Bug 17529 - Low-level packages to update for Mageia 6
Summary: Low-level packages to update for Mageia 6
Status: RESOLVED FIXED
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: Cauldron
Hardware: i586 Linux
Priority: release_blocker normal
Target Milestone: ---
Assignee: Base system maintainers
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-01-18 20:46 CET by David Walser
Modified: 2017-01-17 10:29 CET (History)
6 users (show)

See Also:
Source RPM:
CVE:
Status comment: David Walser is going to check a last time.


Attachments
mgrarepo `svn diff' output to bring LVM2 to version 2.02.154 (709 bytes, patch)
2016-05-21 20:01 CEST, Theodoros Kalamatianos
Details | Diff

Description David Walser 2016-01-18 20:46:02 CET
Here are some low-level base system packages, which are only usually touched by a few people (mostly tmb and coling), that can be updated before Mageia 6.

-----------------------------

cloog/cloog-ppl/ppl, as I recall these had some relation to gcc, but I don't see any relationship to it now.  Perhaps these packages aren't needed anymore, but every distro seems to have some form of cloog packaged.

CLooG has three variants:
http://www.cloog.org/

Most distros have either the main cloog (isl) or cloog parma (ppl), some have both.  We have the ppl one right now.  The parma one depends on the ppl library:
http://bugseng.com/products/ppl/

The newest versions are 0.18.4 for cloog (isl), in debian/omdv/gentoo/ubuntu, and
0.16.1 for cloog-parma (ppl), in debian/omdv/ubuntu.

Thomas Backlund previously told me he'll take care of these packages.

-----------------------------

kernel/filesystem/boot-related:

dkms 2.2.0.3.1 omdv
https://abf.rosalinux.ru/openmandriva/dkms

dracut 044 fedora
http://pkgs.fedoraproject.org/cgit/rpms/dracut.git

initscripts 9.65 in fedora and 10.0 upstream
http://pkgs.fedoraproject.org/cgit/rpms/initscripts.git
http://git.fedorahosted.org/cgit/initscripts.git/snapshot/initscripts-10.0-1.tar.xz

kmod 22 fedora
http://pkgs.fedoraproject.org/cgit/rpms/kmod.git

lvm2 2.02.140 fedora
http://pkgs.fedoraproject.org/cgit/rpms/lvm2.git

xfsprogs 4.3.0 fedora
http://pkgs.fedoraproject.org/cgit/rpms/xfsprogs.git

-----------------------------

rpmlint 1.8 fedora (coling)
http://pkgs.fedoraproject.org/cgit/rpms/rpmlint.git

We have a local patch that needs rediffed, and probably Colin understands this best.

Reproducible: 

Steps to Reproduce:
David Walser 2016-01-18 20:46:33 CET

Blocks: (none) => 15527
CC: (none) => mageia

Comment 1 David Walser 2016-01-23 22:32:02 CET
I forgot to mention that ppl's newest version is 1.1.

It looks like cloog/ppl were used by gcc through Mageia 5, but no longer are in Cauldron.  Maybe we can just drop them.

dracut and kmod are updated now.
Comment 2 Theodoros Kalamatianos 2016-05-21 20:00:31 CEST
LVM2 is at version 2.02.154 now. I have been using a build of it on my system for a few days and have encountered no problems, although admittedly it did not have much to do except for assembling my lvmcache volumes on boot.

I am attaching a patch to bring the LVM2 package to this version.

CC: (none) => thkala

Comment 3 Theodoros Kalamatianos 2016-05-21 20:01:51 CEST
Created attachment 7829 [details]
mgrarepo `svn diff' output to bring LVM2 to version 2.02.154
Comment 4 David Walser 2016-05-21 20:19:02 CEST
cloog/ppl dropped and replaced by isl.  We had the latest upstream until upstream isl released 0.17.1, but the gcc documentation doesn't say whether that one is supported yet, so that can wait for Mageia 7.

I believe dracut, kmod, and xfsprogs are currently up to date now.

dkms, initscripts, and rpmlint still need some love from Colin or someone.

lvm2 suggested update in Comment 3, needs review by Thomas.

e2fsprogs 1.43 has two test failures on i586:
http://pkgsubmit.mageia.org/uploads/failure/cauldron/core/release/20160521171414.luigiwalser.duvel.11665/log/e2fsprogs-1.43-1.mga6/build.0.20160521171514.log

binutils could potentially be updated to 2.26, but I'd expect that to go for Mageia 7 instead.

glibc 2.23 would be nice to have for Mageia 6, however.  Up to Thomas.

Otherwise, we're in good shape.
Comment 5 Morgan Leijström 2016-05-29 02:44:53 CEST
On an related matter i wonder if this old package should be in this list of things to get updated:
Bug 16712 - We have libqtkeychain0 version 0.1.0 from 2013, later versions improve KDE5/FC support

CC: (none) => fri

Comment 6 David Walser 2016-05-29 03:35:51 CEST
(In reply to Morgan Leijström from comment #5)
> On an related matter i wonder if this old package should be in this list of
> things to get updated:
> Bug 16712 - We have libqtkeychain0 version 0.1.0 from 2013, later versions
> improve KDE5/FC support

Please have these kind of packages talked about on Bug 17528 instead.
Comment 7 Rémi Verschelde 2016-07-28 10:52:44 CEST
Are there still low-level packages that would be worth upgrading so late in the release process, or can we consider this issue as FIXED/OLD?
Comment 8 David Walser 2016-07-28 13:42:11 CEST
This could still use a look.

I guess we're going with glibc 2.22 for Mageia 6.

It looks like initscripts was updated in SVN and then reverted for some reason.

kmod 23 just came out, so tmb can decide on that one.

lvm2 is updated upstream frequently, so tmb can decide what version to go with.

dkms and rpmlint seem unlikely to be updated any time soon.

binutils has some security issues (Bug 18987), so something should be done there.

I still need to check again for outdated packages, I haven't in over a month.
Comment 9 David Walser 2016-08-09 23:53:24 CEST
CC'ing Pascal, since Thomas is taking a step back.

CC: (none) => pterjan

Rémi Verschelde 2016-09-08 10:06:32 CEST

Assignee: tmb => basesystem

Comment 10 David Walser 2016-09-09 14:18:11 CEST
Marking as blocker for now, as if we update initscripts or glibc, we'll need to do that before the release.

I don't know why tmb decided to stick with 2.22 (probably only he knows) and I also don't know why we didn't follow through with the initscripts update (which apparently had been done in SVN).

Priority: Normal => release_blocker

Comment 11 Rémi Verschelde 2016-09-09 14:19:26 CEST
Maybe we should bring those points to the dev@ ML to reach a decision quickly whether we want to work on those updates, or consider those frozen for mga6.
Samuel Verschelde 2016-09-12 16:16:09 CEST

Status comment: (none) => If glibc or initscripts will be updated, should be done before release

Comment 12 Thomas Backlund 2016-09-12 21:51:03 CEST
(In reply to David Walser from comment #10)
> Marking as blocker for now, as if we update initscripts or glibc, we'll need
> to do that before the release.
> 
> I don't know why tmb decided to stick with 2.22 (probably only he knows) and
> I also don't know why we didn't follow through with the initscripts update
> (which apparently had been done in SVN).


Because normally we dont update core packages like glibc/gcc/binutils when we are past distro rebuild as they all tend to give some fun/subtle "hiccups" and we want a stable core...

CC: (none) => tmb

Comment 13 David Walser 2016-09-12 21:55:38 CEST
(In reply to Thomas Backlund from comment #12)
> (In reply to David Walser from comment #10)
> > Marking as blocker for now, as if we update initscripts or glibc, we'll need
> > to do that before the release.
> > 
> > I don't know why tmb decided to stick with 2.22 (probably only he knows) and
> > I also don't know why we didn't follow through with the initscripts update
> > (which apparently had been done in SVN).
> 
> 
> Because normally we dont update core packages like glibc/gcc/binutils when
> we are past distro rebuild as they all tend to give some fun/subtle
> "hiccups" and we want a stable core...

Thanks for the comment.

With regards to glibc, that definitely makes sense.  We never anticipated this development cycle would take this long, and IIRC the mass rebuild was in January, so we *could* do another one, but I don't know that that's desirable.

As for initscripts, I would think it would have less of an impact, but people like you and Colin who are more intimately familiar with it would know much better than I would about that.
Samuel Verschelde 2016-09-13 10:38:21 CEST

Status comment: If glibc or initscripts will be updated, should be done before release => If glibc or initscripts will be updated, should be done before release. glibc involves rebuilding the whole distro so not to be done lightly.

Comment 14 Nicolas Lécureuil 2016-11-06 15:01:32 CET
i don't think we should update initscript this late.
Maybe as an update after more testing.

So i think we should close this bugreport.

CC: (none) => mageia

Comment 15 David Walser 2016-11-06 15:08:23 CET
I should check the versions of things again since it has been months.
Comment 16 Samuel Verschelde 2016-11-08 11:36:43 CET
(In reply to David Walser from comment #15)
> I should check the versions of things again since it has been months.

Please do, we need to close this bug report one way or the other.
Samuel Verschelde 2016-11-08 12:47:33 CET

Status comment: If glibc or initscripts will be updated, should be done before release. glibc involves rebuilding the whole distro so not to be done lightly. => David Walser is going to check a last time.

Comment 17 David Walser 2016-11-14 15:38:44 CET
dkms 2.3 debian
glibc 2.24 upstream
initscripts 9.65/10.0 fedora/upstream
rpmlint 1.9 fedora
systemd 232 fedora
util-linux 2.29 fedora

I guess are what would be left on this list at this point (and binutils, but I mentioned earlier we have a security bug filed for that one).  They're not going to be updated for Mageia 6, so I guess this is WONTFIX at this point, but it'd be nice to have a Mageia 7 bug for this so hopefully they can be addressed early in the cycle, especially considering things like dkms, initscripts, and rpmlint were not addressed *at all* during the entire Mageia 5 development cycle.  We really need those who understand those packages to address those ones.
Comment 18 Samuel Verschelde 2016-11-14 16:18:29 CET
Thanks for the review. It's indeed late for those. Let's indeed create a bug report with Target Milestone "Mageia 7" and high priority. 

dkms has its own bug report: bug 17198
rpmlint now has one: bug 19777
and initscripts: 19778
and a new tracker to tie them all: 19779

Closing for Mageia 6.

Resolution: (none) => FIXED
Status: NEW => RESOLVED

Samuel Verschelde 2017-01-17 10:29:39 CET

Blocks: 15527 => (none)


Note You need to log in before you can comment on or make changes to this bug.