To spot regressions it would be nice to add more tests around GC.
It's inherently slow and unreliable, the best thing I could think of is having a mainnet node run with latest main and good monitoring/log aggregation to spot those. A node would need to run at least 3 days to yield reliable results.
Another solution would be to do what @lemmih did in his branch, running the GC super often and ignoring the possibility of a fork, but that'd need a special workaround.
I don't think it's very high prio, but let's keep it in the backlog.
To spot regressions it would be nice to add more tests around GC.
It's inherently slow and unreliable, the best thing I could think of is having a mainnet node run with latest main and good monitoring/log aggregation to spot those. A node would need to run at least 3 days to yield reliable results.
Another solution would be to do what @lemmih did in his branch, running the GC super often and ignoring the possibility of a fork, but that'd need a special workaround.
I don't think it's very high prio, but let's keep it in the backlog.