transferSize and TAO redirects about resource-timing HOT 2 CLOSED

Re, "last": yes, I think we could drop that but I'm not sure it's a big deal. For RT{1,2} we're not exposing each individual fetch so I think there is still value in clarifying this behavior. Once we address #21, I think it'll be definitely redundant.

Re, TAO-enabled redirects: if memory serves, I think the complication was with respect to what happens when some redirects have the opt-in and some don't.. because we don't expose which is which it becomes hard to reason about the returned number; so we opted to do all-or-nothing.

from resource-timing.

Re-reading the spec.. Re, "last": I believe this is fine, we're just being explicit about allowing transferSize to be exposed when the last non-redirected fetch passes TAO; this language is consistent with the rest of the spec.

That said, I'm not sure why we restricted the redirect chain case to same-origin only.. The check that we want is that all redirects have TAO opt-in -- i.e. same as all other attributes. I'll push a change to fix that.

from resource-timing.