image download doesn't resume on failure

Image files are relatively stable, and are downloaded via HTTP. It would be very useful if a subsequent download attempt started from where the last one failed, rather than starting over again from 0%. I believe that's most often done using the `Range:` header.

Start by locating the HTTP code that downloads image files and inspect how a failed download is retried. Check whether the downloader already tracks partial files or response metadata, then verify behavior with a failed transfer followed by a retry. Done means a subsequent attempt requests and preserves the already-downloaded portion instead of restarting at 0%.

Written by the indexing model from the issue text.