session_->NeedsCredentials() || pushed_ ||
!url.SchemeIs("https")) {
Note that I removed the pushed_ check here

[Ryan Hamilton, 2013-05-26 15:33:49]

One question.  LGTM otherwise.

https://codereview.chromium.org/15701009/diff/1022/net/spdy/spdy_http_stream.cc
File net/spdy/spdy_http_stream.cc (right):

https://codereview.chromium.org/15701009/diff/1022/net/spdy/spdy_http_stream....
net/spdy/spdy_http_stream.cc:69: int error =
spdy_session_->GetPushStream(request_info_->url, &stream_,
Oh, I see.  it does know that it's using a push stream. Cool.

https://codereview.chromium.org/15701009/diff/1022/net/spdy/spdy_stream.cc
File net/spdy/spdy_stream.cc (right):

https://codereview.chromium.org/15701009/diff/1022/net/spdy/spdy_stream.cc#ne...
net/spdy/spdy_stream.cc:568: CHECK_NE(type_, SPDY_PUSH_STREAM);
I don't think this is right, is it?  When a request is made for a URL which can
be handled by a pushed stream, the SpdyHttpStream does not know that, does it? 
I thought it would actually call SendRequestHeaders.

EDIT: Ah, but actually looking at SpdyHttpStream shows that I'm wrong :>

https://codereview.chromium.org/15701009/diff/1022/net/spdy/spdy_stream.h
File net/spdy/spdy_stream.h (right):

https://codereview.chromium.org/15701009/diff/1022/net/spdy/spdy_stream.h#new...
net/spdy/spdy_stream.h:318: // bidirectional streams; for request/response
streams, it must be
What if a bidirectional stream knew (because of some higher level protocol) that
this was in-fact the last data it would send.  Could it use
NO_MORE_DATA_TO_SEND?

[akalin, 2013-05-27 08:13:35]

Committing soon

https://codereview.chromium.org/15701009/diff/1022/net/spdy/spdy_stream.h
File net/spdy/spdy_stream.h (right):

https://codereview.chromium.org/15701009/diff/1022/net/spdy/spdy_stream.h#new...
net/spdy/spdy_stream.h:318: // bidirectional streams; for request/response
streams, it must be
On 2013/05/26 15:33:49, Ryan Hamilton wrote:
> What if a bidirectional stream knew (because of some higher level protocol)
that
> this was in-fact the last data it would send.  Could it use
> NO_MORE_DATA_TO_SEND?

Yeah, I think we eventually want to allow that. For now, I'm just matching the
current behavior as closely as possible.

[commit-bot: I haz the power, 2013-05-27 08:13:59]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/akalin@chromium.org/15701009/1022

[commit-bot: I haz the power, 2013-05-27 12:53:10]

Change committed as 202401