Refactor text() et al to share more infrastructure with Fetch

Andrea implemented these methods in Firefox and ended up using a shared abstraction with our code for Fetch.

If I look at https://fetch.spec.whatwg.org/#concept-body-consume-body and https://w3c.github.io/FileAPI/#text-method-algo side-by-side it seems quite doable to have less duplication here.

(We might want to point out the one major difference though, which is that File API will create an endless series of streams for the same underlying data where in Fetch that'd require explicit cloning.)

Lee el algoritmo de Fetch «concept-body-consume-body» junto con el «text method algorithm» de File API enlazado en la issue, comenzando por los pasos compartidos y sus supuestos sobre el manejo de streams. El trabajo estará terminado cuando text() y los métodos relacionados reutilicen una infraestructura más común, preservando el comportamiento de File API para streams repetidos en lugar de requerir un cloning explícito.