06:05 <sebastian_> I'm sure it's a common, boneheaded mistake, but I really can't figure my way around using map_stateful. I'm trying the following: https://pastebin.com/raw/bLCZp0s0 and receiving these errors: https://pastebin.com/raw/LMX7MSGh

06:05 <sebastian_> I can't figure out if the base of the problem is a capability issue, or just a misuse of basic language constructs

06:05 <sebastian_> any help is much appreciated! Thanks

06:13 <sebastian_> I'll be offline for a bit, but I'll come back and check the log

06:23 <doublec> sebastian_: see http://playground.ponylang.org/?gist=30e03ade5b5ba474166fe9bce2fa05d3

06:24 <doublec> sebastian_: map_stateful maps from a State to a State, so I changed the result from 'None' to 'State'. This means I have to return the modified state.

06:24 <doublec> sebastian_: collect() takes a generic argument that defaults to an Array. Since you use a list we need to be explicit about it being a List, hence: collect[List[State]]

13:47 <kjekac> hi all! currently trying to write some Pony code after reading the tutorial. is "classes are for data, actors are for behavior" a useful rule-of-thumb?

13:50 <kjekac> sorry, realize that's an unfortunate choice of words. I don't mean "behaviors" as in the language construct, but as in "this is what my program should do" (as opposed to "this is how my program should store and handle its data")

14:32 <SeanTAllen> kjekac: kind of?

14:32 <SeanTAllen> classes are good for "do" but not "communication"

14:32 <SeanTAllen> so it depends on how you define "what my program should do"

14:33 <SeanTAllen> that said, it sounds like a reasonable rule of thumb to start with

14:33 <SeanTAllen> one thing that helped me when i started was i added documentation to parts of the standard library and got to see what was an actor and what was a class.

14:33 <SeanTAllen> that might be helpful for you as well.

15:09 <kjekac> SeanTAllen: ah, that makes sense, especially that classes are for "do" but not "communication", but I also realize that I need to refine my notions of what those two concepts constitute.

15:11 <kjekac> for example, I realized just now, that if multiple actors should be able to write to some data structure, that has to be handled by an actor, even though "write" feels more like "do" than like "communication" to me, at least intuitively.

15:11 <kjekac> add documentation is probably a good idea if I have further problems moving forward. :)

15:12 <kjekac> might come back with some other specific questions though.

15:17 <sebastian_> doublec: Thanks for the tips! That all makes sense, and works inside the Pony Playground, but when I try it locally I still get an error: https://pastebin.com/raw/9xMU7f7G

15:19 <sebastian_> Is there some difference between flags used by the Playground and just running ponyc? Or a substantial version difference?

15:19 <sebastian_> This is my `pony --version`: https://pastebin.com/raw/x9hW5ztj

15:30 <SeanTAllen> sebastian_: playground is on 0.19.1. List of changes is here: https://www.ponylang.org/blog/2017/09/0.19.1-released/

15:31 <SeanTAllen> the code in question is taking advantage of new lambda inference capabilities that aren't in 0.19.0

15:31 <SeanTAllen> if you upgrade locally, you should be good.