I don't know if gpg stores any kind of date as part of the metadata of the 
encrypted content but encrypting a timestamp seems like the only resiliant 
approach. Any external mean is likely to fail in some situation.

The timestamp could be as simple as a date 20-05-2016. Unless having hour or 
even minute resolution is a requirement.

The only possible issue is that given some of the content being encrypted is 
known, I don't know if that makes the encryption less secure. In modern 
cryptography it shouldn't but I'm not an expert.

Renato

-------- Original Message --------
From:Grégoire Détrez <[email protected]>
Sent:Fri, 20 May 2016 12:10:18 +0200
To:Evgenii Sovetkin <[email protected]>
Cc:password-store <[email protected]>
Subject:Re: [pass] list of passwords with its age

>On Fri, May 20, 2016 at 11:46:08AM +0200, Evgenii Sovetkin wrote:
>> On 2016-05-20 11:33 AM, Grégoire Détrez wrote:
>> > On Fri, May 20, 2016 at 10:45:46AM +0200, Evgenii Sovetkin wrote:
>> > > I want to obtain a sorted list of my passwords together with its age
>> > > (number of days since the last change).
>> > > 
>> > > With git it seems the following loop will do part of the job:
>> > > 
>> > > git ls-tree -r --name-only HEAD | while read filename
>> > > do
>> > >     echo "$(git log -1 --format="%ad" --date=relative -- $filename) 
>> > > $filename"
>> > > done
>> > > 
>> > > Unfortunately, if I rename a password (or even rerun git init) then it
>> > > would seem that I have changed the passwords, though I have not...
>> > > 
>> > > Is there a neater solution one have in mind? Is there a way in git to
>> > > show time passed from the *significant* content change, not simply the
>> > > latest commit?
>> > 
>> > You could use use git-blame to get the last commit that changed the
>> > first line of each file, something like:
>> > 
>> >     git blame -L 1,1 $filename --porcelain | sed -n 's/^committer-time //p'
>> > 
>> > 
>> > /ǵ
>> 
>> Thanks for the suggestion! It solves the problem with the renamed
>> passwords, but it doesn't really work, with the reencrypted passwords
>> (I rerun git-init recently).
>
>I guess you mean pass-init?  I far as I understood, as long as you can
>still decrypt the files, git-blame should be robust to re-encrypting
>password.  You need to properly setup the git attributes to compute
>clear-text diff (it seems that the latest versions of pass do that
>automatically otherwise I described how configure it manually in an
>older thread on this list).
>
>
>/ǵ
>
>_______________________________________________
>Password-Store mailing list
>[email protected]
>http://lists.zx2c4.com/mailman/listinfo/password-store
_______________________________________________
Password-Store mailing list
[email protected]
http://lists.zx2c4.com/mailman/listinfo/password-store

Reply via email to