Remove variables only used by internal directives

[alexlafroscia]

This PR removes the variables only used by internal directives. See the related issue for more details about why this is important.

Closes #308

[changeset-bot]

🦋 Changeset detected

Latest commit: 840803c

The changes in this PR will be included in the next version bump.

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[AlecAivazis]

Thanks for getting this out so quickly!

[alexlafroscia]

@AlecAivazis is there a reason this project is using such an old version of TypeScript? I am running into a problem with the code I am refactoring to, where the TypeScript version provided by my editor's language tools are OK with the code (since it defaults to TypeScript 4.6) but the tests are failing because the version used by the project (3.9.5) doesn't seem to like it.

On 4.6 the instanceof check narrows the type correctly

On 3.9.5 it doesn't

Unfortunately I can't use Array.isArray here due to microsoft/TypeScript#17002 (I think; for some reason Array.isArray doesn't narrow the type to remove the possibility that parent is an Array, and since the type from graphql is possibly a readonly Array, that seems likely to be the reason why)

[AlecAivazis]

Like I mentioned in the other PR, I am more than happy to update the typescript version if we can get the build to pass. Feel free to update it here instead of a separate PR if you prefer that way you don't have to juggle too many things going in at once

[alexlafroscia]

@AlecAivazis gotcha! Juggling a bunch of branches doesn't matter to me; keeping things separate is fine by me.

In case I can't figure out why the TypeScript version bump breaks CI, feel free to merge this PR without the TypeScript update if that's what you want to do! I would love to get a Houdini version published with this bug fix in it so I can fix my app at work 😬

[alexlafroscia]

Since I was able to get #312 passing, if you're happy with both PRs, let's merge that one first; I'll then rebase this one on main again to I can remove the silly @ts-expect-error comment that's currently necessary for this code to work correctly.

[AlecAivazis]

@alexlafroscia its merged now! This should be good to go once you rebase and we get the 👍 from github

[alexlafroscia]

@AlecAivazis I rebased the PR on main -- looks like we're all set here!

[AlecAivazis]

Looks good! Thanks for adding a test 👍