Question regarding ordering for rename requests

I am developing a new language server and I'm having issues understanding how multiple rename requests should be handled and in which order.

Suppose the client sends a rename request and the server starts computing the results, while it's computing, the client sends another rename request, in this case, what should do the server do?

• Should it try to fullfill the first rename request and then process the second request such that the requests are queued at the server?
• Should the request be cancelled? In tha case: Who should cancel the request? and which of the two request should be cancelled?

Note that the same questions can be asked about all requests that send a WorkspaceEdit as a result.

Nell’issue non sono indicati file, test o punti di ingresso. Inizia esaminando le indicazioni del Language Server Protocol su request-cancellation e WorkspaceEdit, quindi chiarisci l’ordine previsto o il comportamento di annullamento per le richieste di Rename sovrapposte e se la stessa regola si applica anche alle altre richieste che producono WorkspaceEdit.