OnStreamEnd is now called instead of the sentinel call of OnStreamFrameData(stream_id, nullptr, 0, …

OnStreamEnd is now called instead of the sentinel call of OnStreamFrameData(stream_id, nullptr, 0, true). Protected by the flag FLAGS_spdy_on_stream_end.

Merge internal change: 116272960

BUG=

[rjshade, 2016-03-10 22:24:01]

Description was changed from

==========
OnStreamEnd is now called instead of the sentinel call of
OnStreamFrameData(stream_id, nullptr, 0, true). Protected by the flag
FLAGS_spdy_on_stream_end.

Merge internal change: 116272960

BUG=
==========

to

==========
OnStreamEnd is now called instead of the sentinel call of
OnStreamFrameData(stream_id, nullptr, 0, true). Protected by the flag
FLAGS_spdy_on_stream_end.

Merge internal change: 116272960

BUG=
==========

[rjshade, 2016-03-10 22:24:02]

rjshade@chromium.org changed reviewers:
+ bnc@chromium.org, rch@chromium.org

[Ryan Hamilton, 2016-03-10 23:21:14]

https://codereview.chromium.org/1777163003/diff/20001/net/quic/quic_headers_s...
File net/quic/quic_headers_stream.cc (right):

https://codereview.chromium.org/1777163003/diff/20001/net/quic/quic_headers_s...
net/quic/quic_headers_stream.cc:70: // or SYN_REPLY frame that had the fin bit
set.
Hm. I think we actually need to do something here. If you look at
QuicHeadersStream::OnControlFrameHeaderData, it looks to me like that code
expects len == 0. Do the tests pass with this CL (and flag = true)?