naive MG test on CloseChannel, potential naive behavior on tProxy

[plebhash]

It's not clear what pool-sri-test-close-channel tries to achieve.

There's nothing in the specs stating that a proxy must close the entire connection when it receives a CloseChannel from Upstream.

This feels like a naive test. What if the proxy has other active channels?

Also relevant:

stratum/roles/translator/src/lib/upstream_sv2/upstream.rs

Lines 397 to 400 in 61b5cf0

	Mining::CloseChannel(_m) => {
	error!("Received Mining::CloseChannel msg from upstream!");
	handle_result!(tx_status, Err(NoUpstreamsConnected));
	}

[plebhash]

from @Fi3 over Discord:

not sure
depend what we want to test
and what tproxy is expected to do
the thing that make most sense impo
is to close downstream connection
reopen the channel with upstream
and wait for the miner to reconnect
tproxy from what I know can not open more then one channel
so tproxy having more channle opened is impossible
when if tproxy will support more then one channel opened with upstream we can reason about it