The tiles are grouped based on multiple servers, so the fact that your browser requested a specific tile doesn't directly tell them which site you actually visited.
Their last blog entry is from 18 May: https://blog.mozilla.org/
Edit: they have a different blog here that mentions it: https://blog.mozilla.org/futurereleases/
Why is it so hard to understand that leaking any bits not necessary for the retrieval and presentation of the link I clicked on is absolutely unacceptable? Anything that sends data that could not be learned form the server logs is spyware and will be treated as such.
Some companies like to claim that this spying is necessary because the data is useful. I'm sure it is, just like I'm sure a thief find the goods they stole to be useful. For similar reasons claims - like your explanation of this firefox misfeature - relating to the amount of bits leaked are not actualy a defense.