9:55 a.m.
Hi!
There is a build issue with xine on F8; the same problem will afaics
show up in mythtv for F8(¹). Does anybody have a few spare cycles and is
willing to look into this? I'm not sure if it's a general yum/mock
problem, something specific to our builders or something wrong in the
Fedora repos.
I tend to say "builders", as it doesn't happen on ppc (which uses Fedora
as host for the builder; the x86 builders use centos):
http://buildsys.rpmfusion.org/logs/fedora-8-rpmfusion_free/441-xine-0.99....
CU
knurd
(¹) happened in the livna buildsys a few days ago already
10:40 a.m.
On Tue, 19 Aug 2008 11:55:42 +0200, Thorsten Leemhuis wrote:
Hi!
There is a build issue with xine on F8; the same problem will afaics
show up in mythtv for F8(¹). Does anybody have a few spare cycles and is
willing to look into this? I'm not sure if it's a general yum/mock
problem, something specific to our builders or something wrong in the
Fedora repos.
I tend to say "builders", as it doesn't happen on ppc (which uses Fedora
as host for the builder; the x86 builders use centos):
http://buildsys.rpmfusion.org/logs/fedora-8-rpmfusion_free/441-xine-0.99....
CU
knurd
(¹) happened in the livna buildsys a few days ago already
What happens here is that lirc-0.8.3-1.fc8 is not seen. Instead,
the old lirc-0.8.2-2.fc8 from stock F8 is taken.
Also for ppc, btw!
--
$ rpm -qa 'lirc*'
lirc-devel-0.8.3-1.fc8
lirc-libs-0.8.3-1.fc8
lirc-0.8.3-1.fc8
10:53 a.m.
New subject: [Fwd: Build Error (Job 441): xine-0_99_5-2_fc8 on fedora-8-rpmfusion_free]
Thorsten Leemhuis wrote:
Hi!
There is a build issue with xine on F8; the same problem will afaics
show up in mythtv for F8(¹). Does anybody have a few spare cycles and is
willing to look into this? I'm not sure if it's a general yum/mock
problem, something specific to our builders or something wrong in the
Fedora repos.
I tend to say "builders", as it doesn't happen on ppc (which uses Fedora
as host for the builder; the x86 builders use centos):
http://buildsys.rpmfusion.org/logs/fedora-8-rpmfusion_free/441-xine-0.99....
<snip>
Transaction Check Error:
file /usr/lib/liblirc_client.so.0.2.0 conflicts between attempted installs of
lirc-0.8.2-2.fc8 and lirc-libs-0.8.3-1.fc8
I believe this is due to a rather old yum on the buildsys. Between F-8 release
and F-8 updates lirc was split in to lirc and lirc-libs. xine BuildRequires
lirc-devel, which requires /usr/lib/liblirc_client.so.0.2.0, the old yum on the
builder has the old behavior when multiple packages provide a missing dep, to
install all providing packages, so in this case both lirc-0.8.2-2.fc8 (from F-8
release) and lirc-libs-0.8.3-1.fc8 (from F-8 updates) it should of course never
install an older lirc when a newer is available in updates (which does not
contain the lib) but it does.
Possible work arounds:
1: Add a BuildRequires on lirc-libs hoping it gets into the transaction before
the lirc-devel BuildRequires so that yum doesn't go looking for
/usr/lib/liblirc_client.so.0.2.0 as that is already there.
2: Add a BuildRequires on lirc, so that the newest lirc becomes part of the
transaction, this does not depend on dep resolv orderering (I hope) so is
probably the better workaorund.
Regards,
Hans
11:21 a.m.
New subject: [Fwd: Build Error (Job 441): xine-0_99_5-2_fc8 on fedora-8-rpmfusion_free]
On 19.08.2008 12:53, Hans de Goede wrote:
Thorsten Leemhuis wrote:
> There is a build issue with xine on F8; the same problem will afaics
> show up in mythtv for F8(¹). Does anybody have a few spare cycles and is
> willing to look into this? I'm not sure if it's a general yum/mock
> problem, something specific to our builders or something wrong in the
> Fedora repos.
>
> I tend to say "builders", as it doesn't happen on ppc (which uses
Fedora
> as host for the builder; the x86 builders use centos):
Just a note I forgot in my initial mail: the same xine spec file works
on F9 and devel.
>
http://buildsys.rpmfusion.org/logs/fedora-8-rpmfusion_free/441-xine-0.99....
<snip>
> Transaction Check Error:
> file /usr/lib/liblirc_client.so.0.2.0 conflicts between attempted installs of
lirc-0.8.2-2.fc8 and lirc-libs-0.8.3-1.fc8
I believe this is due to a rather old yum on the buildsys.
Do you really believe that we set up a new builder with a totally
ancient version of yum? ;-) K, sure, the one from CentOS5 is not that
new, but should be good enough IMHO.
Possible work arounds:
1: Add a BuildRequires on lirc-libs hoping it gets into the transaction before
the lirc-devel BuildRequires so that yum doesn't go looking for
/usr/lib/liblirc_client.so.0.2.0 as that is already there.
2: Add a BuildRequires on lirc, so that the newest lirc becomes part of the
transaction, this does not depend on dep resolv orderering (I hope) so is
probably the better workaorund.
Sounds like a good idea. I'll give it a try later.
CU
knurd
5:59 p.m.
New subject: [Fwd: Build Error (Job 441): xine-0_99_5-2_fc8 on fedora-8-rpmfusion_free]
On 19.08.2008 13:21, Thorsten Leemhuis wrote:
On 19.08.2008 12:53, Hans de Goede wrote:
> Thorsten Leemhuis wrote:
> Possible work arounds:
> 1: Add a BuildRequires on lirc-libs hoping it gets into the transaction before
> the lirc-devel BuildRequires so that yum doesn't go looking for
> /usr/lib/liblirc_client.so.0.2.0 as that is already there.
Seems that did the trick (at least on ppc). thx for the pointer.
CU
knurd
11:23 a.m.
New subject: [Fwd: Build Error (Job 441): xine-0_99_5-2_fc8 on fedora-8-rpmfusion_free]
mkdir -p /var/lib/mock/
Thorsten Leemhuis wrote:
Hi!
There is a build issue with xine on F8; the same problem will afaics
show up in mythtv for F8(¹). Does anybody have a few spare cycles and
is willing to look into this? I'm not sure if it's a general yum/mock
problem, something specific to our builders or something wrong in the
Fedora repos.
I tend to say "builders", as it doesn't happen on ppc (which uses
Fedora as host for the builder; the x86 builders use centos):
http://buildsys.rpmfusion.org/logs/fedora-8-rpmfusion_free/441-xine-0.99....
CU
knurd
(¹) happened in the livna buildsys a few days ago already
5980
days inactive
5980
days old
rpmfusion-developers@lists.rpmfusion.org
5 comments
4 participants
participants (4)
-
Hans de Goede -
Itamar -
Michael Schwendt -
Thorsten Leemhuis