fix(types): enhance reusability of pres

[IJuanI]

Support for pre handlers whose type definition is only a subset of the actual route type definition.

When trying to reuse pre handlers in a typescript environment like the following

const doSomething: Lifecycle.Method<any> = async ({ params }) => {
  console.log(params.preParam);
};

server.route<{ Params: { param: string, preParam: number } }>({
    method: 'GET',
    path: '/route',
    handler: async ({ params }) => console.log(params.param),
    options: {
        pre: [ doSomething ]
    }
});

server.route<{ Params: { otherParam: string, preParam: number } }>({
    method: 'GET',
    path: '/route',
    handler: async ({ params }) => console.log(params.otherParam),
    options: {
        pre: [ doSomething ]
    }
});

The type definition of the pre handler does not have any possible valid value, hence needing to drop type checking using the any keyword in order for typescript not to throw an error.

This happens because the definition of Lifecycle.Method expects the request type to equal to the type provided in the Method generic, but in this scenario there're 2 distinct request types.

Note: Using unions of both request types is also invalid, I would need to dig deeper into why, but seems related to the usege of the keyof operator in another type definitions.

The changes in this PR update the Method type definition to require the request type to extend the type provided in the generic instead of checking for it to be equal.

This change allows to resolve the above scenario with type definitions like the following:

const doSomething: Lifecycle.Method<{ Params: { preParam: number } }> = async ({ params }) => {
  console.log(params.preParam);
};

As all route definitions contain a preParam: number param, they're successfully extending the Method's generic, passing type checking

[Marsup]

Thanks! Sorry this went under the radar for so long, I'm releasing it now.

[Marsup]

Sorry, I had to revert it. It's breaking the final types of the request.

[IJuanI]

Hey @Marsup, thanks for taking a look!

This patch has been running in my company for almost a year with no typing issues. Any chance you could either re-run the failing CI, share the error you found, or provide testing instructions?

Thank you!