On Tue, Sep 13, 2016 at 11:56 AM, Jacob Hoffman-Andrews <[email protected]>
wrote:

>
> > Anything that isn't tested Will be broken.  We can't fix that in the
> > protocol, but we can fix it by adding tests :)
> We can also fix it by deleting unnecessary things, which is the best and
> most reliable way to avoid bugs. That's why I think this is the right
> approach.
>

Note that you can synthesize the "TOS-as-bool" case from the
"TOS-as-string" case -- the bool just reflects whether the string has the
right value.  In that sense, having the string there would actually
simplify the CA's job.

So string or bool is just a question of whether you make the
breaking-change case or the non-breaking-change case easier (where breaking
== requires re-agreement).  If it's a bool, then the server has to go unset
it on all the accounts when it makes a breaking change (instead of just
changing the string and breaking equality).  If it's a string, then the
sever has to update the strings whenever it makes a non-breaking change to
the URL.

Of course, these options are not mutually exclusive.  You could have
something like a "last agreed" URL field and an "agreement compatible with
current TOS" field.  Don't really think this is a good idea, just bringing
it up :)

Net/net, I'm a little sad to lose the "thing you last agreed to"
information.  But if it's not acceptable to the CA any more anyway, why
does it matter?  (The client can also keep a copy.)  And it does seem like
we should optimize for the non-breaking-change case.  So I can live with
just a bool.

--Richard
_______________________________________________
Acme mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/acme

Reply via email to