[Duplicity-talk] A question on the duplicity remove-older-than option

[Joshua Bacher]

Hello all,

First of all i have to admit, i am totally new to duplicity, but 
what i have seen from it, makes me love this tool.

There is one feature that is not totally clear to me, and i would
really like to know how this is implemented. It is the 
remove-older-than  option that you may pass to duplicity.

I would like the know more about the mechanism behind this flag. 

The documentation says that this would lead to the deletion of backups
that are older then time specified.

I don't know if i get it right. If we have a file at time t1 that is
added to the repository (i call it like this since the whole backup
reminds me strongly on source code repositorys like cvs and svn)

So if we have this file it would be inserted completely in the first
insertion. I call this file f1_t1, meaning that this is the current
file. Every change will result in adding its diff to the repository,
right? i call the diff f1_d1 (the number indicates the incrementation
rank) So if i add some changes it goes like this:

f1_t1 -> f1_d1 -> f1_d2 -> .... -> f1_dn

How is the deletion of older files achieved? Just deleting the diffs
would be suicide. 
But the manual is somehow incomplete here, and i am really hot to know. 
Does duplicity just patches f1_t1 with all the diffs, or are several 
diffs merged to one big diff?

I had a short look on the source but it takes some time to get into it, 
could you help me understand the way how you achieve the deletion?

thank you anyway joshua bacher

signature.asc
Description: Digital signature