Sean T Allen

release is underway.

All done

I think the commit comment for this should be changed. This doesnt at the moment remove any refcap ceremony. It removes an unneeded variable. With my noted change, it could...

I discussed this with @sylvanc. He agrees with my assessment that this is a time of check/time of use where that has to mean that the actor was being executed...

There's a TOCTOU in `ponyint_sched_unmute_senders` plus how we decide when sending whether to schedule an actor or not. For example in `pony_sendv` we do: - A1: ponyint_actor_messageq_push IF que was...

Interestingly, I can't find a way to have B2 in `ponyint_sched_unmute_senders` that doesn't result in a bad interleaving. However, I have an idea of how to do this so that...

I forgot about a protection that is built in so the above TOCTOU isn't valid. So I am back to where I was when I did the last review. It...

I spent time last night looking at what assumptions I was making about lower level things working correctly and then trying to verify them against Verona which has some similar...

There's a deep exploration and discussion of this issue in VUG #12: https://vimeo.com/695067236

here's a minimal TSAN report from @mjp41: https://gist.github.com/mjp41/3747d988209ca1d0d18a5f09ef9d67c7 as referenced 3 days ago in https://github.com/ponylang/ponyc/issues/4069#issuecomment-1083313798