Dhall-haskell: Marshalling HashMap

[sjakobi]

In analogy to #1391 and #1392, we could also have support for Interpreting and Injecting HashMaps from unordered-containers.

Not sure whether we should also add a newtype RecordHashMap for Injecting to Record.

[sjakobi]

Not sure whether we should also add a newtype RecordHashMap for Injecting to Record.

As pointed out in https://github.com/dhall-lang/dhall-haskell/issues/1392#issuecomment-540872043, Inject wouldn't currently allow this.

[tenniscp25]

I would love to work on this.

[Gabriel439]

@tenniscp25: Yeah, feel free to go for it! 🙂

[jiegillet]

Please check #1412, especially the tests, it should be very simple to add one for HashMaps

[tenniscp25]

Thanks @jiegillet. Your tests help a lot.

Btw, do I need to run tests with nix? I tried running stack test from master and got a few errors.

[jiegillet]

You don't need nix. Make sure that you have the latest version of stack first.

[tenniscp25]

I'm using stack 2.1.3, which to my understanding is the latest version. 🤔

[sjakobi]

@tenniscp25 What kind of errors are you seeing?

[jiegillet]

Indeed. What errors do you get?

[tenniscp25]

Sorry. It was my fault. I forgot to update the git submodule.. 😞

[tenniscp25]

Are these doctest errors expected?

https://gist.github.com/tenniscp25/7622e5d199ac1f9292c9fe1f3e664dab

[sjakobi]

Are these doctest errors expected?

The doctests in dhall are supposed to work – at least with stack. I believe we have some in dhall-lsp-server that are broken. @quasicomputational might be able to help with your errors.

For now I suggest you just use stack test dhall:tasty.

[Gabriel439]

Fixed by #1426