Samuel Moelius
Samuel Moelius
> A couple of things here: > > * This can calculate possible borrowers multiple times per function. It also fully recalculates the liveness bitset each time as well. What's...
> That would unconditionally add another traversal over every function. I thought what I was proposing would conditionally add another traversal (i.e., if all of the conditions were met so...
> You suggested to switching the lint to use `check_body` with the current `check_expr` moved into a visitor, correct? That leaves the visitor with all the lints, plus the new...
> I'll check to see if we can change rustc to allow storing `TyCtxt` bounded items in lint passes as that would allow caching the MIR result directly. Thanks! Did...
Thanks, @Jarcho!
I saw you PR was merged. But it looks like it didn't make it in time for the most recent rustup? If I'm reading everything right, the current toolchain (nightly-2022-09-08)...
I think this is ready for review again, @Jarcho. (Rebasing was a bit of bear, so hopefully I did not screw it up.)
I think the three most recent comments have been addressed.
> I think the only thing remaining is the previous point about changing `f1(&x); f2(&x);` to `f1(&x); f2(x);` Meaning _don't_ suggest that, right? It's possible there's a problem, but I...
> So it is. I was going by the change in `clippy_dev/src/new_lint.rs` still being there. (there's no need to change this). It should have occurred to me to fix that....