George Thomas

Actually, this should be an equivalent definition: ```hs inputSafe :: Dhall.FromDhall a => Dhall.Decoder a -> Text -> IO (Maybe a) inputSafe decoder t = handle (\(e :: Dhall.ExtractErrors Dhall.Src...

@Gabriel439 Ooh, I like that a lot, especially if there's precedent.

Hmm. Github's syntax highlighting appears to be slightly broken - wrong colour for the opening bracket in `2.dhall`.

Ah, thanks, that suddenly makes a lot of sense. Is there any particular reason for it to be the way it currently is? Keeping normalisation separate seems preferable on the...

@Gabriel439 That would be great! There's still a potential blowup from diamond dependencies. But that seems like something I can experiment with downstream. e.g. introducing a new let-binding with a...

Has there been any progress towards this? It's still my main frustration with Dhall. (I don't like to pester maintainers - I'm just genuinely curious whether there are any threads...

> @georgefst: Yes, I plan to fix the diamond dependency issue, too. The idea is that if an import is protected by an integrity check then we don't have to...

This is probably my fault. I'd never even thought to test it.

> Yeah, I think the real problem here is that `pretty-simple` doesn't have any tests checking the whole laziness thing! We have at least one! I remember from https://github.com/cdepillabout/pretty-simple/pull/67#issuecomment-651701234.

Hmm, yeah, I would like to get away from using a custom setup if it's at all possible. There are a lot of issues around them, and this isn't the...