bodhi 0.9.3 deployed to production

[Peter Robinson]

On Thu, Nov 15, 2012 at 6:15 AM, Adam Williamson <awilliam at redhat.com> wrote:
> On Thu, 2012-11-15 at 05:52 +0100, Kevin Kofler wrote:
>> Luke Macken wrote:
>> > A new bugfix release of Bodhi has just been deployed to production.
>> >
>> >     https://admin.fedoraproject.org/updates
>> >
>> > Bugs and enhancement requests can be filed here:
>> >
>> >     http://bodhi.fedorahosted.org
>>
>> This seems to be closing bugs as CURRENTRELEASE rather than ERRATA now, is
>> that intentional? If yes, why?
>
> There seems to be a mania for CURRENTRELEASE lately. I don't know where
> it's coming from.
>
> The policy clearly states ERRATA:
>
> https://fedoraproject.org/wiki/BugZappers/BugStatusWorkFlow#CLOSED
>
> Now admittedly the policy is more or less something I pulled out of my
> ass a few years ago and I have my suspicions as to how many people a)
> have ever read it (few) or b) care (fewer). But hey, it's there.
>
> In the end we rarely do anything with the resolutions anyway, so I
> stopped caring as much as I used to, but at least our official tools
> should probably follow what are nominally our policies.
>
> Note that Fedora's policy differs *significantly* from Red Hat's here,
> as Fedora's update workflow is entirely different from RHEL's.
> https://bugzilla.redhat.com/page.cgi?id=fields.html - the 'help' page
> you get by clicking on various bits of a bug page - is RHEL's policy,
> not Fedora's. There is a link to the Fedora policy at the top of it.
> Unfortunately, most of the links to the fields.html page jump into the
> middle of it with anchors, so no-one sees the link.
>
> The other problem with our policy is that it's somewhat of a hack job,
> because it involves taking the resolutions that were written for RHEL's
> update process and make sense in that context and applying them the
> Fedora update process, in which context they really don't make sense at
> all. So several of the choices are just arbitrary decisions I made when
> I was writing the smegging thing. This is of course one of the drawbacks
> of sharing a bug tracker with RHEL, but then the point is always made
> that the effort of having our own would outweigh the benefits.
>
> Still, even in RHEL's policy, CURRENTRELEASE is clearly wrong for a
> straightforward 'this was a bug that got fixed and we pushed an update'
> case, so I don't know why people are suddenly plumping for it, other
> than that it somehow 'sounds right'.

Unless it's a bugfix that is pushed to multiple releases and the one
your after isn't the one that made it stable first at which point
errata IMO makes slightly more sense. After all in our case
CURRENTRELEASE changes roughly every 6 months so it becomes a moving
target. In RHEL CR makes more sense because it would be the current
dot release in the X cycle, where X is 4/5/6/7 etc.

Peter