Hello! I'm back from vacation today, but I somehow wound up double booked.
So @WG-traits I won't be around at our usual time today (i.e., in 10 minutes)
But maybe we can leave some brief notes
Over vacation, I was working on "the chalk refactoring" and I made a lot of progress. The fact that I consider this vacation is perhaps something I should be discussing with a mental health professional, but never mind that.
I also found it was a more complex issue than I initially thought and I think it'd be great to go over some of what I found and what design we want
I've been talking a lot about trying to have a recurring "design zoom call" (recorded, of course) to go over this sort of thing -- maybe this is a good first candidate?
I think the major priority for rutsc-chalk integration remains kind of enumerating clearly the work to be done -- @tmandry / @Aaron Turon, were either of you able to make any progress on those builtin-bound items?
I made a list of other tasks before I left but didn't do a good job enumerating it
But that remains (I think) the key thing we need to do this week.
And I've been talking to @Alexander Regueiro about trying to make progress on object upcasting
I'm trying to decide just what is needed there -- it feels like there is at least some lang team involvement -- so we should discuss the overall procedure.
Anyway, maybe we can run the meeting a bit async? If people want to leave thoughts I'll try to read and catch up :)
I wasn't able to start the work on chalk integration, most of my time has been wrapped up in figuring out the generator optimization stuff (@eddyb and I had to iterate on the design a bit)
but I haven't forgotten, and may have time to look at it this week
I was thinking of starting to do the work on modelling specialization in chalk. We had a meeting about this back in January (oh wow :open_mouth:) where we discussed some initial thoughts about how we might want to go about it
OK. That all sounds good.
How's lifetime constraint stuff in Chalk going?
as I understand that's still the main blocker for usability
Isn't lifetime stuff part of polonius? (I think that's separate from chalk)
We need to handle lifetime constraints during trait solving too
Ah okay. I didn't know that was part of chalk