You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add a new event emitted by the relay chain runtime whenever a candidate is included, which contains the approved peer sent by the parachain over UMP. This will be used by collator protocol for bumping reputations.
It makes sense to emit this on inclusion rather than backing, in order to make sure that backing an invalid candidate and then not making it available would result in a slash. Otherwise, backers could use this to bump the system-wide reputations of select collators without doing the work of creating valid blocks.
An alternative would be to explicitly query the candidates_pending_availability at the previous relay parent, but this would be more wasteful and would use an extra runtime API which potentially returns a lot of extra data.
The text was updated successfully, but these errors were encountered:
I just think it's simpler. You process much less data and you don't need to go back to a block where the candidates were still pending availability.
Anyway, I think either approach is fine really, and it's not a high priority issue.
We can just rely on the candidates_pending_availability runtime API for a while
Add a new event emitted by the relay chain runtime whenever a candidate is included, which contains the approved peer sent by the parachain over UMP. This will be used by collator protocol for bumping reputations.
It makes sense to emit this on inclusion rather than backing, in order to make sure that backing an invalid candidate and then not making it available would result in a slash. Otherwise, backers could use this to bump the system-wide reputations of select collators without doing the work of creating valid blocks.
An alternative would be to explicitly query the
candidates_pending_availability
at the previous relay parent, but this would be more wasteful and would use an extra runtime API which potentially returns a lot of extra data.The text was updated successfully, but these errors were encountered: