16:51 <Colinot> Hey

16:52 <SeanTAllen> hello Colinot

16:52 <Colinot> How are you ?

16:53 <SeanTAllen> Fine. And yourself?

16:53 <Colinot> Good thank you

16:54 <Colinot> I'm looking for someone willing to help me learn pony, i'm a real beginner. Could you help ?

16:55 <SeanTAllen> we are here to help as best as we can when you run into issues.

16:55 <SeanTAllen> we are most able to help if, when you run into problems, you can provide a small sample of code that demonstrates your problem as well as an explanation of what larger goal you are trying to accomplish

16:56 <SeanTAllen> there's also the mailing list as a means of getting assistance

16:56 <SeanTAllen> see: https://www.ponylang.org/learn/#getting-help

17:00 <Colinot> Ok, thank you

19:34 <endformationage> What are the circumstances when an actor stops processing messages? Garbage collection?

19:39 <endformationage> I am begining to think my problem stems from this. An actor of mine simply stops processing messages, and the CPU is at 100%. Messages are still being sent to this stalled actor after this occurs.

19:44 <endformationage> I think before, my removal / addition of the json iso through this actor simply stalled or hastened gc.

19:49 <endformationage> Ah, increasing --ponygcinitial allowed the program to complete. So, definately a GC issue.

19:53 <endformationage> Looks like garbage collection hangs and eats CPU on my system for some reason.

19:56 <endformationage> To be clear, a single core is at 100% (supposedly the gc on that actor)

20:04 <endformationage> FYI, on 0.14.0 [release], compiled with: llvm 3.9.1 -- gcc (GCC) 5.4.0, NixOS 17.03

21:23 <SeanTAllen> what do you mean by "stops processing messages" endformationage?

21:23 <SeanTAllen> how do you know it "simply stops processing messages"?

21:24 <SeanTAllen> can you clarify "I think before, my removal / addition of the json iso through this actor simply stalled or hastened gc." ?

21:25 <endformationage> I made debug calls just after calling behaviors to make sure they were called, as well as just inside behaviors to indicate they are run.

21:29 <endformationage> I used a timer to make the same exact calls to the actor over time. Whether I actually send json, a string, or nothing at all, the described results will still occur. It is just the number of calls that can be made is more or less before the stalling occurs.

21:36 <endformationage> To be precise: I call the apply behavior on the trouble actor, which consumes an iso object (which originally carried the json). This always succeeds.

21:41 <endformationage> This object is processed just fine up till a point, presumably the point at which gc is triggered between message processing.

21:42 <endformationage> The apply method basically passes this object on to a 'one off' actor, along with some other construction.

21:46 <SeanTAllen> do you have a minimal example?

21:46 <SeanTAllen> that reproduces the problem?

21:48 <endformationage> I failed to be able to reproduce last night, but that was before I suspected garbage collection. I will try again.

21:50 <SeanTAllen> how much code is there that can reproduce?

21:51 <endformationage> Not sure if you caught it, I was able to run it fine, through to quiescence by increasing --ponygcinitial. Not sure if that really nails it down to the gc or not.

21:59 <SeanTAllen> i did.

22:18 <endformationage> OK, I think I've reproduced ...

22:38 <endformationage> SeanTAllen: https://gist.github.com/patternspandemic/b0ab0872788831aa5aa52ce071f72ca3

22:43 <endformationage> Err, going to update the example without SSLContext to make it easier to run

23:54 <SeanTAllen> endformationage: this does what for you?

23:54 <SeanTAllen> for me it starts up, prints nothing, then exits

23:55 <SeanTAllen> o i see i need debug

23:56 <SeanTAllen> so this works fine for me endformationage

23:56 <SeanTAllen> what OS, LLVM, etc are you using?

23:57 <SeanTAllen> also how many cpus does your computer have?