I have gone back-and-forth about this pull request and decided to abandon it. Using a non-repeatable payload has a number of drawbacks, either buffering on-disk as in this commit or not calculating the length and MD5. Instead of this approach, we should calculate the MD5 via `HashingInputStream` and check against the eTag after `putBlob` completes. Note that aws-s3 requires a Content-Length but I do not believe we should try to support this via buffering.
--- Reply to this email directly or view it on GitHub: https://github.com/jclouds/jclouds-karaf/pull/23#issuecomment-49090874
