DevHeads.net

RANT: compton-ng…

<a href="https://bugzilla.redhat.com/show_bug.cgi?id=1738293" title="https://bugzilla.redhat.com/show_bug.cgi?id=1738293">https://bugzilla.redhat.com/show_bug.cgi?id=1738293</a>

How did it pass review?

Silently doing Obsoletes of an active package and doing so even
without version which is prohibited by the Packaging Guidelines.
Really?

I am not sure if it is written anywhere, but hardcoding
/builddir/whatsoever is so bad…

There is specific reason why we have --buildtype=plain in %meson
macro, because with release buildtype, debuginfo is unusable…

Have anybody ever talked to the RPM folks about proper way of enabling
LTO for builds? This specific way of setting it makes build
non-reproducible.

Comments

Re: RANT: compton-ng…

By Vitaly Zaitsev ... at 08/13/2019 - 11:37

I can suggest following:

1. Do review process much more easier by running Fedora Review Tool on
Koji using fedpkg. Output can be sent to email for additional manual checks.

2. Do package revocation procedure from non-responsive maintainers much
more simple and easier. Currently we need to wait 3 weeks, post to
mailing list, create a FESCo ticket and then wait, wait, wait.

Re: RANT: compton-ng…

By Artem Tim at 08/13/2019 - 08:51

Fixed <a href="https://src.fedoraproject.org/rpms/compton-ng/blob/master/f/compton-ng.spec" title="https://src.fedoraproject.org/rpms/compton-ng/blob/master/f/compton-ng.spec">https://src.fedoraproject.org/rpms/compton-ng/blob/master/f/compton-ng.spec</a>

Because no one reviewing packages in Fedora. Only one person reviewing and help witch packaging - eclipseo. Many thanks to him.

Also no one want to wait for month+ and pinging non existent maintainer.

Keep more like that and more people will continue switch to SUSE, Arch, whatever where no such bureaucracy and where no terrible tooling from 1990s like RHBZ for reviewing process.

Re: RANT: compton-ng…

By Ernestas Kulik at 08/13/2019 - 09:52

On Tue, 2019-08-13 at 13:51 +0000, Artem Tim wrote:
I see plenty of names in Bugzilla doing package reviews. If you’ve
waited a long time, this list is one place to ask for them (review
swaps for extra points).

Have you even attempted getting in touch with the original author? Have
you attempted proposing switching the upstream of the existing package?

Lots of things to do beforehand, but I can’t see any communication to
have happened.

No need to get defensive about this. Provide feedback ahead of time
and, preferably, help out. Replacing tooling is not just flipping a
switch.

Re: RANT: compton-ng…

By Artem Tim at 08/13/2019 - 11:43

This guy probably didn't know about that since he still trying to push it since 2014
<a href="https://bugzilla.redhat.com/show_bug.cgi?id=1161834" title="https://bugzilla.redhat.com/show_bug.cgi?id=1161834">https://bugzilla.redhat.com/show_bug.cgi?id=1161834</a>
<a href="https://bugzilla.redhat.com/show_bug.cgi?id=1517451" title="https://bugzilla.redhat.com/show_bug.cgi?id=1517451">https://bugzilla.redhat.com/show_bug.cgi?id=1517451</a>
And this is original dev of this app. GJ.

No, and not even because i have previous experience with non-responsive maintainer and lesson learned of some other people who struggle with this process. And not even because maintainer of old compton doesn't exist anymore and this is obvious - <a href="https://src.fedoraproject.org/user/axeld" title="https://src.fedoraproject.org/user/axeld">https://src.fedoraproject.org/user/axeld</a> (see Activity) and his second packages already maintained by Till Maas for a long time. Because i didn't wanted overwrite old compton, i mistakenly put 'obsoletes' instead of 'conflicts'.

Also that's why non responsive maintainer process **finally** changed now and now we need only 1 week to wait <a href="https://pagure.io/fesco/fesco-docs/pull-request/15" title="https://pagure.io/fesco/fesco-docs/pull-request/15">https://pagure.io/fesco/fesco-docs/pull-request/15</a>

Re: RANT: compton-ng…

By Vitaly Zaitsev ... at 08/13/2019 - 11:31

On 13.08.2019 16:52, Ernestas Kulik wrote:
But he was right. Review process in openSUSE is much more easier, than
in Fedora. Currently no one want to do it.

And Fedora Review Tool currently destroying our SSDs by lots of
unnecessary writes.

Re: RANT: compton-ng…

By Adam Williamson at 08/13/2019 - 11:34

On Tue, 2019-08-13 at 18:31 +0200, Vitaly Zaitsev via devel wrote:
And yet this thread demonstrates that without good review, we will get
garbage packages. Review processes exist for a reason.

This is not true. Many people are doing package reviews.

Re: RANT: compton-ng…

By Artem Tim at 08/13/2019 - 11:48

Review processes is one of the best thing in Fedora and i believe this helps with packaging quality overall. The problem is no one literally wanted to do this in Fedora since 99% of packages reviewing one person which i mentioned already before there.

Re: RANT: compton-ng…

By Vitaly Zaitsev ... at 08/13/2019 - 11:42

On 13.08.2019 18:34, Adam Williamson wrote:
Every package can become a garbage, because after package review no one
reviews it again. I see lots of legacy packages, which breaks modern
package guidelines. Some of them we fixed by my PRs.

And every day I see in this mailing list requests for review swaps. Do
you think this is fine?

Re: RANT: compton-ng…

By Adam Williamson at 08/13/2019 - 13:01

On Tue, 2019-08-13 at 18:42 +0200, Vitaly Zaitsev via devel wrote:
This is true, but that's not a reason to stop doing package reviews.
"Things aren't perfect" is never a good excuse for "...so we can make
them worse!"

Yeah, I do.

Re: RANT: compton-ng…

By Vitaly Zaitsev ... at 08/13/2019 - 14:22

On 13.08.2019 20:01, Adam Williamson wrote:
I never asked to stop doing package reviews. Package review is a good
thing. I just asked to simplify this process for example by running
Fedora Review Tool on Koji and then email results to reviewer for
additional manual checks.

Re: RANT: compton-ng…

By Abhiram Kuchibhotla at 08/13/2019 - 13:52

Hey guys, I just saw this so sorry for the late reply.

I'm still around. I haven't been contacted but I see no problems with
moving away from compton in an orderly fashion.
Feel free to take over if you want. :)

Regards,
Abhiram K

On Tue, 13 Aug, 2019, 11:32 PM Adam Williamson, < ... at fedoraproject dot org>
wrote:

Re: RANT: compton-ng…

By Bob Mauchin at 08/13/2019 - 09:48

On Tuesday, 13 August 2019 15:51:41 CEST Artem Tim wrote:

I've only checked that Compton was obsolete, which it seems due to inactivity
on GitHub. However I shouldn't have assumed that you discussed this with
Compton maintainer, that's on me. You are obsoleting his or her package, you
should have talked about it with them. You wouldn't want your package to be
obsoleted without you being consulted, would you? If he or she is not
responding, you should launch a non-responsive maintainer process, not ignore
it and go your own way.^

Best regards,

Robert-André

Re: RANT: compton-ng…

By Kevin Kofler at 08/13/2019 - 07:26

Igor Gnatenko wrote:
This was submitted even as an update to Fedora 30!
<a href="https://bodhi.fedoraproject.org/updates/FEDORA-2019-8f0fdfb37e#comment-999789" title="https://bodhi.fedoraproject.org/updates/FEDORA-2019-8f0fdfb37e#comment-999789">https://bodhi.fedoraproject.org/updates/FEDORA-2019-8f0fdfb37e#comment-9...</a>

Kevin Kofler

Re: RANT: compton-ng…

By Panu Matilainen at 08/13/2019 - 04:26

On 8/13/19 11:17 AM, Igor Gnatenko wrote:
Rpm can't help it if gcc/linker argument makes builds non-reproducable.
That said, see <a href="https://github.com/rpm-software-management/rpm/pull/809" title="https://github.com/rpm-software-management/rpm/pull/809">https://github.com/rpm-software-management/rpm/pull/809</a>,
gcc upstream is aware of the issue and supports -flto=auto now, which is
what builds should be using for reproducability.

Re: RANT: compton-ng…

By Kamil Dudka at 08/13/2019 - 05:19

On Tuesday, August 13, 2019 11:26:12 AM CEST Panu Matilainen wrote:
By reproducibility you mean that the number of jobs can be (in some cases)
controlled by RPM macros?

Does the number of jobs have any impact on the resulting binary?

Kamil

Re: RANT: compton-ng…

By Panu Matilainen at 08/13/2019 - 05:22

On 8/13/19 1:19 PM, Kamil Dudka wrote:
See the above URL.

It doesn't affect the actual link stage, but it affects the stored
optflags in the package (and the actual binaries if that was enabled)

Re: RANT: compton-ng…

By Kamil Dudka at 08/13/2019 - 05:38

On Tuesday, August 13, 2019 12:22:17 PM CEST Panu Matilainen wrote:
OK, got it. The issue is that -flto=N can be stored in the resulting binary
RPMs and the N can differ across builds. So you want to use -flto=auto to
always store the same flag, even though the flag may have different semantics
in different builds of the same package.

Sorry for not understanding it originally. It is good idea to provide some
context when you post to such generic list as fedora-devel. Not all readers
are experts in all topics and/or have capacity to read in depth all referred
documents.

Thank you for clarifying it!

Kamil

Re: RANT: compton-ng…

By Panu Matilainen at 08/13/2019 - 07:26

On 8/13/19 1:38 PM, Kamil Dudka wrote:
Not all, of course. I did assume that people actually interested in the
subject would go and read it ;)

The plot thickens though (again, its all in the referred PR), and now
that the cat is out of the bag I really do need to clarify it further here:

Contrary to what was said above, gcc does not, and apparently will not,
have an -flto=auto option. It was proposed for gcc 10, but instead the
basic form -flto will default to that behavior. Seems sane to me.

Or that's how it currently stands, AFAICS. We're still talking about
unreleased software so you never really know until it ships.

Re: RANT: compton-ng…

By Vitaly Zaitsev ... at 08/13/2019 - 05:58

On 13.08.2019 12:38, Kamil Dudka wrote:
Supported only by GCC 10+.