Right now we don't normalize impl bounds during method probing, which means that the winnow stage may not work great when those bounds involve associated types. I think with the new setup we should be able to do this but didn't have time to test. Search for the FIXME in probe.rs
Right now we don't normalize impl bounds during method probing, which means that the winnow stage may not work great when those bounds involve associated types. I think with the new setup we should be able to do this but didn't have time to test. Search for the FIXME in
probe.rs