On 03/11/2010 02:52 PM, Mat Booth wrote:
> Saw this in today's updates:
>
>
> Cleanup : cyrus-sasl-2.1.23-4.fc12.i686
> 195/254
> groupdel: group 'saslauth' does not exist
> Non-fatal POSTUN scriptlet failure in rpm package cyrus-sasl
> warning: %postun(cyrus-sasl-2.1.23-4.fc12.i686) scriptlet failed, exit status 6

This case is a nice example demonstrating several defects in "applying
karma votes for QA":


1. The update package was sitting in "updates-testing" since 2010-02-22.


2. It did receive +3 karma points before being pushed to "updates"
c.f. https://admin.fedoraproject.org/updates/cyrus-sasl-2.1.23-6.fc12
=> There are people who claim to have tested it and not having noticed
anything unusual.


3. During today's update I was immediately "greeted" by the same error
message Mat cites above and BZ'ed it.
c.f. https://bugzilla.redhat.com/show_bug.cgi?id=572399


4. Despite the fact this package had been pushed to "updates", I had
been able to cast a karma vote on the package in bodhi.
c.f. https://admin.fedoraproject.org/updates/cyrus-sasl-2.1.23-6.fc12
=> A malfunction in bodhi


5. Unlike in many other similar cases before, the packager responded
almost immediately and tried to provide an updated package (Thanks!).
However, due to the fact update-candidates are not immediately pushed to
"updates-testing", this package is not available in "updates-testing".
=> "karma voting on packages in updates-testing" is not applicable in
situations like these (being directly affect by a bug, the bug still
"being hot")

Seemingly, other people who were affected by this bug did pick up the
package from other sources (Most probably directly from bodhi) and
provided feedback through BZ
c.f. https://bugzilla.redhat.com/show_bug.cgi?id=572399


Think about it, FESCO!

Ralf




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

On Thu, 2010-03-11 at 16:35 +0100, Ralf Corsepius wrote:

> This case is a nice example demonstrating several defects in "applying
> karma votes for QA":
>
>
> 1. The update package was sitting in "updates-testing" since 2010-02-22.
>
>
> 2. It did receive +3 karma points before being pushed to "updates"
> c.f. https://admin.fedoraproject.org/updates/cyrus-sasl-2.1.23-6.fc12
> => There are people who claim to have tested it and not having noticed
> anything unusual.

They didn't claim that. They claimed that it 'works fine here' and
'works fo[r] me' (the other message just says 'thanks', but we can count
it as 'works for me' as that's what the radio button says).

The point is: this update *does* work. The error message is non-fatal.
The software works. So what they claim is correct. What you claim is
also correct.

What this highlights is, indeed, a defect, though - the same one I
raised at the FESco meeting: we don't have a definition of what exact
criteria a package should meet to get a +1. Should we vote down updates
which have non-fatal scriptlet errors? That question doesn't yet have a
clear answer. It doesn't, however, mean the initial testers were 'not
carefully enough', as you claim on the ticket. It just means you were
working to different criteria.
--
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

Adam Williamson <awilliam@redhat.com> writes:

> The point is: this update *does* work. The error message is non-fatal.
> The software works. So what they claim is correct. What you claim is
> also correct.

If the user/group saslauth is not needed by cyrus-sasl, why has it been
added in the first place?

Andreas.

--
Andreas Schwab, schwab@redhat.com
GPG Key fingerprint = D4E8 DBE3 3813 BB5D FA84 5EC7 45C6 250E 6F00 984E
"And now for something completely different."
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

On 03/12/2010 09:58 AM, Andreas Schwab wrote:
> Adam Williamson<awilliam@redhat.com> writes:
>
>> The point is: this update *does* work.
A package which installs with errors does NOT work properly

>> The error message is non-fatal.
>> The software works. So what they claim is correct. What you claim is
>> also correct.
>
> If the user/group saslauth is not needed by cyrus-sasl, why has it been
> added in the first place?
Exactly. If the package "works properly" without specialized
user/groups, then something is wrong with it.

Finally: Yes, the package "works properly for me", because it's
disabled, not in use and only pulled in as mandatory installation bloat
by other packages

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

On Fri, 2010-03-12 at 09:58 +0100, Andreas Schwab wrote:
> If the user/group saslauth is not needed by cyrus-sasl, why has it
> been
> added in the first place?
>
>

Packaging bug or some leftover, but it appears the user/group isn't used
for anything. So the package functions, but there are still some
packaging bugs.

--
Jesse Keating
Fedora -- Freedom² is a feature!
identi.ca: http://identi.ca/jkeating
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

On Fri, 2010-03-12 at 07:58 -0800, Jesse Keating wrote:
> On Fri, 2010-03-12 at 09:58 +0100, Andreas Schwab wrote:
> > If the user/group saslauth is not needed by cyrus-sasl, why has it
> > been
> > added in the first place?
> >
> >
>
> Packaging bug or some leftover, but it appears the user/group isn't used
> for anything. So the package functions, but there are still some
> packaging bugs.

It is the user which can be used if the admin uncomments the appropriate
line in /etc/sysconfig/saslauthd.
--
Tomas Mraz
No matter how far down the wrong road you've gone, turn back.
Turkish proverb

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

On Fri, Mar 12, 2010 at 09:58:58AM +0100, Andreas Schwab wrote:
> Adam Williamson <awilliam@redhat.com> writes:
>
> > The point is: this update *does* work. The error message is non-fatal.
> > The software works. So what they claim is correct. What you claim is
> > also correct.
>
> If the user/group saslauth is not needed by cyrus-sasl, why has it been
> added in the first place?
>
The rpm changelog references this bug:
https://bugzilla.redhat.com/show_bug.cgi?id=185614

but there isn't much info in the bug report from the package maintainer to
tell what is going on. (Is that feature finished? Then why isn't the bug
closed? Have all the recommendations in the bug been performed? Then the
saslauth user and group are being used on the filesystem by the package and
the daemon [to write out its run file])

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