So I think we should do another sprint week -- which really I think means like 3 dedicated mornings? maybe 4? :) Or at least I would like to do that. Probably early to mid June.
Maybe we can start planning a bit the dates and agenda?
As usual I feel completely lost and don't remember where we were, but it'd be cool to line up some goals etc
early to mid june sounds good to me
as for agenda, what I had in mind was:
- (go over some of the recent work I've done (which I have yet to open PRs for anyway))
- a bit of planning / design: I'd like to talk with you all about our plans, especially about some of the feature parity work we're doing, which might have an unavoidable cost, which is fine for a reference implementation say, but maybe not for production. We're redoing some dataflow work, in a more general setting which cannot take advantage of the dataflow setting, i.e oversimplifying, there's no super easy way to take advantage of computing over a CFG rather than an arbitrary relation.
- talk about the "Locations::All" fix I had in mind and prototyped, and how it impacts our rules
- I walked through the valley of the shadow of death while looking at NLLs performance on the whole clap crate, so maybe talk about location-insensitivity, how I think we could do it outside of datafrog since it's in my mind actually just a case of graph traversal here, etc (which I also have prototyped)
- (if, and it's a big if, I manage to actually achieve anything with them: maybe talk about the other ideas I had wrt polonius' model itself)
- maybe some design work for niko, about the equality variant :) check the couple rule I "had" to add to it, maybe we try to understand why one or two rustc tests failed with the variant, those kinds of things
- actually emit the move errors to validate our rules, I haven't been able to look at this yet so it'd be reassuring to do so
@Jake Goulding is that a good :astonished: emoji or a bad :astonished: emoji ? :)
It's me being a spoiled brat, really... I'm super excited to see Polonius happen, but I haven't seen any progress here so I figured that everyone was busy with other things. I'm happy to see that the work just wasn't being logged here (which I don't have any right to expect in the first place).
ah sorry about that, I should have thought more about the people following the work here
Not at all. If the choice is between y'all making 2x progress vs 1x progress and 1x progress updates, I'm OK with the former.
I guess I would be a good candidate for looking at emitting move errors, and documenting the new rules in the Book
(I'm still here, but I have done exactly nothing since my thesis, so it's exactly what it looks like)
Early June may work for me, but I also have vacation basically late June to early August
Last update: Jun 20 2021 at 00:30UTC
(as I'll take vacation days off of work for this, the exact scheduling is open for me, to whenever y'all are available)