We should have a leaderboard issue on the r-a issue tracker for people to post their observed peak memory usages of rust-analyzer. So far the highest a friend reported to me was 80gb I think

That's great to hear that those PR improved perf as I've hoped to!

You mean: when can we expect the Amos builds their own build system video/blog post?

The book is very outdated

A bit annoying but curling this with a classic token does the trick docs.github.com/en/rest/acti...

You'll have to send a request to one of the rest api endpoints to mark everything as read, Ive gotten the same issue

Oh damn this is a huge quality of life improvement

Yes, but keep this cursed knowledge to yourself.

Tt munchin data carrying enums sounds pretty difficult, not sure if you can reasonably do that

The c programmer mindset

Woah congrats!

It seems like we can mostly ignore those new bounds and thing should just work™ again but we'll see

> However, are there problems with a proc-macro returning more than one item?

That is also just fine, derives are not even required to output a trait impl. It's just convention as that was what they initially were meant for. Nowadays its more just a macro kind that cannot change its input.

Note that `const _` is exemtp from the non-local lint, as its currently the only way to generate code without polluting the surrounding scope due to the lack of stable declaration site hygiene. So the first one is be preferred if you don't want to leak generated code out of its scope.