bye bye normalizes-to

The old solver does not support the concept of normalizes-to, we always have the fallback to equating as a rigid projection behavior. It should be safe to maintain that behavior with the new solver as well and then maybe change it at some point in the future.

how projection goals work in the new solver

One-step normalization can either assemble no candidates, keep the term unconstrained, or actually constrain it in some way.

how projection goals work in the old solver

If we fail to project, equate the term with a rigid projection. In case of ambiguity, we stall. We make inference progress if selection results in a single candidate. We do not evaluate nested goals during selection if there is only a single candidate, which is different in the new solver.

If selection succeeds ignoring nested goals, but fails once we evaluate them, this means that there was a single impl candidate and no candidates from the ParamEnv. This case would error in the old solver and result in a 'rigid projection' in the new one. Given that such a projection is not WF we error either way.

issue

AMBIG w constraints -> ERROR: do not want to emit an error like "failed to normalize <partially inferred assoc ty> to <rigid projection>". This should only happen if the projection is not well-formed

issue

Projection unconstrained term hack. How can the parent goal differentiate between "normalizes to itself" and "failed to normalize". One should result in overflow, the other one should not.

idea: normalizes-to which only expects and infer var as it's term

structurally normalizing in the solver

alias-relate

With Projection, it should be possible to use structurally_normalize_ty(lhs) == structurally_normalize_ty(rhs). This has good perf :3

structurally normalizing outside of the solver

deeply normalizing outside of the solver