Provide more accurate rename range in prepareRename LSP response

[auvred]

Strada uses node.getStart, while Corsa uses node.Pos. This causes the textDocument/prepareRename method to return a range with leading trivia. As a result, that range doesn't match the one later returned returned by textDocument/rename.

[Jake Bailey (jakebailey)]

This seems fine but, no tests change?

[auvred]

This seems fine but, no tests change?

TextDocumentPrepareRenameInfo is used in fourslash only in VerifyRenameSucceeded and VerifyRenameFailed, and neither validates the response it returns:

typescript-go/internal/fourslash/fourslash.go

Lines 4297 to 4325 in 1de8d68

	func (f *FourslashTest) VerifyRenameSucceeded(t *testing.T, preferences *lsutil.UserPreferences) {
	if preferences != nil {
	defer f.ConfigureWithReset(t, *preferences)()
	}
	params := &lsproto.PrepareRenameParams{
	TextDocument: lsproto.TextDocumentIdentifier{
	Uri: lsconv.FileNameToDocumentURI(f.activeFilename),
	},
	Position: f.currentCaretPosition,
	}
	prefix := f.getCurrentPositionPrefix()
	result := sendRequest(t, f, lsproto.TextDocumentPrepareRenameInfo, params)
	if result.Range == nil && result.PrepareRenamePlaceholder == nil && result.PrepareRenameDefaultBehavior == nil {
	t.Fatal(prefix + "Expected rename to succeed, but prepareRename returned null")
	}
	// Also verify that textDocument/rename produces edits, since prepareRename is optional.
	renameResult := sendRequest(t, f, lsproto.TextDocumentRenameInfo, &lsproto.RenameParams{
	TextDocument: lsproto.TextDocumentIdentifier{
	Uri: lsconv.FileNameToDocumentURI(f.activeFilename),
	},
	Position: f.currentCaretPosition,
	NewName: "RENAME_SUCCEEDED_TEST",
	})
	if renameResult.WorkspaceEdit == nil || renameResult.WorkspaceEdit.Changes == nil || len(*renameResult.WorkspaceEdit.Changes) == 0 {
	t.Fatal(prefix + "prepareRename succeeded but textDocument/rename returned no changes")
	}
	}

Should I update them to save the results to the baseline?

[Jake Bailey (jakebailey)]

It would be good, though, I'm not sure how that plays into our existing diffing. Probably it's not worth it, I can see that this is obviously correct.