02:41 <TwoNotes> Praetonus, ponyint_mpmcq_pop https://gist.github.com/pdtwonotes/e8c0cf04ab366373ce848a3b7b024c26

03:25 <DanC_> I'm giving this RFC process a try: https://github.com/dckc/rfcs/blob/ffi-taming/text/0000-ffi-taming.md https://github.com/ponylang/rfcs/pull/28

03:25 <DanC_> I'm not confident I have the bandwidth to advocate it in the relevant forums, so I hope I can find a co-champion.

03:30 <DanC_> SeanTAllen, any chance you're around to chat about this?

03:52 <doublec> DanC_: I like the idea

03:53 <DanC_> :)

03:53 <DanC_> do you participate in the sync meetings and such?

03:53 <doublec> I'm not sure how painful it will be passing ambient auth down to places where ffi is used though

03:53 <doublec> eg. I often @printf[None](...) for debugging

03:54 <DanC_> ah... it's somewhat traditional to make an exception to this policy for logging/tracing. Perhaps I should add that.

03:54 <doublec> But I remember reading a paper about Oz-e (which never eventuated) that warmed me on the idea.

03:55 <doublec> DanC_: unfortunately the sync meetings clash with work meetings for me

03:55 <DanC_> yeah...

12:47 <SeanTAllen> DanC_: i was not around. i'm on US east coast, tend to go to bed early, rise early.

12:51 <TwoNotes> It looks like there may be a fix for issue #1000. I wonder if there are other places that do _atomic operations that should be looked into?

12:52 <TwoNotes> Or maybe some sort of test that really beats on those sections of the RT

14:10 <TwoNotes> Need a test that forces the work-stealing code to run a lot.

14:26 <jemc> TwoNotes: intuitively, to force lots of work stealing, I'd think you could do something like have an actor that makes a bunch of behaviour calls from a single central behaviour, but hogs its own scheduler thread

14:27 <jemc> to hog its own scheduler thread, you could try doing "bad things" like a big `while` loop in you behaviour that makes a bunch of behaviour calls then calls `sleep` with FFI in each iteration :)

14:28 <jemc> I'd think in that case that pretty much all of the work would get stolen

21:35 <lisael> Hi, I've just posted a new issue on github, I had a segfault while benchmarking String

21:35 <lisael> https://github.com/ponylang/ponyc/issues/1186

21:39 <lisael> I have to go to bed, too, I'm in Europe. I'll be here tomorrow if someone needs info.

21:39 <lisael> Bye!

22:44 <SeanTAllen> thanks lisael