bodhi 0.4.10 features

Jon Stanley jonstanley at gmail.com
Mon Jan 28 12:56:51 UTC 2008


On Jan 28, 2008 12:19 AM, Tom Lane <tgl at redhat.com> wrote:
> "Jon Stanley" <jonstanley at gmail.com> writes:
> > ... The important thing to take
> > away from what was decided at FESCo is that the checkbox to not
> > auto-close bugs is/will be no more.
>
> Umm ... say what!?
>
> This is entirely not sane.  It presumes that bodhi and the release bots
> always know better than the human maintainer whether or not a bug should
> be auto-closed.

When the human maintainer knows better, it was decided that there is
nothing preventing them from clicking the "reopen" button in Bugzilla.
 There are times when that will be appropriate, but those really
should be few and far between.

What use case is there for NOT auto-closing once an update has been
pushed (assuming that one bug is ONE problem in ONE release - which is
something that this was designed to "enforce"?   I'm really interested
in hearing - Luke and I came up with no valid ones at FUDCon,

> If you do this, it will result in maintainers omitting
> bug numbers from bodhi entries altogether, anytime they are not sure if
> the bug is definitely dealt with, or if it is related but not solely
> confined to the particular release being made.

Not quite following you here - if the bug turns out not to be dealt
with, then the bug can always be re-opened.  Remember - ONE bug == ONE
problem in ONE release.  You should be able to auto-close that, and
re-open if that particular problem is not dealt with.

> PLEASE do not do this.  It has zero advantage and significant downside.

The advantage is significantly improved reporter usability and brand
perception - I can now say (which I could not before) - this is what
happens to a bug once you file it, here is the states that are valid
for it to be in.  When the maintainer believes that it's fixed, the
update system WILL auto-close this bug.  I'm still not clear as to the
downside, though.

> The only real effect will be to force maintainers to omit relevant
> information from release announcements.

I *really* hope that doesn't happen




More information about the fedora-devel-list mailing list