Skip to main content

maybe_evaluate_root_goal_for_proof_tree_with_higher_recursion_limit

Function maybe_evaluate_root_goal_for_proof_tree_with_higher_recursion_limit 

Source
fn maybe_evaluate_root_goal_for_proof_tree_with_higher_recursion_limit<D, I>(
    delegate: &D,
    goal: Goal<I, I::Predicate>,
    span: I::Span,
    initial_result: &mut (Result<NestedNormalizationGoals<I>, NoSolution>, GoalEvaluation<I>),
)
where D: SolverDelegate<Interner = I>, I: Interner,
Expand description

The old solver doesn’t check depth requirement when looking up cache while the next solver does so. Thus the next solver is more prone to overflow. To mitigate breakages, we re-evaluate the overflowed goal with doubled recursion limit and emit a FCW if it succeeds. See the doc comment on RECURSION_DEPTH_EXCEEDING_LIMIT and #159228 for more details.