JDC is not doing fallback to solo-mining

[GitGab19]

During testing, I noticed that sometimes JDC disconnects from upstreams because of some shares refused by the pool (this needs to be investigated better).
In this case, our default config for JDC doesn't contain a second upstream, so JDC should fallback to solo-mining.
But it's not doing that, instead it just hangs there.

[GitGab19]

For a better context, here's a screenshot I did when I noticed the bug during testing:

[GitGab19]

To reproduce the issue, just try to fake the prev_hash used by the pool to check shares sent by downstream, by changing line 1159 in protocols/v2/roles-logic-sv2/src/channel-logic/channel_factory.rs with this:

//let prev_blockhash = crate::utils::u256_to_block_hash(referenced_job.prev_hash.clone());
let prev_blockhash = crate::utils::u256_to_block_hash([220, 157, 233, 68, 198, 232, 73, 141, 133, 12, 97, 197, 97, 244, 132, 146, 127, 171, 248, 91, 23, 159, 215, 209, 2, 0, 0, 0, 0, 0, 0, 0].into());

[plebhash]

To reproduce the issue, just try to fake the prev_hash used by the pool to check shares sent by downstream, by changing line 1159 in protocols/v2/roles-logic-sv2/src/channel-logic/channel_factory.rs with this:

//let prev_blockhash = crate::utils::u256_to_block_hash(referenced_job.prev_hash.clone()); let prev_blockhash = crate::utils::u256_to_block_hash([220, 157, 233, 68,

it would be good to reproduce this with MG

[plebhash]

partially solved by #1001