mikros/rust
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Rollup merge of #114524 - compiler-errors:more-ice-in-uncertainty, r=jackh726

Browse Source 
Also ICE when goals go from Ok to Err in new solver

We were just using `?` here, silently downgrading the goal's response from (presumably) maybe to error -- that seems concerning, since this whole check is for detecting goal instability 😅

r? `@lcnr` or `@BoxyUwU`
This commit is contained in:
Matthias Krüger 2023-08-06 03:56:10 +02:00 committed by GitHub
commit bf4b1d7805
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 8 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
compiler/rustc_trait_selection/src/solve/eval_ctxt.rs
View File

@ -391,13 +391,19 @@ impl<'a, 'tcx> EvalCtxt<'a, 'tcx> {
debug!("rerunning goal to check result is stable");
self.search_graph.reset_encountered_overflow(encountered_overflow);
let (_orig_values, canonical_goal) = self.canonicalize_goal(goal);
let new_canonical_response = EvalCtxt::evaluate_canonical_goal(
let Ok(new_canonical_response) = EvalCtxt::evaluate_canonical_goal(
self.tcx(),
self.search_graph,
canonical_goal,
// FIXME(-Ztrait-solver=next): we do not track what happens in `evaluate_canonical_goal`
&mut ProofTreeBuilder::new_noop(),
)?;
) else {
bug!(
"goal went from {certainty:?} to error: re-canonicalized goal={canonical_goal:#?} \
first_response={canonical_response:#?},
second response was error"
);
};
// We only check for modulo regions as we convert all regions in
// the input to new existentials, even if they're expected to be
// `'static` or a placeholder region.