This is partially due to the fact that our analysis query

does not invoke the optimized_mir query on all items in the current crate, and also because we don't run monomorphization when building bin crates.

• run optimized_mir query in analysis
  • this may cause slowdowns for cargo check because we'd be doing optimizations now
  • All const propagation lints are emitted in optimized_mir, but we may add more in the future.
• run monomorphization where possible.
  • Not sure how useful this is, it will only give us additional lints in very specific situations. (e.g. when we subtract something form an associated constant, we may get an overflow error depending on the value of the constant).
  • This will also be significantly more expensive than just running optimized_mir

cc @Centril