[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Gnu-arch-users] Re: signatures and checking
From: |
Tom Lord |
Subject: |
[Gnu-arch-users] Re: signatures and checking |
Date: |
Mon, 26 Jan 2004 17:56:59 -0800 (PST) |
> From: Tom Lord <address@hidden>
> > From: Miles Bader <address@hidden>
> > Surely you can just treat the two cases separately:
> > Ask gpg (or whatever) to deliver the contents in case (2), by
having a
> > `.arch-params/signing/*.contents' script, and just parse them
directly in
> > case (1) (which is when you _don't_ have the script).
> > This would do the right thing, be simple to implement, not require
any
> > awk scripts, and not require any new options to gpg....
> Interesting idea but no.
> That would create a partial exploit in which non .check'ing clients
> saw different checksum data from .check'ing clients.
> I don't immediately see any obvious way to turn that partial exploit
> into a complete one --- but it seems needlessly fragile, nevertheless.
I suppose one way to do essentially that safely would be:
If there's a .check file, parse the checksum twice -- once raw and
once as the output of .check. Make sure they agree.
But, really -- I think it's simpler just to toss a script into the
distro and update the HOWTO while waiting for gpg to get fixed.
-t
Re: [Gnu-arch-users] signatures and checking, Andrew Suffield, 2004/01/26
Re: [Gnu-arch-users] signatures and checking, Robert Collins, 2004/01/27