This is a great improvement - I've definitely run into issues before where you'd have a model file like:

User
    profileName Text

UserProfile
    name Text

and it would fail to compile.

Thanks for the PR!

Would you be willing to add a test module here? Mostly so we can avoid breaking this with a regression in the future

Probably something like:

#if __GLASGOW_HASKELL__ >= 902
{-# LANGUAGE NoFieldSelectors #-}
{-# LANGUAGE DuplicateRecordFields #-}
#endif    

module NoFieldSelectorTest where

#if __GLASGOW_HASKELL__ >= 902

import ...

mkPersist [persistLowerCase|

something that triggers a bug

|]

spec :: Spec
spec = it "compiles" True

#else

spec :: Spec 
spec = do
    it "only works with GHC 9.2 or greater" $ do
        pendingWith "only works with GHC 9.2 or greater"    

#endif

Added (using your example almost as-is). Unfortunately it turns out that disambiguating fields isn't enough to solve the User vs UserProfile problem because it also defines type constructors (UserProfile).

Alas! Good point on that 😄

I'll get this merged and released once CI is done. This should be a patch-level bump.

Released as persistent-2.13.3.4, thank you!

Thanks! This is a big step for my ambition for greater records!