You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Jun 6, 2023. It is now read-only.
Catching non-determinism bugs might benefit from lots of runs. With the current flow we get a new chance to find a non-determinism bug every time CI runs a VM integration test. If we run these tests more we will get more chances. We could do this with a persistent background CI job running actors tests on the main branch, say every hour.
This might not do much if test runs are very similar to each other. A nice extension would be to investigate the go runtime and its entropy. A good place to start is understanding the non-determinism bug that halted mainnet last year and understanding if there is a way to configure CI/go runtime/tests to catch this particular bug. This could then give us ideas on generalizing this process as well as ensuring that this particular type of bug will be caught quickly. This also gives us a concrete case for checking if this whole approach is feasible.
The text was updated successfully, but these errors were encountered:
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Catching non-determinism bugs might benefit from lots of runs. With the current flow we get a new chance to find a non-determinism bug every time CI runs a VM integration test. If we run these tests more we will get more chances. We could do this with a persistent background CI job running actors tests on the main branch, say every hour.
This might not do much if test runs are very similar to each other. A nice extension would be to investigate the go runtime and its entropy. A good place to start is understanding the non-determinism bug that halted mainnet last year and understanding if there is a way to configure CI/go runtime/tests to catch this particular bug. This could then give us ideas on generalizing this process as well as ensuring that this particular type of bug will be caught quickly. This also gives us a concrete case for checking if this whole approach is feasible.
The text was updated successfully, but these errors were encountered: