Automation ignores FTBFS guidelines

[Till Hofmann]

Hi all,

So I've just been notified that tolua++ has been retired, which is a 
dependency of one of my packages (fawkes). BZ: 
https://bugzilla.redhat.com/show_bug.cgi?id=1736911

This would have been fine (as no action has been taken), if the 
automation had actually followed the FTBFS guidelines [1]. But it 
hasn't, in many ways:
1. "If an FTBFS or FTI bug remains in the NEW state for at least 1 week, 
any concerned party can set a NEEDINFO for the maintainer to respond and 
send an e-mail reminder with the Bugzilla link to 
<component_name&gt;-maintainers(a)fedoraproject.org, cc’ing the devel mailing 
list (so there is a public record) and commenting on the bug about doing 
so."
I did not see such an email on devel. Also NEEDINFO was set on Sep 22, 
~10 weeks after retirement.
2. "If the bug remains in NEW state for at least another 4 weeks after 
the second e-mail and comment (= at least 8 weeks in total), the package 
will be orphaned. Orphaning can be requested via a releng issue."
There was no email at all, at least not to devel, or the maintainers of 
the dependencies of tolua++.
3. "The normal Orphaned package that needs new maintainers procedure 
will be followed for the packages orphaned in this way, leading to their 
retirement if nobody adopts them."
This did not happen either. In particular, it was never announced that 
the package was orphaned.
4. In fact, as far as I can tell, the package was never orphaned, but 
directly retired.

To put it differently, the guidelines were completely ignored. The 
package was retired 6 days after the initial FTBFS bug, without any 
announcement. This is in stark contrast to the 14 weeks mentioned in the 
guidelines. To add to this, the retirement isn't even mentioned in the 
bug report. It just silently happened.

The comment "your package has not been built successfully in 31. Action 
is required from you" isn't helping either, as the package has long been 
retired at that time.

I understand and support that FTBFS packages are retired, but this is 
not the way this should work. There was no way I could have heard about 
the issue in time. It effectively requires me to become co-maintainer on 
every package that I depend on. Clearly not something I want to do.

So can we please make sure that guidelines apply to everybody, also and 
especially to scripts?

Thanks,
Till


[1] 
https://docs.fedoraproject.org/en-US/fesco/Fails_to_build_from_source_Fai...

[Hans de Goede]

Hi,

On 22-09-2019 14:37, Till Hofmann wrote:
>  Hi all,
>  
>  So I've just been notified that tolua++ has been retired, which is a dependency of
> one of my packages (fawkes). BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1736911
>  
>  This would have been fine (as no action has been taken), if the automation had actually
> followed the FTBFS guidelines [1]. But it hasn't, in many ways:
>  1. "If an FTBFS or FTI bug remains in the NEW state for at least 1 week, any
> concerned party can set a NEEDINFO for the maintainer to respond and send an e-mail
> reminder with the Bugzilla link to <component_name&gt;-maintainers(a)fedoraproject.org,
> cc’ing the devel mailing list (so there is a public record) and commenting on the bug
> about doing so."
>  I did not see such an email on devel. Also NEEDINFO was set on Sep 22, ~10 weeks after
> retirement.
>  2. "If the bug remains in NEW state for at least another 4 weeks after the second
> e-mail and comment (= at least 8 weeks in total), the package will be orphaned. Orphaning
> can be requested via a releng issue."
>  There was no email at all, at least not to devel, or the maintainers of the dependencies
> of tolua++.
>  3. "The normal Orphaned package that needs new maintainers procedure will be
> followed for the packages orphaned in this way, leading to their retirement if nobody
> adopts them."
>  This did not happen either. In particular, it was never announced that the package was
> orphaned.
>  4. In fact, as far as I can tell, the package was never orphaned, but directly retired.
>  
>  To put it differently, the guidelines were completely ignored. The package was retired 6
> days after the initial FTBFS bug, without any announcement. This is in stark contrast to
> the 14 weeks mentioned in the guidelines. To add to this, the retirement isn't even
> mentioned in the bug report. It just silently happened.
>  
>  The comment "your package has not been built successfully in 31. Action is required
> from you" isn't helping either, as the package has long been retired at that
> time.
>  
>  I understand and support that FTBFS packages are retired, but this is not the way this
> should work. There was no way I could have heard about the issue in time. It effectively
> requires me to become co-maintainer on every package that I depend on. Clearly not
> something I want to do.
>  
>  So can we please make sure that guidelines apply to everybody, also and especially to
> scripts? 
I've no opinion on whether the process was followed correctly here (I did not look
closely
at that part of this email).

But I do have something to say about the specific example used. tolua++ being retired
is not a problem for fawkes, as fawkes depends on compat-tolua++, which is maintained
by me and has NOT been retired.

Regards,

Hans

[Till Hofmann]

On 9/22/19 3:01 PM, Hans de Goede wrote:
>  
>  But I do have something to say about the specific example used. tolua++ 
>  being retired
>  is not a problem for fawkes, as fawkes depends on compat-tolua++, which 
>  is maintained
>  by me and has NOT been retired.
>   
Thanks for pointing this out! I assumed compat-tolua++ was a subpackage 
of tolua++ (I hadn't actually looked at the spec file yet).

[Miro Hrončok]

On 22. 09. 19 14:37, Till Hofmann wrote:
>  Hi all,
>  
>  So I've just been notified that tolua++ has been retired, which is a dependency 
>  of one of my packages (fawkes). BZ: 
>  https://bugzilla.redhat.com/show_bug.cgi?id=1736911 
I've closed this bug because it "was already retired", not because of that
bug.

The package was actually retired more than a month ago in here:

https://bugzilla.redhat.com/show_bug.cgi?id=1676145#c12

Based on this rule:

"7. A week before the mass branching, any packages which still have open FTBFS 
bugs from the previous release will be retired."

>  To put it differently, the guidelines were completely ignored. The
> package was 
>  retired 6 days after the initial FTBFS bug, without any announcement. 
No, they were not ignored at all.

>  The comment "your package has not been built successfully in 31.
> Action is 
>  required from you" isn't helping either, as the package has long been retired at
>
>  that time. 
Yes, that's why I closed the bugzilla. See https://pagure.io/releng/issue/8478

>  So can we please make sure that guidelines apply to everybody, also
> and 
>  especially to scripts? 
They do.

-- 
Miro Hrončok
--
Phone: [phone]
IRC: mhroncok

[Miro Hrončok]

On 22. 09. 19 15:40, Miro Hrončok wrote:
>  "7. A week before the mass branching, any packages which still
> have open FTBFS 
>  bugs from the previous release will be retired." 
Whether this is a reasonable thing or not has been discussed in:

https://lists.fedoraproject.org/archives/list/[e-mail].o...

-- 
Miro Hrončok
--
Phone: [phone]
IRC: mhroncok

[Till Hofmann]

On 9/22/19 3:40 PM, Miro Hrončok wrote:
>  On 22. 09. 19 14:37, Till Hofmann wrote:
> > Hi all,
> >
> > So I've just been notified that tolua++ has been retired, which is a 
> > dependency of one of my packages (fawkes). BZ: 
> > https://bugzilla.redhat.com/show_bug.cgi?id=1736911
>  
>  I've closed this bug because it "was already retired", not because of 
>  that bug.
>  
>  The package was actually retired more than a month ago in here:
>  
>  https://bugzilla.redhat.com/show_bug.cgi?id=1676145#c12
>  
>  Based on this rule:
>  
>  "7. A week before the mass branching, any packages which still have open 
>  FTBFS bugs from the previous release will be retired." 
I see, so the package was already FTBFS in F30. So the F31 FTBFS was 
actually a duplicate. It would have helped if it was marked as such, but 
I understand that's easier said than done.

>  
> > To put it differently, the guidelines were completely ignored. The 
> > package was retired 6 days after the initial FTBFS bug, without any 
> > announcement.
>  
>  No, they were not ignored at all.
>  
> > The comment "your package has not been built successfully in 31. 
> > Action is required from you" isn't helping either, as the package has 
> > long been retired at that time.
>  
>  Yes, that's why I closed the bugzilla. See 
>  https://pagure.io/releng/issue/8478
>  
> > So can we please make sure that guidelines apply to everybody, also 
> > and especially to scripts?
>  
>  They do.
>   

You're right, as this was FTBFS in F30 already, it's all good.

So the guidelines allow package retirement without any announcement on 
devel. That's certainly unexpected, at least to me. I also don't think 
this is a good idea, because maintainers of packages that depend on the 
retired package are still clueless until it's too late. Or do they get a 
separate notification?

I still find it somewhat inconsistent that basically every rule in the 
FTBFS guidelines is about notifying people, but (7) allows retirement 
without any prequisites other than being FTBFS for 2 releases -- no 
announcement, no orphaning.


Kind regards,
Till

[Miro Hrončok]

On 22. 09. 19 16:01, Till Hofmann wrote:
>  So the guidelines allow package retirement without any announcement
> on devel. 
>  That's certainly unexpected, at least to me. I also don't think this is a good 
>  idea, because maintainers of packages that depend on the retired package are 
>  still clueless until it's too late. Or do they get a separate notification?
>  
>  I still find it somewhat inconsistent that basically every rule in the FTBFS 
>  guidelines is about notifying people, but (7) allows retirement without any 
>  prequisites other than being FTBFS for 2 releases -- no announcement, no orphaning.

Yes, I agree. That's why I've started:

https://lists.fedoraproject.org/archives/list/[e-mail].o...

-- 
Miro Hrončok
--
Phone: [phone]
IRC: mhroncok

[Till Hofmann]

On 9/22/19 4:08 PM, Miro Hrončok wrote:
>  On 22. 09. 19 16:01, Till Hofmann wrote:
> > So the guidelines allow package retirement without any announcement on 
> > devel. That's certainly unexpected, at least to me. I also don't think 
> > this is a good idea, because maintainers of packages that depend on 
> > the retired package are still clueless until it's too late. Or do they 
> > get a separate notification?
> >
> > I still find it somewhat inconsistent that basically every rule in the 
> > FTBFS guidelines is about notifying people, but (7) allows retirement 
> > without any prequisites other than being FTBFS for 2 releases -- no 
> > announcement, no orphaning.
>  
>  
>  Yes, I agree. That's why I've started:
>  
>
> https://lists.fedoraproject.org/archives/list/[e-mail].o...
>
>  
>   
Ah yes, I now remember reading this. I'll add to if I have anything to add.

Apologies for blaming this on the scripts!

Kind regards,
Till