Force thunk in bufferedTxs’ binding in acknowledgeTxIdsInMempool
Semigroup operation for
union, which is
actually not implemented in such a way that values of the resulting
Map are forced to WHNF.
An odd thing that I noticed is that the
union function that is exposed
Data.Map.Strict module is the same as that exposed from the
Data.Map.Lazy module. Internally, this function uses lazy
and therefore provides no guarantee that the values of the resulting
Map will be in WHNF. It seems misleading to me that this function
would be exposed from
Data.Map.Strict without providing that guarantee
which is promised by the module (see its documentation).
679: CAD-742: Chairman’s remainder of maintenance: topology & log analysis r=deepfire a=deepfire
- Generate & document the legacy topology, instead of having it hard-coded.
- Add more detailed tracing of headers during ChainDB updates.
- Fix the Chairman log analysis scripts to once again be capable of graphing the blockchain.
- Refactor the Chairman log analysis scripts to a different model, where all intermediate information is dumped in a directory, keyed by node id.
- Chairman: deduce the delegate <-> node Id mapping from the set of keys (de-hardcode).
scripts/cluster-test.sh(but not the CI test itself) perform log analysis at the end. Although we could make that a separate derivation..
- This PR does not result in breaking changes to upstream dependencies.
[x] This PR contains all the work required to resolve the linked issue.
[ ] The work contained has sufficient documentation to describe what it does and how to do it.
[x] The work has sufficient tests and/or testing.
[x] I have committed clear and descriptive commits. Be considerate as somebody else will have to read these.
[x] I have added the appropriate labels to this PR.
Co-authored-by: Kosyrev Serge [email protected]