Comment 3 for bug 33876

Revision history for this message
Celso Providelo (cprov) wrote : Re: [Bug 33876] separate binary NEW entries for each architecture

Colin Watson wrote:

[snip]

> (As a distro admin, I don't really care about the details of the
> database backend, and generally I don't think I should have to. queue
> should be able to hide details that don't matter.)

It's relative, bad drivers assume the same concept, they "don't really care"
about the essential details, we know the result.
Anyway, let's discuss at the required abstraction level.

[snip]

> They should show up as a single aggregated record in queue, and I should
> be able to override them all at once. At present, if three architectures
> have built binaries when I process NEW, then the NEW approval will only
> apply to those three builds, and when the other three build the binaries
> then they'll show up in NEW again, which creates extra unnecessary work
> for me.

It follows the initial specification that don't assume we have the same
overrides across architectures. if you say we should assume that, I can
implement this in the 'find_and_apply_overrides' method ASAP.

> katie didn't have this misfeature; once I did NEW processing for a given
> binary then it applied to later builds of the same binary on other
> architectures.

Right, I'll save my commendations about Katie ;)

--
Celso Providelo <email address hidden>
Canonical Ltd - http://www.canonical.com