Guilty as charged, but I'd say it's quite well bounded.
> The storage and API in juju/state/charm.py will also have to change
> accordingly, for instance.
That's the core of the problem: that I thought I'd done enough damage in the last week or so, and that it would be important *not* to break charm-state storage and thus have to bump topology.VERSION. I'll go ahead with the nicer but more disruptive approach you propose in a stacked branch.
[1,2]
> Hmmm.. this is feeling a bit like a hack.
Guilty as charged, but I'd say it's quite well bounded.
> The storage and API in juju/state/charm.py will also have to change
> accordingly, for instance.
That's the core of the problem: that I thought I'd done enough damage in the last week or so, and that it would be important *not* to break charm-state storage and thus have to bump topology.VERSION. I'll go ahead with the nicer but more disruptive approach you propose in a stacked branch.
[3,4]
Good points; I'll fix them in this branch.