Tools with signatures

John Meinel john at arbash-meinel.com
Mon Apr 8 13:10:41 UTC 2013


If the juju tool only knows the "official juju" public key. How does that
not validate the tools are from a trusted source? It doesn't help the
--upload-tools case, but those are put in your private bucket anyway.

It is true that we shouldn't trust anything that is signed. Though I'll
note xkcd: http://xkcd.com/1181/

We could trust a specific CA or web or whatever, but why not just have a
key that is needed to release the tools?

John
=:->
On Apr 8, 2013 7:08 AM, "Tim Penhey" <tim.penhey at canonical.com> wrote:

> Hi all,
>
> I have a task card that I don't feel confident starting without more
> understanding of the problem, and how the solution would help with that.
> And how the tool release process would work with this too.
>
> Firstly, which is the problem that we are trying to solve?
>  * We have an unmodified tarball as it was defined at the source
>  * The tools are from a trusted source
>
> The traditional method for the first is to provide an md5 (or other)
> hash that you can grab as well as the source file.  However if we are
> trying to solve the second problem, this doesn't help.
>
> So my first question really is what problem are we trying to solve?
>
> How will this differ between tools from a public bucket, and those that
> are uploaded locally?
>
> Who is considered trusted?  How would we sign the tools to ensure that
> they come from us?  I can think of several ways, but I'd prefer to hear
> some others input first.
>
> Tim
>
> --
> Juju-dev mailing list
> Juju-dev at lists.ubuntu.com
> Modify settings or unsubscribe at:
> https://lists.ubuntu.com/mailman/listinfo/juju-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/juju-dev/attachments/20130408/0eab3a67/attachment.html>


More information about the Juju-dev mailing list