FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Redhat > Fedora Development

 
 
LinkBack Thread Tools
 
Old 02-26-2010, 02:07 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Josh Boyer wrote:
> Nobody said disallow direct-to-stable pushes completely, entirely, with no
> exceptions. That would indeed be absurd.

But the proposed exception procedures which were floated were so burdensome
and slow that they made the entire exception procedure effectively useless.
For example: Calling an emergency FESCo meeting? Yeah sure! It's faster to
just wait for a push and queue for stable then. And in practice nobody will
want to bother FESCo just to get his trivial bugfix out faster. So it's as
if the exception procedure didn't exist at all.

For an exception procedure to be effective, it needs to be possible for one
person out of a group of several people to ACK a push. (rel-eng would be a
good candidate, they already do very similar tasks. Or possibly the union of
rel-eng and QA, as for the freeze overrides for critical path packages. Or
even FESCo as long as 1 FESCo member is enough to approve it, not a vote.
And no, I wouldn't blanket-approve everything as I have been accused of in
the meeting, please quit the paranoia!)

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-26-2010, 02:09 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Till Maas wrote:
> Imho it takes too long to get packages into updates-testing, if people
> are really interested in testing packages, they often seem to get
> packages directly from Koji, e.g. on this update I got 3 positive Karma
> points (one of them was anonymous) within 76 minutes after submitting
> the update:
> https://admin.fedoraproject.org/updates/F12/FEDORA-2010-0604
>
> It did not seem very useful to delay this update that also fixed several
> very annoying bugs any further.

Good point. Indeed, packages are often tested sufficiently before they even
enter updates-testing. Even if pushes become more frequent, it can still
happen if testing is called for on a fast medium like IRC and the fix
touches many people.

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-26-2010, 02:12 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Michael Schwendt wrote:
> Doesn't sound right. FE could push to stable always and much more quickly,
> too. What was missing was a convenient interface for packagers which they
> could use to decide between testing and stable or whether not to push a
> build at all. It was necessary to submit special requests by email, since
> by default every new build would become a test update. Packagers wanted
> more control, and some even wanted no (!) delays by means of automatically
> pushing to the repos.

Actually, FE didn't have testing at all, all updates were stable. Only the
newer Plague setups (EPEL, RPM Fusion) included a testing repo.

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-26-2010, 02:15 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Patrice Dumas wrote:
> I may be remebering wrong, but an argument for bodhi against those who
> wanted a simpler push mechanism (like wwhat was in the fedora extra days)
> and argued that bodhi will add more unecessary delays was that there
> always was the possibility to push to stable for packagers.

Indeed.

And actually, initially, Bodhi didn't allow direct stable pushes except for
security updates. People complained a lot about that (and for good reasons,
which are also being brought up in this thread) and so direct stable pushes
got added. So why now try to go back to something that didn't work?

> Bringinig down productivity of good packagers for a few bad ones, is,
> in my opinion, not a good move.

+1

> Also, I find it quite ridiculous to have to give use cases, if people
> in FESCo have never needed a direct push to stable, they necessarily
> haven't done enough packaging -- though seeing who is in FESCo, it looks
> quite strange to me since some members are seasoned packagers and some
> even were here before bodhi.

Yeah, it quite surprised me that I was the only one seeing a need for this
feature in FESCo!

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-26-2010, 02:17 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Michael Schwendt wrote:
> Sometimes, not always. Don't forget that reviewers don't review builds
> for all dists, but packagers often publish mass-builds for multiple dists
> without prior testing.

In practice that is not often a source of trouble. (Though new packages are
somewhat more likely to be affected than bugfixes to existing ones.) Most
often what works on Fedora n also works on Fedora m. It's not like the
reviewer tested on Slackware or OS X. ;-)

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-26-2010, 02:18 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

drago01 wrote:

> On Fri, Feb 26, 2010 at 1:43 PM, Michael Schwendt <mschwendt@gmail.com>
> wrote:
>> [...]
>> Unconvincing, though. History has shown that some packagers still managed
>> to push new packages that suffered from broken deps [..]
>
> Well than the review process failed ...


Indeed.

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-26-2010, 02:20 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Jesse Keating wrote:

> On Fri, 2010-02-26 at 14:55 +0100, Kevin Kofler wrote:
>> > The possibility to publish hot-fixes is most important.
>>
>> +1. Not being able to push those out quickly would really suck.
>
> What sucks more is recent "hot-fixes" which were even more broken than
> the issue they were trying to fix. They were pushed directly to stable
> and broke a significant number of systems because of a scenario the
> maintainer didn't imagine or test.

Those weren't "hot-fixes", they were fixes for an issue which wasn't really
a regression. They were also quite invasive and risky, very much unlike the
trivial fixes I'm talking about. The maintainer simply underestimated the
risk, maybe he also overestimated the urgency. But that's not an issue with
the process.

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-26-2010, 02:22 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Jaroslav Reznik wrote:
> Maybe some package rating included in PackageKit would be nice - for
> stable packages it's indicator that this package is worth to install, for
> testing package it would mean it's working (but again - who's going to
> rate it in pkgkit once installed).

That won't solve the problem that people aren't using updates-testing in the
first place. We can't force them to.

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-26-2010, 02:25 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Michael Schwendt wrote:
> Nonsense. Such arbitrary rules will only drive off packagers. The field in
> an update request may be empty because the list of bugzilla tickets is
> sufficient and because the package %changelog adds further details.
>
> It would be wrong to shut the door for everyone just because some
> packagers don't include any info at all. FESCo's first step ought to
> be to send a memo to all packagers. Bodhi could warn about missing
> Notes when no bugzilla numbers have been entered. And if that still
> doesn't help, individual packagers could be talked to.

Yes, there too, it's a people problem, it needs a social solution. Technical
"solutions" will cause both false positives and false negatives and just
cause more trouble.

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-26-2010, 02:40 PM
Kevin Kofler
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

Josh Boyer wrote:
> The time period is mere speculation on your part.

It's not just mere speculation, the idea has been brought up by nirik,
citing EPEL as precedent:
[begin quote (from the meeting log)]
Feb 23 21:40:50 * nirik notes the maintainer also requested a push to
stable in epel, but the epel policy of 2 weeks in testing was observed
instead.
[snip not directly related discussion]
Feb 23 21:53:23 * nirik personally thinks the epel process has been
working nicely...
[snip not directly related discussion]
Feb 23 21:53:41 <skvidal> nirik: I think time-based is probably a hang
up - but....
[end quote]

> You've left out parts that were discussed in the meeting as options (like
> mechanisms to allow direct-to-stable pushes with FESCo/rel-eng/QA karma)

That was my suggestion. All I got was negative comments (AIUI, nobody else
wanted anything less than a majority of FESCo to be able to approve direct
stable pushes, at least nobody said otherwise in the meeting), and even
outright accusations of proposing ad personam rules:
[begin quote (from the meeting log)]
Feb 23 21:58:17 <notting> Kevin_Kofler: of course you'd say 1 FESCo
member, becuase from your statements, it sounds like you would intentionally
sabotage the process and approve everything
[end quote]

> Transparency in process is great and I think it is extermely important.
> What you've done is not transparency. What you've started is a smear
> campaign against a draft policy that hasn't even been written yet. Way to
> be a class-A dickhead.

Transparency means asking for feedback BEFORE writing the policy. The sooner
you involve the community, the better. Putting out a policy as "take it or
leave it", or worse "take it, you have to, we voted it through already" is
not transparent.

Kevin Kofler

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 

Thread Tools




All times are GMT. The time now is 07:43 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org