regarding tests: tests\fsharpqa\Source\Warnings\env.lst is now a dumping ground for tests that check error messages. I think it would be ok to add it there

thanks @forki, I'll pay a look and try to add few tests supposed to trigger that new message.

Mmmh reading details @dsyme left in the ancient conversation and my proposed fix is not really handling things at the level he was intending the changes to happen but later down the road, although it looks simpler (if that doesn't raise false positives), maybe it doesn't follow the normal conventions for deciding of errors, but there are few other conditionals to decide which error message to raise where I've added the logic.

To add a better error message, the logical place to put it would be here: https://github.com/fsharp/fsharp/blob/6ac6318858da25fe4f1d4728901c5824e9f3bf3a/src/fsharp/typrelns.fs#L1191

"IsExactMatch" would need to be duplicated and adjusted to a "IsExactMatchUpToInstantiatedUnitReturnType". If an override exists satisfying that predicate, then a better error message can then be given, much as for code already present to give a better error message when there is an override satisfying IsPartialMatch or IsNameMatch.

I'm trying to see if other cases would trigger the initial / new error messages, this compiles fine with and without the change:

type Foo<'t> =
  abstract member Bar : 't -> int

type Bar() =
  interface Foo<unit> with
    member x.Bar _ = 1

I'll look if fsharpqa also has tests checking no error / warning shows up to add that case.

If anyone knows of additional cases I should account for, that would be great.

@KevinRansom or @dsyme could we get a review on proposed fix and some direction for the error message?

@dsyme
This looks good to me, but I may be missing something, could you take a look please.

Kevin

Thank you for taking care of this

Kevin