RFC: Potential fix for #4808

[hvr]

See #4808 (comment)

Please include the following checklist in your PR:

• Patches conform to the coding conventions.
• Any changes that could be relevant to users have been recorded in the changelog.
• The documentation has been updated, if necessary.
• If the change is docs-only, [ci skip] is used to avoid triggering the build bots.

Please also shortly describe how you tested your change. Bonus points for added tests!

[phadej]

Test would be great.

I can only shake my head around this. If only we could have versioned APIs to call older Setup.hs.

[23Skidoo]

Looks OK, but I'd like to see a test. We can merge it as-is to 2.0 to get the point release out faster, and leave the PR for master open.

[hvr]

@23Skidoo I'm not sure yet how such a test would look like; we can probably do easy roundtrip tests (and it would help if FlagAssignment was a newtype, so we could attach parse/display instances to it).

(relatedly, @ezyang mentioned in #4710 (comment) we don't have many communication tests with different lib:Cabal versions)

So which kind of tests did you (& @phadej) have in mind specifically?

[bgamari]

What is the status of this? For the record, this is something that 8.2.2 is blocking on.

[phadej]

@hvr, I'm ok for this going to 2.0 as is, but for master could we change FlagAssignment to the newtype and do roundtrip stuff?

[phadej]

See my comment.

[phadej]

cherry-picked on 2.0. @bgamari could you verify it unblocks GHC-8.2.2?

[hvr]

but for master could we change FlagAssignment to the newtype and do roundtrip stuff?

oh yes; in fact I have already started on refactoring it into a newtype (see #4849) to see how much places I'd have to touch; it's not bad at all :-)

[bgamari]

Thanks @phadej. That should do it.