Skip to content

don't return already errored pending circuit when searching new circuit matching spec

trinity-1686a requested to merge trinity-1686a/arti:fix/210 into main

fix #210 (closed)

see root cause analysis there

I've run the test on a very busy system (stress -c 96 on 8c16t cpu) 10k times, and did not get a single error (before, this setup would create errors about two thirds of the time).

Merge request reports