Merge diffs so no new complete backup is required

On 15.06.2014 23:02, gamana wrote:
> New question #250273 on Duplicity:
> https://answers.launchpad.net/duplicity/+question/250273
>
> Why isn't there a feature in duplicity that lets you merge the chain of diffs into one backup and start the new diffs based on that? The integrity could be checked using checksums.
> This feature would be especially helpful for a slow connection. Just having a backup of all my data lasts 3 months and I don't want to start a complete new backup a short time later.
>

how about nobody hacked it so far?

also it wouldn't really make sense. usually backups are stored remotely and "merging" would mean to redownload all volumes, merge locally, reupload. duplicity strives to be bandwidth efficient.

btw. currently you can reclaim space by keeping fulls of chains which works pretty good.

..ede/duply.net

I'm currently looking at Duplicity to replace rdiff-backup, and this is also my concern. I know I will want to remove old backups after a while, and having to make regular full backups is a less favorable option. It would be great if duplicity worked like rdiff-backup in this regard - keeping the newest backup as the full one and make diffs for the old ones.