Fix shorthand identifier import usage

[jondlm]

↪️ Pull Request

Fixes #9140 by adding a new SWC visitor to the JS transformer. It visits all identifiers, checks if their IDs match any of the previously detected imports. If so it adds them to the used_imports set.

💻 Examples

See #9140 for an outline of the bug this fixes.

🚨 Test instructions

I added an integration test that fails before this change and passes after. I couldn't find unit tests for the JS transformer so I'm assuming an integration test is appropriate here. I used @lettertwo's new fsFixture helper to colocate the test files with the actual test. That worked quite nicely!

✔️ PR Todo

• Added/updated unit tests for this change
• Filled out test instructions (In case there aren't any unit tests)
• Included links to related issues/PRs

[jondlm]

I couldn't think of any cases where this new visitor would be problematic since SWC seems to imply that checking the id here helps us uniquely identify the correct use accounting for scope. Let me know if I'm wrong! I am very new to using Rust with SWC.

[jondlm]

Perhaps @mischnic could take a look at this one when you have time?

[mischnic]

Seems like object shorthand is one of the few cases where there's an actual identifier that isn't/can't be an expression (all of the other cases are handled by visit_expr)