How do i make harper understand capital case is actually multiple words?

[redsuperbat]

Id like harper-ls to understand MasonLspconfigSettings as Mason Lspconfig Settings instead of a single word. Like it does with camel case and snake case. How can I configure this?

[jessevdp]

+1 — I'm also looking for this.

I've recently started using Harper to spell check comments in my code. I couldn't find a setting to make it respect CamelCasing.

To add an example.. I might have a comment like the following:

# rubocop:disable Style/DataInheritance

Where harper will complain:

Did you mean to spell “DataInheritance” this way?

[elijah-potter]

I would argue that this is not a spell check problem, but rather a masking problem. It looks like you're using a structured comment, which Harper should ignore (and does for Javadoc, JSDoc, etc.), but isn't.

As someone more familiar with Ruby than I am, is there a spec for those doc comments?

[hippietrail]

I would argue that this is not a spell check problem, but rather a masking problem. It looks like you're using a structured comment, which Harper should ignore (and does for Javadoc, JSDoc, etc.), but isn't.

As someone more familiar with Ruby than I am, is there a spec for those doc comments?

I think they're trying to say that they're expecting "DataInheritance" to be spellchecked as if it were "data inheritance" by first converting camelcase into words that can be spellchecked individually.

But their choice of example happens to resemble an unrelated recent masking feature we added that turns off linting for an item.

[jessevdp]

I agree that my example is perhaps a bit misleading. I suppose it's a structured comment of sorts, though I don't think Ruby necessarily has a standard for those.

There are plenty of other instances where I might write documentation in code comments and mention a class name that's camel cased where I'd like it to spellcheck the individual words.

"DataInheritance" -> "data inheritance"

[hippietrail]

I seem to remember when reading the original paper on Harper (I think) a reference to dealing with identifiers. But I don't remember ever coming across such a concept in dealing with the codebase or using the VS Code extension.

Is there actually support for identifiers?

I'm guessing that we'd mostly track them so as to not cause distraction by flagging mentions of them in comments since they're often not directly English words but artificial abbreviations or compounds etc. But an option to turn on spellchecking would also be helpful. I've definitely done PRs on projects for misspelled words in identifiers before.