odersky

The way it is written it's clearly a cycle. The RHS needs to be type-checked, which requires typing `this`, which requires typing the self type `Context`. I believe #20251 would...

I guess the question is why does it need the base classes of Context when running the namer on TyperOps? Maybe the next nightly will give more info. Could it...

reaches.scala still has errors?

It's just a single test that fails, and looking at the history we never really understood why it works now and failed previously. So it's worth digging some more here....

> Intersection types should affect compilation times with no more then 20% overhead. Intersection types can give a worst case exponential blowup of compilation times. So this is a nice...

test performance please

(let's do it again, to see whether this is a fluke or a trend)

> If we always update the lastDenotation then we lose the denotation in the truncated, tab-completion run (we replace an outdated denotation with NoDenotation). Yes, but I am not sure...

Right. But then I wonder why that was not caught in the test case?

Ah, I only ran testCompilation and that does not include the repl tests. How does it work for you with the latest patch? (I force pushed with a fix not...