Mass closing EOL bugs should not close bugs with pending updates

[Orcan Ogetbil]

On Sun, Feb 17, 2013 at 4:53 PM, Christoph Wickert
<christoph.wickert at gmail.com> wrote:
> Am Sonntag, den 17.02.2013, 16:12 -0500 schrieb Orcan Ogetbil:
>> On Sun, Feb 17, 2013 at 3:26 PM, Christoph Wickert wrote:
>> > Am Sonntag, den 17.02.2013, 14:46 +0100 schrieb Tadej Janež:
>> >> On Sun, 2013-02-17 at 12:02 +0100, Christoph Wickert wrote:
>> >> >
>> >> > I found that a couple of F16 bugs were closed by endoflife at fp.o even
>> >> > though there were pending updates for F17 and F18 to fix them. As a
>> >> > result, the bugs are now closed WONTFIX even they were or are going to
>> >> > be fixed.
>> >>
>> >> What you describe is another example of strange behavior of the Fedora
>> >> EOL Closure script.
>> >> I discovered two related problems that I described three days ago:
>> >> http://lists.fedoraproject.org/pipermail/devel/2013-February/178649.html
>> >>
>> >> Since then I found a page that describes the Fedora 16 EOL Closure
>> >> procedure:
>> >> https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Fedora18#Fedora_16_EOL_Closure
>> >>
>> >> It says that the bugs with "version == Fedora 16" and "status != CLOSED"
>> >> are subject to automatic closure. Could you give an example of a bug
>> >> that you described?
>> >
>> > https://admin.fedoraproject.org/updates/FEDORA-2013-2359/lxpanel-0.5.12-1.fc18
>> > and
>> > https://admin.fedoraproject.org/updates/FEDORA-2013-2359/lxpanel-0.5.12-1.fc17
>> > fix several bugs, among them two very old and annoying ones:
>> > https://bugzilla.redhat.com/show_bug.cgi?id=782431 and
>> > https://bugzilla.redhat.com/show_bug.cgi?id=785906
>> >
>> > As you can see the bugs were already ON_QA before they were closed
>> > WONTFIX.
>> >
>>
>> But those are bugs filed against Fedora 16. Will Fedora 16 receive the
>> fix at this point? No. Hence WONTFIX is correct.
>
> No it's not. The bug is resolved in a later release of Fedora, thus
> CURRENTRELEASE or NEXTRELEASE are correct. WONTFIX implies it was not
> fixed it all.
>

Hmm, I always interpreted that the "RELEASE" in CURRENTRELEASE or
NEXTRELEASE refers to the Release tag of the corresponding package,
and not to the Fedora release version. The bug header looks like this

Product:	Fedora
Component:	lxpanel
Version(s):	16
Platform:	x86_64

In my interpretation, I would _not_ close this bug if I fix it only
- for another product than "Fedora" , e.g. for "RHEL". RHEL might
suffer from the same bug as well, but this complaint was for Fedora.
We cannot ignore Fedora.
- for another component. It does not make sense to fix a kernel bug
and close this bug report, does it?
- for another platform. If the bug is filed for i686 and my fix only
fixes x86_64, my fix is not related to this bug report. (Of course the
fix might fix both i686 and x86_64, in which case the bug can be
closed.)
- for another Fedora version. I believe the bug is Fedora version
specific (else why do we have a Version tag?). If a bug is filed for
version 16, it is for version 16. Other versions might also have the
same bug, but this is irrelevant from the bug report's perspective.

Your interpretation treats everything the same but the "Version" part
differently. Any particular reason?

>> Either the person who filed the bug, or the assignee could have bumped
>> the bug's Fedora version in the given timeframe, but they did not. I
>> think the 28 day period was sufficient amount of time to react.
>
> I agree, but I don't think we can rely on the bug reporters, nor on the
> maintainers.

May I ask why? What is the heads'up message for, that was posted 28
days ago? Do we not need that message?

> Therefor I suggest to not mass close bugs which are already
> ON_QA, in fact, that's what the bugzappers documentation says, too.
>
> Best regards,
> Christoph
>

Best,
Orcan