Module wf

Core logic responsible for determining what it means for various type system primitives to be “well formed”. Actually checking whether these primitives are well formed is performed elsewhere (e.g. during type checking or item well formedness checking).

Structs

WfPredicates 🔒

Enums

Elaborate 🔒

Controls whether we “elaborate” supertraits and so forth on the WF predicates. This is a kind of hack to address #43784. The underlying problem in that issue was a trait structure like:

Functions

clause_obligations

Returns the requirements for clause to be well-formed.

extend_cause_with_original_assoc_item_obligation 🔒

Points the cause span of a super predicate at the relevant associated type.

object_region_bounds

Given an object type like SomeTrait + Send, computes the lifetime bounds that must hold on the elided self type. These are derived from the declarations of SomeTrait, Send, and friends – if they declare trait SomeTrait : 'static, for example, then 'static would appear in the list.

obligations

Returns the set of obligations needed to make arg well-formed. If arg contains unresolved inference variables, this may include further WF obligations. However, if arg IS an unresolved inference variable, returns None, because we are not able to make any progress at all. This is to prevent cycles where we say “?0 is WF if ?0 is WF”.

trait_obligations

Returns the obligations that make this trait reference well-formed. For example, if there is a trait Set defined like trait Set<K: Eq>, then the trait bound Foo: Set<Bar> is WF if Bar: Eq.

unnormalized_obligations

Compute the predicates that are required for a type to be well-formed.