Re: [Duplicity-talk] Integrity check

[edgar . soldin]

> But you would catch that during a verify run, wouldn't you? In essence you
>    

yeah, but with unnecessary computing and too much output.

> simply propose to leave away the last step of verify, right? Should be simple
> enough to implement...
>    

exactly my point ...

> Decrypting it on the remote really is silly but gpg --verify would have its
> uses.
>    

unfortunately my tests showed that a encrypted signed gpg file can't be 
verified by 'gpg --verify'. This only works for detached signatures. You 
actually have to decrypt it for that purpose. And therefor you will need 
your most valuable gpg data (keys,passphrases) on the remote machine. 
Again. I wouldn't put them on a non trusted system.

> It may not necessarily a slow rather than an expensive line. Think non-flatrate
>    

The reason why I started thinking of it was ... Let's say I want to 
remove the incrementals of the last month and start with a fresh set of 
incrementals against my old full. Don't I want to make sure that this 
old full is still intact? So actually I think integrity checking is done 
seldom but needed in some cases.


... ede