Can I add language support to harper at runtime?

[andremarianiello]

If I wanted to use harper on Haskell files, is there a way to integrate harper with the tree-sitter-haskell grammar at runtime? Perhaps I could point harper to my installation of the parser for dynamic loading? If not, what is the process of getting language support statically added to harper?

[grantlemons]

I don't think it's possible to integrate it at runtime, but adding support for parsing comments should be pretty trivial since the framework already exists for that based on tree-sitter.

[elijah-potter]

Hey @andremarianiello, we've added very rough initial support for Haskell. Give it whirl and let us know if there are any issues.

@grantlemons, would you mind looking into whether structured comment support as described here is applicable to Haskell?

[grantlemons]

Hey @andremarianiello, we've added very rough initial support for Haskell. Give it whirl and let us know if there are any issues.

@grantlemons, would you mind looking into whether structured comment support as described here is applicable to Haskell?

Are you sure this is the right link? I don't see anything about structured comment support.

[andremarianiello]

My very basic test with Haskell worked nicely, and will definitely be useful going forward in my Haskell projects. Unfortunately I wasn't specific enough in my initial description, because I initially chose harper to help me grammar check blog posts that I write with literate haskell. As expected, I get no hints from harper on literate haskell (.lhs) files. Would you accept a PR that adds literate haskell support? I saw it was removed in grantlemons@c100c6d

EDIT: Orthogonal idea: if harper supported analyzing specific ranges of text (does it?), I could configure my neovim lsp client to send only ranges with the @spell capture group to harper. Then harper wouldn't have to do any parsing at all, just grammar checking. And it would "support" all languages that are sufficiently supported by my client.

[grantlemons]

My very basic test with Haskell worked nicely, and will definitely be useful going forward in my Haskell projects. Unfortunately I wasn't specific enough in my initial description, because I initially chose harper to help me grammar check blog posts that I write with literate haskell. As expected, I get no hints from harper on literate haskell (.lhs) files. Would you accept a PR that adds literate haskell support? I saw it was removed in grantlemons@c100c6d

EDIT: Orthogonal idea: if harper supported analyzing specific ranges of text (does it?), I could configure my neovim lsp client to send only ranges with the @spell capture group to harper. Then harper wouldn't have to do any parsing at all, just grammar checking. And it would "support" all languages that are sufficiently supported by my client.

It shouldn't be too difficult to support literate haskell, I just took it out because I wasn't sure how our comment parsing would mesh with how tree-sitter-haskell views literate haskell (since it's effectively swapped). I didn't actually test though, so there might be an easy fix already.

As for the @spell capture group, that's a bit of a bandaid, and I don't think we really want to add any specific support for things like that, but it sounds feasable just on your end. Best of luck!