Hi @pnkfelix, I'm looking at #62212
The presence of inference variables makes me suspicious that something bad is going there
do you know why they appear?
I assume it’s an artifact of projecting the associated item, but I don’t really know beyond that
I'll like to look at it. Hope I'll have some time tomorrow.
hmm, interesting: normalization itself does not attempt to resolve type variables?
are they considered orthogonal concerns? Do we at least ensure that type vars never map to an unnormalized type form?