Code review comment for lp:~doanac/utah/server-cleanups

On 04/18/2013 03:50 PM, Max Brustkern wrote:
> I see two things that look like they'll work differently.
>
> If you try to upload files and some of them don't exist, the exception will be raised before uploading any of them instead of after. That seems like it's probably fine.

yeah, the results from the run don't change, so it seemed harmless.

> Also, the handling of exceptions related to weird provisioning possibilities for Libvirt goes away. I think that makes the class slightly less robust for situations that we're not seeing at all right now. I wonder if we will see them in the future. Probably not. Seems reasonable, but I haven't tested anything yet. If you have, we're probably in good shape.

the main difference was something I thought didn't make sense. In the
original code, you could specify !new, and if _load failed, we'd call
_create and try and correct your mistake for you. In the new code, that
would fail. alternatively it did a similar logic for new=True. I found
the logic confusing, not sure if it has practical uses?