creating topic to discuss compiler-team#287
Hey @T-compiler/meeting -- so in leadup for this meeting @pnkfelix and I put some effort into this document which has a rough draft of compiler team priorities + goals. It's got some comments that try to explain the various sections but please take a look. Also maybe worth looking at the lang team ideas and priorities doc too, as I hope to reconcile the two of them somewhat.
I'm sure stuff is missing so feel free to add notes or thoughts into the doc (but identify, if you do, those additions to make them easy to spot)
I filled in the const-eval and mir-opt placeholders
I'll try to summarise where const generics is at a high level soon. Lazy normalisation was the main blocker, and now that the first PR has been merged, hopefully we can make some solid progress towards making it more stable.
@varkor I was doing some work today on I think removing some of the other big blockers (on the topic of lazy norm...)
I think I finally figured out how to manage that...
...but that's for another topic...
I wanted to say that @matklad's post on the next few years of rust-analyzer seems pretty germane here, if you've not read it
So to circle back to this, @pnkfelix, first off I sent you the wrong hackmd earlier. This is the correct hackmd, and I edited the issue to reflect that.
But separately, I was thinking after our lang-team meeting, that I think there is an interesting mix of "top-down" and "bottom-up" that the lang-team meeting didn't quite steer, and I'd like to try and shape this coming one a bit differently. I think the right way for us to be working is more about "identify the priorities" and have people tell us what actions they expect to take that fit in that framework
but it's also maybe a good place for us to brainstorm a bit about specific areas and what we can do there
So to circle back to this, pnkfelix, first off I sent you the wrong hackmd earlier. This is the correct hackmd, and I edited the issue to reflect that.
you know, I did skim that (the erroneous hackmd) and said "this doesn't look quite right" but I couldn't put my finger on what was wrong. Now I realize it was the notes for a different roadmap discussion. Sort of sad/funny how much stuff still applies today...
I had that same experience :)
Well, it makes sense, doesn't it? Some of the best things aren't built in a couple of years, but over many
I bet NLL was a thing in the to-do list before 1.0
"borrow scopes should not always be lexical" #6393
(dates from 2013)
Wisdom from old times