check individual mempool per message vs. shared mempools usage
The following discussion from !8526 (merged) should be addressed:
-
@pspacek started a discussion: (+1 comment) Now the hard question: Was this caused by some recent change in the mempool usage? @ondrej?
If so, can we have, say, a Cocinelle check for the correct order of operations? Chasing down these shutdown issues one by one is nightmare and consumes QA time, so if we can have an automated check I'm all for it.
If an automated check is not feasible please could you manually check places affected by (presumed) recent changes to see if there are other place with similar bugs?
Thank you!