Discussion:
The use of Kernel:HEAD
(too old to reply)
Stephan Kulow
2014-06-25 08:02:25 UTC
Permalink
Raw Message
Hi,

This is pretty much history repeating, but I feel *very* disappointed
with the way you treat Factory. Factory tumbles from one RC to another.

And while I try to get 3.15.1 in myself, you replace Kernel:HEAD with
rc2? What's the thinking behind that? I guess it's not too far stretched
to claim that no one is maintaining the kernel-source package
in Factory. All packaging activities are done by brainless scripts ;(

Greetings, Stephan
Takashi Iwai
2014-06-25 09:29:52 UTC
Permalink
Raw Message
At Wed, 25 Jun 2014 10:02:25 +0200,
Post by Stephan Kulow
Hi,
This is pretty much history repeating, but I feel *very* disappointed
with the way you treat Factory. Factory tumbles from one RC to another.
And while I try to get 3.15.1 in myself, you replace Kernel:HEAD with
rc2? What's the thinking behind that? I guess it's not too far stretched
to claim that no one is maintaining the kernel-source package
in Factory. All packaging activities are done by brainless scripts ;(
IMO, a stable rolling release should take Kernel:stable instead.
So it's a question again what is FACTORY.


Takashi
Stephan Kulow
2014-06-25 09:33:41 UTC
Permalink
Raw Message
Post by Takashi Iwai
At Wed, 25 Jun 2014 10:02:25 +0200,
Post by Stephan Kulow
Hi,
This is pretty much history repeating, but I feel *very* disappointed
with the way you treat Factory. Factory tumbles from one RC to another.
And while I try to get 3.15.1 in myself, you replace Kernel:HEAD with
rc2? What's the thinking behind that? I guess it's not too far stretched
to claim that no one is maintaining the kernel-source package
in Factory. All packaging activities are done by brainless scripts ;(
IMO, a stable rolling release should take Kernel:stable instead.
So it's a question again what is FACTORY.
Well, I don't mind what kernel is in factory as long as it works - if
that's rc4 or rc7 or final doesn't matter to me. Neither as factory
maintainer nor as factory user.

But I want it maintained and submitted when it makes sense. What bothers
me is this "I commit to git and I'm done - shall random scripts handle
the rest" mentality.

If that will go away in Kernel:stable, I'm fine with changing.

Greetings, Stephan
Takashi Iwai
2014-06-25 12:47:35 UTC
Permalink
Raw Message
At Wed, 25 Jun 2014 11:33:41 +0200,
Post by Stephan Kulow
Post by Takashi Iwai
At Wed, 25 Jun 2014 10:02:25 +0200,
Post by Stephan Kulow
Hi,
This is pretty much history repeating, but I feel *very* disappointed
with the way you treat Factory. Factory tumbles from one RC to another.
And while I try to get 3.15.1 in myself, you replace Kernel:HEAD with
rc2? What's the thinking behind that? I guess it's not too far stretched
to claim that no one is maintaining the kernel-source package
in Factory. All packaging activities are done by brainless scripts ;(
IMO, a stable rolling release should take Kernel:stable instead.
So it's a question again what is FACTORY.
Well, I don't mind what kernel is in factory as long as it works - if
that's rc4 or rc7 or final doesn't matter to me. Neither as factory
maintainer nor as factory user.
But I want it maintained and submitted when it makes sense. What bothers
me is this "I commit to git and I'm done - shall random scripts handle
the rest" mentality.
All Kernel:* repos _are_ maintained. Kernel:HEAD might be broken
sometimes, but it's usually fixed quickly once when recognized. Not
much different from the normal packages.

I do wonder, however, whether you got actually a submission from
Kernel:HEAD to FACTORY. That is, is Kernel:HEAD automatically
submitted, or did you take it manually?
Post by Stephan Kulow
If that will go away in Kernel:stable, I'm fine with changing.
A similar breakage can happen on Kernel:stable, too, but maybe less
likely because there are less activities there, and the code base is
supposed to be stabler.

The only drawback would be that we'll miss the integration test of the
latest Linus kernel by this change. But, if FACTORY is aimed to be
really stably usable, Kernel:stable is definitely a safer choice.


Takashi
Stephan Kulow
2014-06-25 13:05:44 UTC
Permalink
Raw Message
Post by Takashi Iwai
Post by Stephan Kulow
But I want it maintained and submitted when it makes sense. What bothers
me is this "I commit to git and I'm done - shall random scripts handle
the rest" mentality.
All Kernel:* repos _are_ maintained. Kernel:HEAD might be broken
sometimes, but it's usually fixed quickly once when recognized. Not
much different from the normal packages.
But I was asking about maintenance of the kernel in factory not in Kernel:*
Post by Takashi Iwai
I do wonder, however, whether you got actually a submission from
Kernel:HEAD to FACTORY. That is, is Kernel:HEAD automatically
submitted, or did you take it manually?
The latest (automatic) submission is from 2014-06-06, so I submitted it
myself as request 238317.
Post by Takashi Iwai
Post by Stephan Kulow
If that will go away in Kernel:stable, I'm fine with changing.
A similar breakage can happen on Kernel:stable, too, but maybe less
likely because there are less activities there, and the code base is
supposed to be stabler.
The only drawback would be that we'll miss the integration test of the
latest Linus kernel by this change. But, if FACTORY is aimed to be
really stably usable, Kernel:stable is definitely a safer choice.
Other development projects submit late RCs too if they are confident
about it. But they actually think about if to submit or not. And that's
what I'm talking about. What I'm missing is the taking care and taking
responsibility for the factory kernel. The most important part - the
releasing to factory - is left to scripts.

Greetings, Stephan
Stefan Seyfried
2014-06-27 19:01:47 UTC
Permalink
Raw Message
Post by Takashi Iwai
The only drawback would be that we'll miss the integration test of the
latest Linus kernel by this change. But, if FACTORY is aimed to be
really stably usable, Kernel:stable is definitely a safer choice.
We don't get that testing anyway:

susi:~ # zypper se -s -r factory-oss kernel-default|grep ^i
i | kernel-default | package | 3.15.rc7-1.2 | x86_64 | factory-oss

Which means it is seriously out of date (serious in the sense of that
ugly local root exploit...)

I am using Kernel:HEAD on top of factory for a long time now, one reason
being that factory alone is too boring and I want to do the testing of
Linus' RC but am too lazy to compile my own, but the other reason as
that I was tired of never getting any security relevant kernel fixes in
Factory.
--
Stefan Seyfried

"For a successful technology, reality must take precedence over
public relations, for nature cannot be fooled." -- Richard Feynman
Jeff Mahoney
2014-06-27 13:32:16 UTC
Permalink
Raw Message
Post by Stephan Kulow
Hi,
This is pretty much history repeating, but I feel *very*
disappointed with the way you treat Factory. Factory tumbles from
one RC to another.
And while I try to get 3.15.1 in myself, you replace Kernel:HEAD
with rc2? What's the thinking behind that? I guess it's not too far
stretched to claim that no one is maintaining the kernel-source
package in Factory. All packaging activities are done by brainless
scripts ;(
Kernel:HEAD is autogenerated from our master git branch and will
always contain the latest upstream kernel, starting with -rc2. That is
not going to change. If that's a problem for Factory, then perhaps
Factory shouldn't take it automatically regardless of contents.

- -Jeff

- --
Jeff Mahoney
SUSE Labs
Stephan Kulow
2014-06-27 14:39:31 UTC
Permalink
Raw Message
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Post by Jeff Mahoney
Post by Stephan Kulow
Hi,
This is pretty much history repeating, but I feel *very*
disappointed with the way you treat Factory. Factory tumbles
from one RC to another.
And while I try to get 3.15.1 in myself, you replace Kernel:HEAD
with rc2? What's the thinking behind that? I guess it's not too
far stretched to claim that no one is maintaining the
kernel-source package in Factory. All packaging activities are
done by brainless scripts ;(
Kernel:HEAD is autogenerated from our master git branch and will
always contain the latest upstream kernel, starting with -rc2. That
is not going to change. If that's a problem for Factory, then
perhaps Factory shouldn't take it automatically regardless of
contents.
That's exactly what I'm talking about! Factory doesn't "take" anything,
maintainers submit something - but there is no maintainer for the
factory kernel only scripts.

Greetings, Stephan

- --
Ma muaß weiterkämpfen, kämpfen bis zum Umfalln, a wenn die
ganze Welt an Arsch offen hat, oder grad deswegn.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iEYEARECAAYFAlOtgiMACgkQwFSBhlBjoJb83QCfXwwJ4MJ/uFHf6LAaI2US+Aj7
WiMAn2TjY+5jeld/wQVnK1P5u/7t9+FR
=q9KJ
-----END PGP SIGNATURE-----
--
To unsubscribe, e-mail: opensuse-kernel+unsubscribe-***@public.gmane.org
To contact the owner, e-mail: opensuse-kernel+owner-***@public.gmane.org
Stephan Kulow
2014-06-27 15:01:09 UTC
Permalink
Raw Message
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Post by Stephan Kulow
That's exactly what I'm talking about! Factory doesn't "take"
anything, maintainers submit something - but there is no maintainer
for the factory kernel only scripts.
So, I went forward and setup the devel prj for Factory in yet another
kernel recompilation prj: devel:openSUSE:Factory:kernel

In there I can copypac the kernel I want from the prj I need and push
it. Possibly with yet another script :)

Greetings, Stephan

- --
Ma muaß weiterkämpfen, kämpfen bis zum Umfalln, a wenn die
ganze Welt an Arsch offen hat, oder grad deswegn.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iEYEARECAAYFAlOthzUACgkQwFSBhlBjoJb7YwCglf2wEVH2Xf1QzkxImae1s00m
ihsAniBcexSefbjm/4ry6jPgLBHI4lpC
=ICC0
-----END PGP SIGNATURE-----
--
To unsubscribe, e-mail: opensuse-kernel+unsubscribe-***@public.gmane.org
To contact the owner, e-mail: opensuse-kernel+owner-***@public.gmane.org
Loading...