Hey @Taylor Cramer -- let me know when you're around
@Taylor Cramer ohai =)
so I've been trying to close the various issues, as you've probably noticed.
The lifetime elision one (https://github.com/rust-lang/rust/issues/63388) has me the most worried, though
the current behavior is definitely incorrect
what you described is what the code does (looking for exactly one unnamed lifetime parameter, and using that as the elided lifetime)
I'm not sure how to fix that
I mean obviously we need to change the approach
well I'd say @centril's desire for lots of tests is clearly validated :)
where is the other lifetime elision code to be found...
middle/resolve_lifetime.rs, around line 2100
it seems...not entirely implausible that code could be factored out
well, maybe kinda implausible, as it uses a visitor I think
I am not happy to be down to the wire like this :(
ah well I guess a HIR visitor doesn't require a tcx