Compile time regression in #44186

[Mark-Simulacrum]

Up to a 6% regression in performance is hiding somewhere in this range: 7eeac1b...890c87b. I suspect 5c279a4, but it could be something else as well -- that's the only one that stands out to me though.
cc @alexcrichton

[arielb1]

Looks like 5c279a4 - we're hashing the in-scope traits for a bunch of more things in <NodeId as HashStable>::hash_stable

[arielb1]

Umm no the problem is that we're hashing literally every HirId ever