8:25 a.m.
Hi!
Our rawhide trees seem to be in a mostly sane state now. Here is a list
of steps I'm aware of before the Everything repos for F12 can be created:
- move the nvidia-driver to the nonfree-updates-testing repositories, as
requested by kwizart (#928)
- push the new em8300-kmod to the updates repos, once they exist; move
vdr-dxr3 there, as it depends on em8300-kmod (#920)
- push a older picard-freeworld, if one shows up in time (#932)
Anything else? If yes, then please speed up, as I'd would like to create
the Everything repos on Friday -- updating them afterwards is a lot of
manual work, hence I'd like to avoid that.
CU
knurd
8:38 a.m.
The old picard was broken for me (fingerprinting crashed it), but the
version in rpmfusion works. Fedora already has a version matching the
current picard-freeworld in Koji
(http://koji.fedoraproject.org/koji/buildinfo?buildID=140199).
I suppose picard-freeworld still needs to be rolled back until the
update is pushed in Fedora, I just wanted to mention that picard
0.12.1 *is* available in case anybody else is having the same trouble
with 0.11 as I was.
Another possibility if a 0.11 picard-freeworld doesn't show up might
be to temporarily add picard 0.12.1 to rpmfusion, since packages are
already built.
On Tue, Nov 10, 2009 at 3:25 AM, Thorsten Leemhuis <fedora(a)leemhuis.info> wrote:
Hi!
Our rawhide trees seem to be in a mostly sane state now. Here is a list
of steps I'm aware of before the Everything repos for F12 can be created:
- move the nvidia-driver to the nonfree-updates-testing repositories, as
requested by kwizart (#928)
- push the new em8300-kmod to the updates repos, once they exist; move
vdr-dxr3 there, as it depends on em8300-kmod (#920)
- push a older picard-freeworld, if one shows up in time (#932)
Anything else? If yes, then please speed up, as I'd would like to create
the Everything repos on Friday -- updating them afterwards is a lot of
manual work, hence I'd like to avoid that.
CU
knurd
1:05 p.m.
Thorsten Leemhuis wrote:
Anything else?
Can we get gsview moved to nonfree ?
https://bugzilla.rpmfusion.org/show_bug.cgi?id=856
Anything I can do to help make that happen?
-- Rex
1:57 p.m.
On Nov 10, 2009, at 3:25 AM, Thorsten Leemhuis wrote:
Hi!
Our rawhide trees seem to be in a mostly sane state now. Here is a
list
of steps I'm aware of before the Everything repos for F12 can be
created:
- move the nvidia-driver to the nonfree-updates-testing
repositories, as
requested by kwizart (#928)
- push the new em8300-kmod to the updates repos, once they exist;
move
vdr-dxr3 there, as it depends on em8300-kmod (#920)
- push a older picard-freeworld, if one shows up in time (#932)
Anything else? If yes, then please speed up, as I'd would like to
create
the Everything repos on Friday -- updating them afterwards is a lot of
manual work, hence I'd like to avoid that.
Just want to make sure the MythTV 0.22 release packages I pushed
through the build system yesterday make it into the release repo.
--
Jarod Wilson
jarod(a)wilsonet.com
10:37 p.m.
2009/11/10 Thorsten Leemhuis <fedora(a)leemhuis.info>:
Hi!
Our rawhide trees seem to be in a mostly sane state now. Here is a list
of steps I'm aware of before the Everything repos for F12 can be created:
- move the nvidia-driver to the nonfree-updates-testing repositories, as
requested by kwizart (#928)
One of the task that would help to solve this problem
with selinux
will be to review nvidia-xconfig and nvidia-settings, because thoses
will not be hitten by the selinux denials. (whereas the pre-built
equivalent will, as rpmlint shows).
For the record, the nvidia pre-built binaries have a compatibility
mode from linux 2.4.1 (whereas fedora 11 has linux 2.6.18). It also
means that the flag to verify that the binary doesn't requires
execstack is off (because it doesn't exist at all). I might workaround
it for libraries IIRC but I'm not aware of such workaround with
binaries.
Nicolas (kwizart)
3:04 p.m.
On Thursday, 12 November 2009 at 23:37, Nicolas Chauvet wrote:
2009/11/10 Thorsten Leemhuis <fedora(a)leemhuis.info>:
> Hi!
>
> Our rawhide trees seem to be in a mostly sane state now. Here is a list
> of steps I'm aware of before the Everything repos for F12 can be created:
>
> - move the nvidia-driver to the nonfree-updates-testing repositories, as
> requested by kwizart (#928)
One of the task that would help to solve this problem with selinux
will be to review nvidia-xconfig and nvidia-settings, because thoses
will not be hitten by the selinux denials. (whereas the pre-built
equivalent will, as rpmlint shows).
For the record, the nvidia pre-built binaries have a compatibility
mode from linux 2.4.1 (whereas fedora 11 has linux 2.6.18). It also
means that the flag to verify that the binary doesn't requires
execstack is off (because it doesn't exist at all). I might workaround
it for libraries IIRC but I'm not aware of such workaround with
binaries.
How about using execstack -c on them?
Regards,
R.
--
Fedora http://fedoraproject.org/wiki/User:Rathann
RPMFusion http://rpmfusion.org | MPlayer http://mplayerhq.hu
"Faith manages."
-- Delenn to Lennier in Babylon 5:"Confessions and Lamentations"
5:04 p.m.
2009/11/13 Dominik 'Rathann' Mierzejewski <dominik(a)greysector.net>:
On Thursday, 12 November 2009 at 23:37, Nicolas Chauvet wrote:
> 2009/11/10 Thorsten Leemhuis <fedora(a)leemhuis.info>:
> > Hi!
> >
> > Our rawhide trees seem to be in a mostly sane state now. Here is a list
> > of steps I'm aware of before the Everything repos for F12 can be created:
> >
> > - move the nvidia-driver to the nonfree-updates-testing repositories, as
> > requested by kwizart (#928)
> One of the task that would help to solve this problem with selinux
> will be to review nvidia-xconfig and nvidia-settings, because thoses
> will not be hitten by the selinux denials. (whereas the pre-built
> equivalent will, as rpmlint shows).
>
> For the record, the nvidia pre-built binaries have a compatibility
> mode from linux 2.4.1 (whereas fedora 11 has linux 2.6.18). It also
> means that the flag to verify that the binary doesn't requires
> execstack is off (because it doesn't exist at all). I might workaround
> it for libraries IIRC but I'm not aware of such workaround with
> binaries.
How about using execstack -c on them?
I forgot about this one, this will probably
help.
But they will need to stay in updates-testing for now. As this need
to be really tested.
If unsure prefer to use setsebool -P allow_execstack on
BTW nVidia has released support for legacy driver and xorg-server 1.7 yesterday.
/me working on it right now. (thoses will needs execstack too)
Nicolas (kwizart)
5:27 p.m.
2009/11/14 Nicolas Chauvet <kwizart(a)gmail.com>:
2009/11/13 Dominik 'Rathann' Mierzejewski
<dominik(a)greysector.net>:
> On Thursday, 12 November 2009 at 23:37, Nicolas Chauvet wrote:
>> 2009/11/10 Thorsten Leemhuis <fedora(a)leemhuis.info>:
>> > Hi!
>> >
>> > Our rawhide trees seem to be in a mostly sane state now. Here is a list
>> > of steps I'm aware of before the Everything repos for F12 can be
created:
>> >
>> > - move the nvidia-driver to the nonfree-updates-testing repositories, as
>> > requested by kwizart (#928)
>> One of the task that would help to solve this problem with selinux
>> will be to review nvidia-xconfig and nvidia-settings, because thoses
>> will not be hitten by the selinux denials. (whereas the pre-built
>> equivalent will, as rpmlint shows).
>>
>> For the record, the nvidia pre-built binaries have a compatibility
>> mode from linux 2.4.1 (whereas fedora 11 has linux 2.6.18). It also
>> means that the flag to verify that the binary doesn't requires
>> execstack is off (because it doesn't exist at all). I might workaround
>> it for libraries IIRC but I'm not aware of such workaround with
>> binaries.
>
> How about using execstack -c on them?
I forgot about this one, this will probably help.
But they will need to stay in updates-testing for now. As this need
to be really tested.
It won't works for i386 binaries whereas it works for
x86_64; and also
works for libraries of both arches:
+ find /builddir/build/BUILDROOT/xorg-x11-drv-nvidia-190.42-3.fc12.i386/usr/lib/nvidia
-name '*.so.*' -type f -exec execstack -c '{}' ';'
+ execstack -c
/builddir/build/BUILDROOT/xorg-x11-drv-nvidia-190.42-3.fc12.i386/usr/lib/xorg/modules/extensions/nvidia/libglx.so.190.42
+ execstack -c
/builddir/build/BUILDROOT/xorg-x11-drv-nvidia-190.42-3.fc12.i386/usr/lib/xorg/modules/drivers/nvidia_drv.so
+ execstack -c
/builddir/build/BUILDROOT/xorg-x11-drv-nvidia-190.42-3.fc12.i386/usr/bin/nvidia-settings
/builddir/build/BUILDROOT/xorg-x11-drv-nvidia-190.42-3.fc12.i386/usr/bin/nvidia-smi
execstack:
/builddir/build/BUILDROOT/xorg-x11-drv-nvidia-190.42-3.fc12.i386/usr/bin/nvidia-settings:
Reshuffling of objects to make room for
program header entry only supported for shared libraries
execstack:
/builddir/build/BUILDROOT/xorg-x11-drv-nvidia-190.42-3.fc12.i386/usr/bin/nvidia-smi:
Reshuffling of objects to make room for
program header entry only supported for shared libraries
error: Bad exit status from /var/tmp/rpm-tmp.m2qSy6 (%install)
RPM build errors:
Bad exit status from /var/tmp/rpm-tmp.m2qSy6 (%install)
Child returncode was: 1
EXCEPTION: Command failed. See logs for output.
9:08 p.m.
2009/11/10 Thorsten Leemhuis <fedora(a)leemhuis.info>:
Hi!
Our rawhide trees seem to be in a mostly sane state now. Here is a list
of steps I'm aware of before the Everything repos for F12 can be created:
- move the nvidia-driver to the nonfree-updates-testing repositories, as
requested by kwizart (#928)
- push the new em8300-kmod to the updates repos, once they exist; move
vdr-dxr3 there, as it depends on em8300-kmod (#920)
- push a older picard-freeworld, if one shows up in time (#932)
Anything else? If yes, then please speed up, as I'd would like to create
the Everything repos on Friday -- updating them afterwards is a lot of
manual work, hence I'd like to avoid that.
OK for me once we have the last Cg
and mock-rpmfusion* in GA repository.
About what can go in updates-testing.
At this time I would prevent 190.42-3 to go out of the buildsys. but
even if that's done
it will have to be in updates-testing as the execstack -c is
incomplete and unsafe for various reason.
Now i have 173xx and 96xx ready to be imported. So I will import now,
but it will be the same as the main serie. (prevent to goes out exept
to go eventually on updates-testing).
Nicolas (kwizart)
ps: I've saw that firewing1 missed to remove the faac BuildRequirement
on avidemux. I've warned him.
IIRC there is one package that failed a rebuild against our ffmpeg,
this need to be removed for now as the reason why it has failed is
very dubious.
6:05 p.m.
On 14.11.2009 22:08, Nicolas Chauvet wrote:
2009/11/10 Thorsten Leemhuis <fedora(a)leemhuis.info>:
> Our rawhide trees seem to be in a mostly sane state now. Here is a list
> of steps I'm aware of before the Everything repos for F12 can be created:
[...]
> Anything else? If yes, then please speed up, as I'd would like to create
> the Everything repos on Friday -- updating them afterwards is a lot of
> manual work, hence I'd like to avoid that.
OK for me once we have the last Cg and mock-rpmfusion* in GA repository.
Just for completeness: I think it's totally ^w very bad to build
packages on the very last minute for the F12 repos without strict need,
as they now will be in there for the whole lifetime of F12 now, even if
they contain serious bugs (¹).
Nevertheless I pushed them, as that obviously was what you wanted. But
But I'd vote that we for the next release have a real repo freeze for a
week or two, where only those things get changed in the last minute
where the benefit outweighs the risk.
Just my 2 cent.
Cu
knurd
(¹) that's not completely trues, as it's possible to get them out there,
just a lot of manual work that I would have to do
5276
days inactive
5281
days old
rpmfusion-developers@lists.rpmfusion.org
9 comments
6 participants
participants (6)
-
Dominik 'Rathann' Mierzejewski -
Jarod Wilson -
Justin -
Nicolas Chauvet -
Rex Dieter -
Thorsten Leemhuis