-
Notifications
You must be signed in to change notification settings - Fork 1.6k
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
A very shallow review because I am not very fluent in this area. Also, I've missed that this is a inprogress
, sorry
ValidatorDiscovery(#[from] polkadot_node_subsystem_util::validator_discovery::Error), | ||
#[error(transparent)] | ||
Util(#[from] polkadot_node_subsystem_util::Error), | ||
} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not sure about the team's policy about that, but many of those can be constructed from different places. Having, say, OneshotRecv
at hand it would probably hard to tell where that error actually originated from.
One thing I realized, doesn't this require an impl guide update? |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I see the changes where we connect to the peers in our group about the PoV, but not the changes where we only notify and request from peers in that group. We will need to change the logic for computing peers_to_send
also.
My current understanding of the protocol is this: The first pathway is
The second pathway is
The path for networking messages: just handle In other words, the only proactive networking actions with connecting to relevant validators are taken on fetches, and we can assume that on distributing path we only receive Or am I missing something? |
Oh, now I see. I guess previously, the However, the call to |
Ah yes I agree. For some time I was thinking that we would re-receive the Also looks like the attributes on functions hide |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good modulo error handling.
bot merge |
Trying merge. |
It doesn't look like I've broken any existing tests or the only existing integration test so far. however I am most certain that some corner cases are not yet accounted for but for that I am needing fresh eyes and I will also be re-reading the code so marking at as ready for review now.
Fixes #1688