Great! Some remarks:

1. I am not sure this really closes Type spread failed on empty variant #6952, as there is still the question of how to create an empty variant type (Type spread failed on empty variant #6952 (comment))
2. Can you undo the reformatting of the CHANGELOG? If we want to reformat it, we should do it in a separate PR, and I think the reformatting is partially incorrect.
3. Why is release.ninja changed? Does the change go away when you re-run make lib?

Great! Some remarks:

1. I am not sure this really closes Type spread failed on empty variant #6952, as there is still the question of how to create an empty variant type (Type spread failed on empty variant #6952 (comment))

This PR disallows using extensible variants.

2. Can you undo the reformatting of the CHANGELOG? If we want to reformat it, we should do it in a separate PR, and I think the reformatting is partially incorrect.

Yes, sorry, didn't see I reformatted it.

3. Why is release.ninja changed? Does the change go away when you re-run make lib?

No idea why it changed. I'll try re-running that and see what happens.

This PR disallows using extensible variants.

Yes, but @glennsl also wrote

In OCaml, an empty variant type would be type a = |. That doesn't seem to be supported in Rescript though.

Do we want a syntax for an empty variant type in ReScript? If so, should we open a separate issue for that?

This PR disallows using extensible variants.

Yes, but @glennsl also wrote

In OCaml, an empty variant type would be type a = |. That doesn't seem to be supported in Rescript though.

Do we want a syntax for an empty variant type in ReScript? If so, should we open a separate issue for that?

Oh, right. Feels like a separate topic to this, so yes, a separate issue would be good.

@cknitt seems to have done it. Would you look again?

Now all the format changes are back.

@cknitt check again.