On Fri, Feb 20, 2015 at 4:09 PM, Daniel Micay <danielmicay@xxxxxxxxx> wrote: > On 20/02/15 10:04 AM, Martti Kühne wrote: > > You should really just tell upstream to sign their releases, because it > wipes out the attack vector instead of just making it possible to audit > whether a MITM attack on the original. packager occurred like hashes. > > The hashes provide no security for the initial packaging work and no > defense against an attack that's done by compromising the upstream > sources, which is far more realistic than a targeted MITM attack on a > specific Arch Linux packager. > Since security-by-reading-the-code isn't an option since a few decades, we're stuck in this, yup. cheers! mar77i