Code review comment for lp:~allenap/maas/more-better-things--bug-1389007

> I think this needs a discussion upstream in order to establish the
> best path forward.
>
> If upstream is willing to accept an upgrade-to-32bit patch, then we
> should invest energy into that because long-term it's much more
> sustainable for us. If they have another suggested solution we should
> use it.
>
> We should only really take this fork in if there's really no other
> viable solution, which should be quite unlikely.

I think it's the right and proper thing to do to talk to the Twisted
project, and see if this can go upstream.

Twisted is a community project; we or I would likely be the de facto
maintainers.

Sustainably, my guess is that we're not in for a lot of work in either
case. AMP in Twisted is essentially done. This branch doesn't add new
features, just a change to an on-the-wire detail.

Even so, being upstream would give others the chance to use the code,
and give greater visibility to bugs and security issues.

Anecdotally, though from my own experience of the Twisted development
process, I can say that it's onerous to get a patch reviewed and landed.

If we get it upstream it could be a long time before that release makes
it back into Ubuntu. It's unlikely to get into the current LTS. We'd
need to carry this code anyway, for a while.

In all, I think we can make it a goal to get this upstream, but I don't
think we should block on it, except just to check there isn't a
ready-made alternative.