Selective commits can lead to incorrect backups

[FinnWoelm]

When selectively applying changes during new commit, the backup created can be incorrect. A new version will be created, which triggers a new backup request as opposed to taking any existing backup.

Here is what should be done: When selectively committing changes, a backup will definitely already exist for the file. It's just that the backup may not have the right name.

For example, we have two backups:

• version 1 named name 1
• version 2 named name 2

On a selective commit, the user may want version 1 named name 2 OR version 2 named name 1 — neither of which exist in the archive.

Therefore, when creating a new version, ...

• check if backup with same name & content already exists. If so, reuse that.
• check if backup with content already exists (but different time). If so, duplicate that content and assign the new name.

[FinnWoelm]

Side note: On selective commit, we should actually disable the ability to create any new backups of remote. It is never needed and if it's happening, then something is going wrong.