00:14 <jdhorwitz> Last Week in Pony, September 16, 2018 - https://www.ponylang.org/blog/2018/09/last-week-in-pony---september-16-2018/

01:43 <jdhorwitz> What is a needed package currently for the Pony ecosystem? In my path of learning Pony better I was thinking of doing a blog series of creating a needed Pony library. Figure I would document myself fumbling through it

03:04 <vaninwagen> jdhorwitz: v5

03:06 <vaninwagen> jdhorwitz: oh my god, where to begin? There is so much missing stuff. First and foremost database drivers, postgres being the first, but there are others too, cassandra, foundationdb, ...

03:11 <vaninwagen> But this sounds like a bery big bunch of work... So, something smaller might be a sax xml parser (i know not the coolest shit), or a wrapper around a common c lib, support for the crypto stuff from libssl, or something like a process pool/actor pool

03:11 <vaninwagen> I keep em coming as they come to mind...

03:28 <jdhorwitz> Great , thanks! Yeah, for this first one, I'm looking for something smaller, not too difficult

06:24 <vaninwagen> What about an elasticsearch client? This mostly involves http and json

06:25 <vaninwagen> It also speaks some binary protocol, but that is very tied to java last time i looked at it

13:19 <jdhorwitz> That's an interesting idea!

17:50 <endformationage> vaninwagen: I think I've found another interesting bit about the generics issue I've struggled with.

17:50 <endformationage> As a starting point, here's the playground: https://playground.ponylang.org/?gist=82c701ea64a307b0cf5bd5c223278e87

17:51 <endformationage> As is, the generated error is the same where we left off.

17:52 <endformationage> First thing I tried was to use type parameter names in `_MappedEvents` that were different from those used by the Events trait.

17:52 <vaninwagen> carry on. Will there be a happy end?

17:53 <endformationage> ... and I got a different error message. This can be seen by swapping out the version of `_MappedEvents` in the code as seen here:

17:53 <endformationage> https://playground.ponylang.org/?gist=6c0f60e9f0263291c1b2def5c7754956

17:56 <endformationage> This made me wonder if there's some kind of issue with sharing type param names between the types.

17:56 <vaninwagen> Definitely possible

17:58 <endformationage> The second thing I did was swap out the trait for an interface, and looked at the results with both versions of the `_MappedEvents` type.

17:59 <vaninwagen> I didnt find the time to look into it more deeply yet :(

17:59 <endformationage> I was surprised to see different errors.

18:03 <endformationage> I suppose I'm not sure whether I should expect traits and interfaces to be interchangable, or at least in the context.

18:04 <endformationage> In any case the different error messages suggest different paths being taken in the compiler?

18:08 <vaninwagen> Exactly, the quest is to find which it is

18:08 <endformationage> Swapping `Events` to an interface with the version of `_MappedEvents` that uses unique type param names is interesting because it also mentions the 'no lower bound' from the original error, and also raises issue of 'not implementing its provides list' (which looks like it does).

18:09 <vaninwagen> This needs some time, a debugging session and lots of ast printing

18:09 <endformationage> .. but perhaps not in the sort of recursive genericly typed nature of the example

18:10 <endformationage> https://playground.ponylang.org/?gist=4fbbc249549ca91d62941d2bf4994701

18:12 <endformationage> As I'm simply not knowledgable or experienced enough in Pony's compiler (or compilers / type systems in general), this is really all I can help with. :/

18:12 <vaninwagen> That is already a lot!

18:13 <vaninwagen> I cant do nothing right now on my phone, but let me dig a little more tomorrow

18:13 <endformationage> I'd feel alot more confident if I was told the general principle of my intentions as shown my the example should be possible in Pony.

18:14 <endformationage> Again I really appreaciate you taking the time to look at this.

18:16 <vaninwagen> From my understanding your examples are very valid and exhibit a compiler bug

18:17 <endformationage> vaninwagen: Do you think I should file this in a separate issue?

18:19 <vaninwagen> Yep

18:20 <endformationage> OK, will do.

18:21 <vaninwagen> Thank you :)