Las
https://hackage.haskell.org/package/generically ?
I've realised that there is an issue with @shepmaster's code: it `Display`s the error even if the error never needs to be displayed. This is probably not that common, but...
@shepmaster an error occurring doesn't mean you have to display it.
It's not very relevant when you're using strings as errors, but when you have an actual error type, you can't take ownership of the error or refer to it, so...
*one* derivation. The RFC elaborates more on the approach.
It seems like this is a duplicate of #231?
Sorry for the late answer! One use case would be exporting a haskell.nix module in one flake, that you then expect downstream users to include in their call to `cabalProject'`...
We could do that, but that would mean downstream has to manually keep it in sync with upstream, as is often the unfortunate case with Plutus.
I didn't realise the input to `cabalProject'` was also a module when making this issue.