Backup with diff file instead of full copy to save disk space?

[QiangF]

It should work like this.

1. sort backup files by suffix, filename-number-(max) filename-number-(max-1) ...... filename-number-min
2. let filename-number-(max) be a full copy, and all other files be diff files only
3. on creating a backup, save the copy to filename-number-(max+1), calculate the diff between filename-number-(max) and filename-number-(max+1) and save it to filename-number-(max)
4. on walking back, apply the diffs sequencially