This sounds fine to me too, though I have some quibbles with the PRs. - We should get rid of "processing" globally (#186 leaves it for authorizations) - It seems duplicative to have "deactivated", "revoked", and "invalid". Maybe we can keep "invalid" as "this failed without ever being valid", but let's decide on one for "this was valid, and is no longer". Between "deactivated" and "revoked", I slightly prefer "revoked", but then it seems like we should change the wording of the sections about deactivation to talk about revocation instead. Not something I want to waste a lot of bikeshed paint on, though.
Roland, if you could update the PRs, sounds like we could go ahead and merge. On Wed, Sep 7, 2016 at 12:48 AM, Ron <[email protected]> wrote: > On Tue, Sep 06, 2016 at 02:18:43PM -0400, Daniel McCarney wrote: > > +1. > > > > Removing 'processing' and 'unknown' makes sense to me. I think the > > remaining statuses are comprehensive enough. > > I agree with this. It was one of the original review nits I posted > to this list many months back. > > 'unknown' was basically an internal state of Boulder that had no real > reason to ever be leaked out into a protocol message, and the > distinction between 'pending' and 'processing' was never defined in > the spec, only as an implementation detail of Boulder (though it did > make some sense how Boulder was using it). > > I still think it would also be useful for the spec to give some > definition for what each of the remaining statuses are supposed to > mean and when a server should send them, and some guidance for what > a client should do if it receives them - even though that should be > a bit more obvious now if we prune away the more ambiguous ones. > > > > On Fri, Aug 26, 2016 at 6:12 PM, Roland Bracewell Shoemaker < > > [email protected]> wrote: > > > > > Currently the specification defines the statuses 'unknown' and > > > 'processing' as valid for both applications and authorizations. Given > > > that both also have 'valid', 'pending', and 'invalid' I can't really > see > > > a reason to keep 'unknown' and for applications in particular the > > > meaning of 'processing' seems quite similar to 'pending' (whereas with > > > an authorization I think it makes sense when a validation is actively > > > being done by the server). > > > > > > Any objections to removing these statuses? > > _______________________________________________ > Acme mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/acme >
_______________________________________________ Acme mailing list [email protected] https://www.ietf.org/mailman/listinfo/acme
