Facundo Domínguez
Facundo Domínguez
I cannot reproduce this in latest LH. It does require to move `seed` to the top level in order to give it a spec.
Maybe these could be archived: `template`, `liquid-base`, `liquid-ghc-prim`, `liquid-containers`, and `proposals`. @ranjitjhala or @nikivazou would need to do the job.
Thanks @mbazzani! This is almost fixing #2285. How about adding some test, and using `div` instead of `/` in the spec?
Superseeded by #2339.
[This bit](https://gitlab.haskell.org/ghc/ghc/-/blob/7a08a025eecb0cca494757fc83099a9decb052ad/compiler/GHC/Types/Name/Occurrence.hs#L408) in the GHC source code could be helpful when deciding how to fix it.
With the latest LH I get this error ``` : error: Uh oh. Test.hs:6:13 Cannot lift Haskell function `foo` to logic Cannot transform lambda abstraction to Logic: \ (ds_d21p ::...
I updated the description to make explicit that at the moment we only want to rebase #719 on top of master to merge it.
I updated the description after rebasing #719 and checking the specification of the eras after Shelley.
It would be possible to transform the source code before feeding it to the desugarer, and change the newtypes to data. This could break code that uses `Data.Coerce.coerce`, but perhaps...
> which causes trouble down the line when we want to interop with SMT Bag theory. Which trouble is caused exactly? The specs would still refer to Bag. Perhaps LH...