DevHeads.net

efivar and mokutil long standing FTBFS

efivar and mokutil fail to build from source. They have been retired, then
unretired and they still fail to build from source.

Following the policy:
<a href="https://docs.fedoraproject.org/en-US/fesco/Fails_to_build_from_source_Fails_to_install/" title="https://docs.fedoraproject.org/en-US/fesco/Fails_to_build_from_source_Fails_to_install/">https://docs.fedoraproject.org/en-US/fesco/Fails_to_build_from_source_Fa...</a>

I kindly ask the maintainers to rebuild them or orphan them if they cannot take
care of them.

<a href="https://bugzilla.redhat.com/show_bug.cgi?id=1674840" title="https://bugzilla.redhat.com/show_bug.cgi?id=1674840">https://bugzilla.redhat.com/show_bug.cgi?id=1674840</a>
<a href="https://bugzilla.redhat.com/show_bug.cgi?id=1675403" title="https://bugzilla.redhat.com/show_bug.cgi?id=1675403">https://bugzilla.redhat.com/show_bug.cgi?id=1675403</a>

Thanks,

Comments

Re: efivar and mokutil long standing FTBFS

By =?UTF-8?B?TWlyb... at 09/10/2019 - 04:47

On 13. 08. 19 19:10, Miro Hrončok wrote:
I kindly ask the maintainers after 3 weeks of no further response to rebuild the
packages.

Re: efivar and mokutil long standing FTBFS

By Peter Robinson at 09/10/2019 - 06:35

On Tue, Sep 10, 2019 at 10:48 AM Miro Hrončok < ... at redhat dot com> wrote:
Peter Jones, added to email, is working on a new version to land, he
is currently focused on RHEL-8 deliverables which are taking his
priority ATM. He says he will have it fixed by the time Fedora 31 goes
to final freeze.

Peter

Re: efivar and mokutil long standing FTBFS

By =?UTF-8?B?TWlyb... at 09/10/2019 - 06:38

On 10. 09. 19 13:35, Peter Robinson wrote:
Thanks for the update.

Re: efivar and mokutil long standing FTBFS

By Zbigniew =?utf-... at 10/23/2019 - 18:00

On Tue, Sep 10, 2019 at 01:38:03PM +0200, Miro Hrončok wrote:
We're in final freeze now. Any progress?

Zbyszek

Re: efivar and mokutil long standing FTBFS

By Leigh Scott at 10/24/2019 - 02:45

I poked mokutil with a couple of upstream fixes, now it fails only on i686.

<a href="https://koji.fedoraproject.org/koji/taskinfo?taskID=38515227" title="https://koji.fedoraproject.org/koji/taskinfo?taskID=38515227">https://koji.fedoraproject.org/koji/taskinfo?taskID=38515227</a>

Re: efivar and mokutil long standing FTBFS

By Zbigniew =?utf-... at 10/24/2019 - 03:17

On Thu, Oct 24, 2019 at 07:45:03AM -0000, Leigh Scott wrote:

mokutil.c:1977:16: error: comparison of integer expressions of different signedness: 'unsigned int' and 'int' [-Werror=sign-compare]
1977 | if (salt_size > settings_len - (next - settings)) {
| ^

That should be fairly easy to fix...

Zbyszek

Re: efivar and mokutil long standing FTBFS

By Leigh Scott at 10/24/2019 - 04:10

I have a fix for efivar as well but it only compiles on 64 bit <a href="https://leigh123linux.fedorapeople.org/pub/patches/0001-fix-ftbfs.patch" title="https://leigh123linux.fedorapeople.org/pub/patches/0001-fix-ftbfs.patch">https://leigh123linux.fedorapeople.org/pub/patches/0001-fix-ftbfs.patch</a>

<a href="https://koji.fedoraproject.org/koji/taskinfo?taskID=38517342" title="https://koji.fedoraproject.org/koji/taskinfo?taskID=38517342">https://koji.fedoraproject.org/koji/taskinfo?taskID=38517342</a>

Re: efivar and mokutil long standing FTBFS

By Peter Robinson at 08/13/2019 - 12:43

Firstly they're just FTBFS in F-30.... This isn't exactly long
standing, if it was F-26 like some that were retired I could
completely understand that statement but F-30 is pushing the rhetoric
a bit here.

What do you mean by retired and unretired and still FTB?

Maybe you're not aware of this thing called RHEL-8, maybe the
maintainer has been snowed under and as they stand they work as
intended and it's not exactly a complete drama if they're not fixed
right away. Or maybe he's drowning in email from the 8+ pointless
repetitive bug updates that have been added to the bugs.

I happen to know he's working on updates to them, but he also has
other priorities.

There really has to be a better way to deal with this, I spoke with
many people, that are drowning in pointless email, like weekly SPAM
reminders are really too much. One group I had a discussion with are
well aware of the problems of their FTB packages but have other
priorities and fix them when they get moments between the tidal waves.

For example the OLPC/sugar stuff I do I'm slowly, as I and they get
time, training others up on the packaging/Spin side of things,
upstream is moving towards python3 and everyone is aware of this and
the issues involved, but while we do this we now have hours of extra
work to do because a whole bunch of the packages have been ripped out
from under us in the process without proper notification if you happen
to miss an email to devel, or in some of the people I'm working with
aren't even on devel. PLEASE make it stop.... I'm actually considering
stopping most of my work in Fedora because it makes me way too anxious
to continue and that coming from me is something, I'd really hate to
think how many silent contributors we're losing from this.

Peter

Re: efivar and mokutil long standing FTBFS

By =?UTF-8?B?TWlyb... at 08/13/2019 - 12:59

On 13. 08. 19 19:43, Peter Robinson wrote:
It has been half a year without any response. Hence it is long standing.
You may think that half a year is not long standing. Fine. Sorry for using that
language when I should have said "6 months old" instead.

They have been retired as part of the FTBFS policy.
Then they have been unretired and retagged because it broke the compose.
Yet they still have not been rebuilt.

I believe that if a maintainer doesn't have 5 minutes to say "I'm swamped with
RHEL 8 (or whatever else), please, can somebody help me with those FTBFS" or
even "I don't care" they are pretty much no-maintainers. Getting the FTBFS fixed
in 6 months is not right-away. If you find the repetitive bug updates pointless,
feel free to propose a better approach.

I don't happen to know that. The lack of communication is making it hard. For
what i know they are not even aware of this issue.

Don't we all have other priorities? Is it so hard to communicate about out
priorities openly?

Are you proposing to stop doing exactly what? Following policies? Having
policies? Or is something that I personally am doing that I should stop? I
gladly head proposals of how to make things better. But "PLEASE make it stop..."
isn't really a constructive proposal.

I realize that getting constant notifications is overwhelming. But when we don't
do that, people are angry about "unannounced" breakage.

Re: efivar and mokutil long standing FTBFS

By Peter Robinson at 08/14/2019 - 05:54

The retirement, which took me by shock and IMO wasn't well advertised,
happened while people were traveling to flock last Wednesday, and
people are still returning a week later. People also have deadlines
and real work to do and there's literally _SO_ much bugzilla mail from
all of this most people just mute it or out right delete it because
they can't damn well keep up [1]

[1] <a href="https://twitter.com/hughsient/status/1161382573228142594" title="https://twitter.com/hughsient/status/1161382573228142594">https://twitter.com/hughsient/status/1161382573228142594</a>

In some cases, and using these cases as a perfect example, they
require specialised information about low level things like UEFI,
secure boot, signing of core boot process where there's specialised
information and I don't even fill up a hand of fingers listing the
people that have this skill set and all of those people are so
overloaded with work that they actively prioritide stuff. This isn't a
basic python binding here.

Well the deluge of a bazillion extra bug emails doesn't help TBH. I
have generally kept up with bug emails and this is now stressing me
out I'm probably going to quit!

Yes! For some it is.

Well you can't currently see the forest for the trees and this
bullshit, for the first time in 15 years in making me think about
quitting the Fedora project, and I spoke with many at Flock are
seriously stressed about it and sick of it. I will be bringing this up
at council because this is not sustainable and whether it be by
retiring all of the packages because everything that depends on the
kernel is retired or scaring away all the part time contributors this
militant level of process is causing active harm to the project.

Peter

Re: efivar and mokutil long standing FTBFS

By =?UTF-8?B?TWlyb... at 08/14/2019 - 06:47

On 14. 08. 19 12:54, Peter Robinson wrote:
Despite our differences, I'd like to say that I would really regret it, if the
actions that me and FESCo/Releng took caused you to quit Fedora.

Maintaining a modern distribution is a complex task, so there will inherently be
disagreements about processes and methodologies, but I believe we all share the
goal of making Fedora better.

I'm glad that you're offering to bring this issue at Council, as you're right,
this is a sensitive topic, and could definitely benefit from more voices so we
can fine tune process to both be positive for Fedora, but also not to alienate
maintainers such as yourself.