Stream: wg-traits

Topic: SLG and the chalk-engine crate


nikomatsakis (Apr 22 2019 at 15:44, on Zulip):

Hi @WG-traits! Reminder that today in ~20 minutes we plan to do a call about how the chalk-engine crate works. This call will be recorded. But the more people that can join us on Zoom the better!

nikomatsakis (Apr 22 2019 at 15:53, on Zulip):

Also, here is a Dropbox Paper document that we'll be populating as we walk through (presently empty)

nikomatsakis (Apr 22 2019 at 17:02, on Zulip):

So, if we were to continue again, (a) we should figure out when is a good time, and (b) it'd be useful to get feedback on where to continue from. My sense is that I should just basically try to cover the actual chalk-engine code in more detail -- so starting roughly from that point?

nikomatsakis (Apr 22 2019 at 17:02, on Zulip):

Er, from roughly after the point where I said "here is a forest" :)

Nick Lawrence (Apr 22 2019 at 17:04, on Zulip):

Yeah - my mic started acting up, but I'd like to know if theres a way we order the tables before evaluating

Nick Lawrence (Apr 22 2019 at 17:05, on Zulip):

since certain evaluations can be used in others (if I understand correctly)

scalexm (Apr 22 2019 at 17:06, on Zulip):

@nikomatsakis after we've covered the actual chalk-engine code, we might want to get to the existing built-in bounds problem (which is somehow related with @Nick Lawrence's question)

scalexm (Apr 22 2019 at 17:06, on Zulip):

plus I should be available this week for a presentation of implied bounds, as I'm on PTO

nikomatsakis (Apr 22 2019 at 17:11, on Zulip):

we might want to get to the existing built-in bounds problem

yep, definitely. That was sort of my goal, but I figured it'd be important to try and explain what the code itself is doing now. (I have a vague proposal for how to handle things like ?T: Sized, at least, though I want to talk out what it means...)

nikomatsakis (Apr 22 2019 at 17:11, on Zulip):

plus I should be available this week for a presentation of implied bounds, as I'm on PTO

ooh, nice! Let's schedule something!

nikomatsakis (Apr 22 2019 at 17:12, on Zulip):

Yeah - my mic started acting up, but I'd like to know if theres a way we order the tables before evaluating

do you mean @Nick Lawrence how to order the subgoals?

nikomatsakis (Apr 22 2019 at 17:12, on Zulip):

That is, given some strand like Foo(T) :- Bar(T), Baz(T)

nikomatsakis (Apr 22 2019 at 17:12, on Zulip):

we wind up having to pick whether to evaluate Bar(T) or Baz(T) first

nikomatsakis (Apr 22 2019 at 17:12, on Zulip):

and it matters quite a bit which we pick

nikomatsakis (Apr 22 2019 at 17:12, on Zulip):

I didn't really get to that yet...

nikomatsakis (Apr 22 2019 at 17:13, on Zulip):

and it matters quite a bit which we pick

for performance and practicality, that is; theoretically it doesn't matter much at all

Nick Lawrence (Apr 22 2019 at 18:01, on Zulip):

@nikomatsakis yes, I might mean that

nikomatsakis (Apr 23 2019 at 20:57, on Zulip):

Video posted

nikomatsakis (Apr 23 2019 at 22:01, on Zulip):

When should we do the next session? Next week? Not sure that I have time this week

nikomatsakis (Apr 23 2019 at 22:01, on Zulip):

Maybe somebody can create a Doodle poll?

Last update: Nov 18 2019 at 00:40UTC