Syncstorewjb

[wjblanke]

Purpose:

Fix peer_has_block recording

Current Behavior:

Doesn't check hash before recording block

New Behavior:

Check hash before recording block

Testing Notes:

None

[almogdepaz]

we have a bad peak cache we used for chip0013 that is currently not used, i think we can identify the peaks that fail due to bad genesis and use the bad peak cache this will help us identify WP's that belong to these bad peaks and avoid validating the WP for those peaks.

we would still need to request the WP since for new peaks that are not in the cache we need the WP to see if the last two epochs contained a block that is in the bad cache in order to fail

im not sure this is needed because the WP should fail on this pretty fast since it is the first thing we validate and that happens before all the vdf heavy lifting , lmk what you think

[emlowe]

With this change, we still ask for the WP, but we ask it from the correct peer - ie the peer that has the heavier weight peak. Without this, we would ask for the WP from a random selection of all our peers, since all the peers have that block height even though it's the wrong header hash - and since most of those peers produce a WP we accept, there is no error in WP.

[github-actions]

File	Coverage	Missing Lines
chia/full_node/full_node.py	50.0%	lines 781

Total	Missing	Coverage
2 lines	Unknown	50%