fix: adds support for function overloads

[renato-bohler]

Description (What)

This PR adds support for function overloads. Function overloads are not supported as arrow functions.

Fixes #17

Justification (Why)

Because otherwise the plugin will try to convert overloaded functions to arrow functions, breaking its intended behavior.

How Can This Be Tested?

1. Checkout this branch
2. Install dependencies
3. Run yarn test
4. The new entries in alwaysValid should pass on the test

[renato-bohler]

FYI: this implementation is missing support for exported overloaded functions. I'll add that soon.

// Currently working:
function foo(): any; // This node has type `TSDeclareFunction`
function foo() {}

// Missing implementation:
export function bar(): any; // This node has type `ExportNamedDeclaration` with a property `declaration` of type `TSDeclareFunction`
export function bar() {}

[renato-bohler]

FYI: this implementation is missing support for exported overloaded functions. I'll add that soon.

That's fixed with 0d1f620. I validated this change on a large codebase.

[JamieMason]

Thanks for all your work on this, will get to this when I can, maybe will be able to Sunday.

[JamieMason]

Merged in f60923a

[bensaufley]

Does this require a change to configuration? I don't see anything but I've updated to 3.3.2 and prefer-arrow-functions/prefer-arrow-functions is still tripping for me. Code looks like this:

export default function useGoalTracking(wait?: string, groupId?: string): GoalMapping;
export default function useGoalTracking(wait: string, groupIds: string[]): { [groupId: string]: GoalMapping };
export default function useGoalTracking(
  wait = 'goals.get',
  groupId: string | string[] | undefined = undefined,
): GoalMapping | { [groupId: string]: GoalMapping } {

Works as expected, except I'm currently having to insert eslint-disable-next-line prefer-arrow-functions/prefer-arrow-functions before the implementation.

Maybe I'm misunderstanding what this change was intended to do? But looking at the tests, I feel like the implication is that my code should work without tripping this rule.

[bensaufley]

Ope, of course as soon as I post this – it looks like export default is what's doing it.

[renato-bohler]

Sounds like another overload scenario I missed.

This should not report an issue.

Related: #31