[SPDY] Remove SpdyStream::Delegate::OnSendBody()'s return value

net/spdy/spdy_proxy_client_socket.cc

Index: net/spdy/spdy_proxy_client_socket.cc
diff --git a/net/spdy/spdy_proxy_client_socket.cc b/net/spdy/spdy_proxy_client_socket.cc
index 2fcb899c9ca05c010db2f858880f12de459f20cb..5bc617cb47ff8f08aebb8cfb53349322084b5ff8 100644
--- a/net/spdy/spdy_proxy_client_socket.cc
+++ b/net/spdy/spdy_proxy_client_socket.cc
@@ -468,18 +468,17 @@ SpdySendStatus SpdyProxyClientSocket::OnSendHeadersComplete() {
   return NO_MORE_DATA_TO_SEND;
 }
 
-int SpdyProxyClientSocket::OnSendBody() {
+void SpdyProxyClientSocket::OnSendBody() {
   // Because we use |spdy_stream_| via STATE_OPEN (ala WebSockets)
-  // OnSendBody() should never be called.
-  NOTREACHED();
-  return ERR_UNEXPECTED;
+  // OnSendBody() must never be called.
+  CHECK(false);

[Ryan Hamilton, 2013/05/22 03:17:30]

CHECK(false)?  This is pretty hardcore.  Are you sure you want to crash the
user's browser in this case?

[akalin, 2013/05/22 08:40:05]

Yeah, I think it's for the best. If this were originally a CHECK(false), I think
we would have caught this bug much sooner!

I don't think I've ever regretted converting something to a CHECK(), as it
provides clear feedback on what goes wrong via canary/dev channel. The worst
thing that could happen is that it starts triggering on the stable channel, but
I think we're pretty good at catching things before they get that far.

 }
 
 SpdySendStatus SpdyProxyClientSocket::OnSendBodyComplete(
     size_t /*bytes_sent*/) {
   // Because we use |spdy_stream_| via STATE_OPEN (ala WebSockets)
-  // OnSendBodyComplete() should never be called.
-  NOTREACHED();
+  // OnSendBodyComplete() must never be called.
+  CHECK(false);

[Ryan Hamilton, 2013/05/22 03:17:30]

ditto

[akalin, 2013/05/22 08:40:05]

ditto.

   return NO_MORE_DATA_TO_SEND;
 }