10:36 <TonyLo> Hi, any suggestions on how to make this more idiomatic Pony? (comments in the Gist should explain my objective, if not clear please let me know): https://gist.github.com/TonyLo1/e5fc19d7a60cde5fe3256d26c5114482

11:21 <TonyLo> Does anyone have an implementation of an interval heap that they are prepared to share?, thx

15:19 <SeanTAllen> i left a comment TonyLo

15:20 <SeanTAllen> and ya, lack of recursive types is a pain point. hopefully someone has a chance to address at some point in the not so distant future

15:24 <SeanTAllen> Video for the latest Pony VUG is now available. Good intro into Pony... https://vimeo.com/202387915

15:30 <TonyLo> @ SeanTAllen, thx for the feedback, yes I tried your suggestion already but it moved the issue somewhere else :) now I get this instead: main.pony:24:34: this parameter must be sendable (iso, val or tag)

15:30 <TonyLo> new val create(rel: Op, term': Array[Term]) =>

15:59 <aav> is there a way to cast I8 to U8 (and back)? or they are seen as fundamentally different types?

16:04 <TonyLo> you can use 'from': http://www.ponylang.org/ponyc/builtin-I8/

16:05 <aav> TonyLo: thanx!

16:07 <TonyLo> or let i: I8 = 50

16:07 <TonyLo> let u: U8 = i.u8()

18:12 <kulibali> why do method parameters adapt the viewpoint of the receiver?

18:32 <jemc> kulibali: can you give an example?

18:35 <TonyLo_afk> @jems, how complete is your zmq port?

18:37 <kulibali> oops, i'm wrong, didn't see that ReadSeq is box by default

18:40 <jemc> TonyLo: it works, and is interoperable with the reference implementation (including CURVE encryption)

18:41 <jemc> however, the API is not necessarily ideal in some respects, and I need to come back to see if it could be designed better

18:41 <jemc> also, it's not been used in any serious projects (that I'm aware of), so it's not really battle-tested

18:41 <jemc> if you're interested in using it in a project, I'd be interested to help support you and get your feedback on the API

19:06 <aav> is there any reason why the right side of shift (<<, >>) operator must be of the same type as the left side?

20:14 <jemc> aav: I think that it might come from an LLVM requirement, but not totally sure without looking

20:14 <jemc> in general, all the numeric binary operators use two operands of the same numeric type

20:16 <jemc> I think I see your implicit point that shifting bits is unlikely to need anything bigger than an 8 bit operand to indicate the number of bits to shift, but at the same time I think it's nice to have all the numeric binary operators work in a consistent way