`Text` vs. `as Text`

[dhall-lang]

`Text` vs. `as Text`

[Gabriella439]

This limitation is intentional. Imported Dhall expressions (such as ./template) have to be closed (i.e. no free variables like value in your example) so that they can be type-checked in isolation.

[angerman]

Is there a way around it? E.g. to inject some textual value verbatim prior to the typechecker?

I‘d rather not have a \(...) -> '' first and '' last line in say a HTML file.

[Gabriella439]

Yes, but only if you're willing to use the Haskell API to build your own tool. The Haskell API provides a typeWith function that you can use to type-check expressions using your own custom context. Similarly, there is a normalizeWith that you can use to inject custom values

[angerman]

Could dhall-to-text gain a primitive for that, hypothetically speaking?

[Gabriella439]

Yeah, I'd be fine with providing a utility function that bundles typeWith+normalizeWith+{adding single quotes before parsing}

[asivitz]

I would love this. I was thinking about making a Dhall compiler for Hakyll, and would probably end up implementing this myself if it weren't in the Dhall lib.

[f-f]

Should we move this issue to dhall-haskell?
As far as I understand it is about adding a flag (--wrap-text?) to dhall-to-text

[Gabriella439]

@f-f: Yes, my understanding is that this is a dhall-haskell-related issue, although I think the request here is for a new function in the Haskell API for ease of templating rather than a new command-line option.

[Nadrieril]

Having the ability to import templates from files sounds cool though.
I imagine the following: if file contains "A ${string} with ${interpolations}", then (./file as Template) { string = "blah", interpolations = "blih" } would normalize to the string "A blah with blih". That would allow big templates to be put separately, while still requiring imported expressions to be closed.

[Gabriella439]

@Nadrieril: See the discussion on the linked pull request:

dhall-lang/dhall-haskell#205

... and also the discussion on this related issue:

dhall-lang/dhall-haskell#676

[Nadrieril]

Oh. Too bad, I liked this idea, but you make convincing points

[ocharles]

dhall-lang/dhall-haskell#205 says

I'll go ahead and close this since it's unlikely to be added to the language standard

Should this issue also be closed?

[Gabriella439]

Yeah, I will close this issue, too