HEAD Fallback Mechanism to GET 0-0

[robertvanloenhout]

Replayweb.page does a HEAD request when it opens a WACZ file. If the HEAD request fails it will do a GET request with range 0-, thus requesting all bytes.
We would like it to retry first with a GET request with range 0-0.
The reason for this is that we would like to enable caching on AWS CloudFront and have WACZ files larger than 30GB. A HEAD request and a GET request for all bytes for these files will fail, but a GET range request for some bytes will succeed.