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 02-27-2010, 05:36 AM
Bruno Wolff III
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On Fri, Feb 26, 2010 at 23:48:47 -0600,
Chris Adams <cmadams@hiwaay.net> wrote:
> Once upon a time, Bruno Wolff III <bruno@wolff.to> said:
> > P.S. I don't use enablerepo. I'll yum install a local copy of the rpm and see
> > what it needs if it doesn't install successfully.
>
> That seems like extra and unnecessary work. You doesn't do anything
> without telling you, so "yum --enablrepo=*testing update foo" is going
> to tell you more about what dependencies are needed than "yum
> localinstall foo.rpm".

If things don't work yes. But I would typically be expecting it to work.
Also my repo config files all point to local copies, which may or may
not be there depending on what I am using. So simple enablerepo won't
always work for me.

> I install packages from updates-testing all the time; I'll do a "yum

So do I. I keep a local mirror and use it by default.

> --enablerepo=*testing check-update" and look for packages that are
> "interesting" to me (things that I use that I might would notice a

For stuff that interests me, I don't wait for it to show up in
updates(-testing)?, but rather pull it from koji. (I spend way too much
time looking at koji build lists.)
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-27-2010, 06:09 AM
Paul Wouters
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On Fri, 26 Feb 2010, Adam Williamson wrote:

> On Sat, 2010-02-27 at 06:03 +0100, Ralf Corsepius wrote:
>
>> 2) Recent dnssec-conf updates all did receive several -1, nevertheless
>> these updates were pushed.
>
> This is indeed a problem. Obviously, relying on the judgment of
> maintainers isn't working.

The karma arrived before it as pushed? I don't think it did. The 1.21-8
release fixed the 1.21-7 release of also checking the "very old" config
file location.

There were some bug reports for dnssec-conf that dated back to earlier
versions and are still present, but this emergency update was not to
fix those and -1 on those features would not cause it to remain blocked
or unpushed, especially in the light of older versions DOS'ing a bunch
of name servers due to a bad trust anchor in combination with a bind
bug.

In fact, dnssec-conf is now being phased out, but this is pending one
more dependancy fix for unbound, which is caught in the new DSO linking
behaviour and needs fixing.

Paul
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-27-2010, 06:25 AM
Adam Williamson
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On Sat, 2010-02-27 at 02:09 -0500, Paul Wouters wrote:
> On Fri, 26 Feb 2010, Adam Williamson wrote:
>
> > On Sat, 2010-02-27 at 06:03 +0100, Ralf Corsepius wrote:
> >
> >> 2) Recent dnssec-conf updates all did receive several -1, nevertheless
> >> these updates were pushed.
> >
> > This is indeed a problem. Obviously, relying on the judgment of
> > maintainers isn't working.
>
> The karma arrived before it as pushed? I don't think it did. The 1.21-8
> release fixed the 1.21-7 release of also checking the "very old" config
> file location.

Sorry, I was replying in haste. I should've made clear that I was
talking more in general, and don't have any specific direct knowledge of
the dnssec case. I know of multiple cases where updates have been pushed
hastily, but I don't have any direct knowledge of the dnssec case
specifically and wouldn't want to cast any aspersions in anyone's
direction there.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-27-2010, 08:57 AM
Ralf Corsepius
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On 02/27/2010 08:25 AM, Adam Williamson wrote:
> On Sat, 2010-02-27 at 02:09 -0500, Paul Wouters wrote:
>
>> On Fri, 26 Feb 2010, Adam Williamson wrote:
>>
>>
>>> On Sat, 2010-02-27 at 06:03 +0100, Ralf Corsepius wrote:
>>>
>>>
>>>> 2) Recent dnssec-conf updates all did receive several -1, nevertheless
>>>> these updates were pushed.
>>>>
>>> This is indeed a problem. Obviously, relying on the judgment of
>>> maintainers isn't working.
>>>
>> The karma arrived before it as pushed? I don't think it did.
I don't know, I voted -1 at a time it previous votes had accumulated to +2.

>> The 1.21-8
>> release fixed the 1.21-7 release of also checking the "very old" config
>> file location.
>>
>
Still installing this *-1.21-8 release immediately brought down my local
bind.

To me, this qualifies as "what ever you did, was insufficient".

> Sorry, I was replying in haste. I should've made clear that I was
> talking more in general, and don't have any specific direct knowledge of
> the dnssec case. I know of multiple cases where updates have been pushed
> hastily, but I don't have any direct knowledge of the dnssec case
> specifically and wouldn't want to cast any aspersions in anyone's
> direction there.
>
Well, to voting is an inadequate means for judging a package's quality,
because bugs showing in individual cases are not co-related to "works
for many" - It's a fundamental flaw of the system.

Ralf

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-27-2010, 02:38 PM
Mike McGrath
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On Sat, 27 Feb 2010, Kevin Kofler wrote:

> Chris Adams wrote:
> > IMHO you're developing the wrong distro. It is statements like yours
> > that contribute to the "Fedora is a rolling beta" perception (and I
> > don't think that's a good perception to have). If you want to target
> > rawhide with rolling releases of KDE, have fun. Once a release is out
> > the door, try not to just throw a new kitchen sink in for the hell of
> > it.
>
> Some people actually LIKE rolling releases, indeed some distros use
> completely rolling releases (e.g. Arch Linux). We are currently somewhere
> inbetween (partly release-based, partly rolling), and IMHO this compromise
> is working great. We get the advantages from a rolling release model, but
> with a lot less surprise breakage as in a true rolling model because
> disruptive changes like libata go only into new releases.
>

If only we had some sort of rolling release, that tracked as closely with
upstream as possible, where the users of said release understood they were
drinking from the firehose. Meanwhile, along side that release we could
have periodic stable releases that don't move so quickly. That way you get
what you want and I get what I want. Oh wait! That's the world we live
in today. Next time a user tells you "I want a newer X" tell them
"Upgrade to rawhide".

-Mike
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-27-2010, 02:44 PM
Mike McGrath
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On Sat, 27 Feb 2010, Mike McGrath wrote:

> On Sat, 27 Feb 2010, Kevin Kofler wrote:
>
> > Chris Adams wrote:
> > > IMHO you're developing the wrong distro. It is statements like yours
> > > that contribute to the "Fedora is a rolling beta" perception (and I
> > > don't think that's a good perception to have). If you want to target
> > > rawhide with rolling releases of KDE, have fun. Once a release is out
> > > the door, try not to just throw a new kitchen sink in for the hell of
> > > it.
> >
> > Some people actually LIKE rolling releases, indeed some distros use
> > completely rolling releases (e.g. Arch Linux). We are currently somewhere
> > inbetween (partly release-based, partly rolling), and IMHO this compromise
> > is working great. We get the advantages from a rolling release model, but
> > with a lot less surprise breakage as in a true rolling model because
> > disruptive changes like libata go only into new releases.
> >
>
> If only we had some sort of rolling release, that tracked as closely with
> upstream as possible, where the users of said release understood they were
> drinking from the firehose. Meanwhile, along side that release we could
> have periodic stable releases that don't move so quickly. That way you get
> what you want and I get what I want. Oh wait! That's the world we live
> in today. Next time a user tells you "I want a newer X" tell them
> "Upgrade to rawhide".
>

<bad form replying to myself, sorry>

Or to put it another way, why aren't you doing this and telling others to
do this? If someone is on F11 still, why do you think they want the
latest and greatest software? If they did, they'd upgrade to f12. And
further still, why wouldn't they be running rawhide? The rolling update
release exists. Why force rolling updates on people that haven't chosen
to run rawhide?

-Mike
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-27-2010, 03:21 PM
Patrice Dumas
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On Sat, Feb 27, 2010 at 02:55:41PM +0100, Kevin Kofler wrote:
>
> New packages which don't Obsolete existing packages or Provide existing
> provided names cannot cause any of the above. (They may technically trigger

Special care should be given to the auto-generated Provides. I remember
a package of mine that messed up the buildroot because of a perl
auto-generated provides that happened because my package had a private
copy of a perl module...

Anyway, I don't think that new packages are very relevant to the issue,
because
* they can sit in testing for some time, they are new package, it is
not as if they fix something
* the review is already a thorough review of the 'update', so when
errors happen, they are, in my opinion, more a failure of the review
than of the update system.

Of course, dependencies of updated packages that must enter rapidly
because the update of the dependent package is important should
certainly require more scrutiny. I remember vaguely that a new package
that entered as a dependency of an updated package caused issues in
the past.

--
Pat
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-27-2010, 03:24 PM
Adam Williamson
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On Sat, 2010-02-27 at 08:45 +0000, Camilo Mesias wrote:
> On Sat, Feb 27, 2010 at 6:12 AM, Adam Williamson <awilliam@redhat.com> wrote:
> > this is a *terrible* idea. We may see users as a 'resource', but they
> > don't see themselves this way. We should not interrupt their usage of
> > their computer to try and exploit them to our ends.
>
> What if it was an opt-in scheme? Users would consent to receive a
> limited number of contacts about their current packages and for their
> trouble would get streamlined access to potential fixes.
>
> I think there's enough in that for the opt-in scheme to be marketed
> successfully, because although some people would see the interactions
> as annoying, others would welcome the chance to participate.

Yup, anything making it easier for people who actively choose to
participate in testing to actually do the testing and provide their
feedback would be fantastic.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-27-2010, 03:27 PM
Adam Williamson
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On Sat, 2010-02-27 at 10:57 +0100, Ralf Corsepius wrote:

> > Sorry, I was replying in haste. I should've made clear that I was
> > talking more in general, and don't have any specific direct knowledge of
> > the dnssec case. I know of multiple cases where updates have been pushed
> > hastily, but I don't have any direct knowledge of the dnssec case
> > specifically and wouldn't want to cast any aspersions in anyone's
> > direction there.
> >
> Well, to voting is an inadequate means for judging a package's quality,
> because bugs showing in individual cases are not co-related to "works
> for many" - It's a fundamental flaw of the system.

Yeah, it's not perfect: there are cases where we have, say, a complex
kernel update which works fine for most people but causes a significant
regression for some particular bit of hardware. We wouldn't want to put
that update out, but it's easy for it to get five +1s before someone
with the specific bit of hardware comes by and gives it a -1...and even
then, +4 looks good if you're not reading the feedback too carefully.

So yeah, I agree it's not a perfect system - detailed suggestions for
improving it would be welcome, I'm sure. I don't think 'not perfect' is
the same as 'useless', though. I think it's pretty easy to make a case
that Bodhi has had a significant positive impact on the overall quality
of the updates that have fully utilized it. It rarely makes things
*worse*
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net

--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
 
Old 02-27-2010, 03:32 PM
Adam Williamson
 
Default FESCo wants to ban direct stable pushes in Bodhi (urgent call for feedback)

On Sat, 2010-02-27 at 11:41 +0100, Till Maas wrote:

> Ok, maybe the question should be then: How much does AutoQA support me
> writing these tests? E.g. this test is pretty simple, but afaics there
> is no easy support for the common tasks that are needed to run the test,
> but not really part of the test, e.g. installing the package or setting
> up a machine.
>
> The Writeing AutoQA Tests wiki page[0] says:
> | I'll say it again: Write the test first. The tests don't require
> | anything from autotest or autoqa. You should have a working test before
> | you even start thinking about AutoQA
>
> But this is not really supportive, because if I want to test a packages,
> I need a framework that creates the initial environment, e.g. a system
> of the Fedora version to be tested with the package installed and there
> needs to be a way to interact with the programs.
>
> Or is this really a test I can easily integrate into AutoQA currently?
> Say we start without locales and commandline arguments, then the test
> would be:
>
> Input: Package to be tested as ${PACKAGE}
> for binary in $(rpm -ql ${PACKAGE} | grep bin); do ${binary} 2>&1 | grep
> "Segmentation fault" && echo "test failed" ; done

It'd probably be best to ask on the autoqa-devel mailing list - you'll
get Will and Kamil there, who know far more in detail than I do
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net

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

Thread Tools




All times are GMT. The time now is 08:59 AM.

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