Tim Watson

Okay, I'm going to update all the library ecosystem that way then. Can I leave you to do this for distributed-process/static, and close this bug?

Thanks, I'll update the API docs. This is handled better by distributed-process-extras, which has a separate unit of measure in the data type.

Could this be related to the issue with network when is compiled with -threaded? Do you both see this for spawnlocal too?

> Can you try export DISTRIBUTED_PROCESS_TRACE_EVENT=y instead, this should reduce this kind of overhead? @qnikst do you mean `DISTRIBUTED_PROCESS_TRACE_EVENTLOG` or `DISTRIBUTED_PROCESS_TRACE_CONSOLE`? Also, if you don't set any trace flags, there...

Also, tracing might still be a bit broken for some cases, as per #265. That is high on my stack to fix, and I'm hoping to get to it once...

@qnikst can you point me to the discussion about `flush` please? I'm going to look into this after the v0.8 release is done.

@qnikst I saw that yeah, question is what are the semantics for deciding when to call `flush`? I wouldn't have thought implementing `flush` is too onerous, but knowing _when_ to...

Also, does this issue persist with the example repository/project you wrote, @teh, if you use a different network-transport library? For example the network-transport-cci library, or network-transport-zeromq? Those libraries may be...

It's possible your os isn't configured to multicast to the interface you're trying to use. What does netstat -nr look like?

So I don't see a multicast range IP address in the routing table there. You need to add the route so that all the servers know about the group, something...