[pallet-broker] Force-unpool provisionally pooled regions before redispatching them #4081
Conversation
seadanda
changed the title
Co-authored-by: Bastian Köcher <[email protected]>
|
The CI pipeline was cancelled due to failure one of the required jobs. |
| @@ -390,6 +390,14 @@ pub mod pallet { | |||
| /// The Region whose contribution is no longer exists. | |||
| region_id: RegionId, | |||
| }, | |||
| /// A region has been force-removed from the pool. This is usually due to a provisionally | |||
| /// pooled region being redeployed. | |||
| RegionUnpooled { | |||
There was a problem hiding this comment.
This is the reason for the major bump, but since there is an event for a region being pooled we should also have one for it being unpooled.
| // the current timeslice if we are already part-way through the region. | ||
| let size = region_id.mask.count_ones() as i32; | ||
| let unpooled_at = end_timeslice.max(region_id.begin); | ||
| InstaPoolIo::<T>::mutate(unpooled_at, |a| a.private.saturating_reduce(size)); |
There was a problem hiding this comment.
Do we leave the WorkPlan as is on purpose?
|
Hmm yea I just updated from master today after a while of waiting for audit, and it seems that #3940 changed the mock for these tests |
|
All GitHub workflows were cancelled due to failure one of the required jobs. |
This PR force removes regions from the pool before allowing them to be redispatched (through
partition/interlace/assign) in the case that a region was pooled withProvisionalfinality.This PR does not account for the case where a pooled region already entitles the benefactor to a contribution reward before the point of it being redispatched. However,
claim_revenueshould be called before trying to redispatch a region anyway.Otherwise there would be ambiguity as to what should be done with the first part of a partitioned region. The first of the two new regions actually has the exact same
region_idas the original region -- leave it pooled/withdraw from pool/claim contribution then withdraw?