On Mon, 2021-02-01 at 11:34 +0000, Richard Purdie wrote:
> On Sun, 2021-01-31 at 11:56 +0000, Richard Purdie via
> lists.openembedded.org wrote:
> > On Sun, 2021-01-31 at 11:41 +0000, Richard Purdie via
> > lists.openembedded.org wrote:
> > > On Wed, 2021-01-20 at 13:03 +0800, Chen Qi wrote:
> > > > Build error occurs as below:
> > > >
> > > > ERROR: igt-gpu-tools-1.25+gitAUTOINC+d16ad07e7f-r0
> > > > do_package_qa:
> > > > QA Issue: /usr/lib/libigt.so.0 contained in package
> > > > igt-gpu-tools
> > > > requires libprocps.so.8(LIBPROCPS_0)(64bit), but no
> > > > providers
> > > > found in RDEPENDS_igt-gpu-tools? [file-rdeps]
> > > >
> > > > Add procps-lib to RDEPENDS to solve this issue.
> > > >
> > > > Signed-off-by: Chen Qi <qi.c...@windriver.com>
> > > > ---
> > > > meta/recipes-graphics/igt-gpu-tools/igt-gpu-tools_git.bb | 2
> > > > +-
> > > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/meta/recipes-graphics/igt-gpu-tools/igt-gpu-
> > > > tools_git.bb b/meta/recipes-graphics/igt-gpu-tools/igt-gpu-
> > > > tools_git.bb
> > > > index 89480d79d2..b012dbf58e 100644
> > > > --- a/meta/recipes-graphics/igt-gpu-tools/igt-gpu-tools_git.bb
> > > > +++ b/meta/recipes-graphics/igt-gpu-tools/igt-gpu-tools_git.bb
> > > > @@ -15,7 +15,7 @@ SRC_URI =
> > > > "git://gitlab.freedesktop.org/drm/igt-gpu-
> > > > tools.git;protocol=https"
> > > > S = "${WORKDIR}/git"
> > > >
> > > > DEPENDS += "libdrm libpciaccess cairo udev glib-2.0 procps
> > > > libunwind kmod openssl elfutils alsa-lib json-c bison-native"
> > > > -RDEPENDS_${PN} += "bash"
> > > > +RDEPENDS_${PN} += "bash procps-lib"
> > > > RDEPENDS_${PN}-tests += "bash"
> > > >
> > >
> > >
> > > You are correct about this and I see the same issue if I build it
> > > locally. What worries me is why we don't see this on the
> > > autobuilder in
> > > world builds?
> > >
> > > Cc'ing Anuj as the maintainer.
> >
> > Looking at the autobuilder logs, I think I can guess the problem, I
> > think its a hash equivalence issue.
> >
> > Errors raised by do_package_qa aren't reflected in the output so it
> > would error once and then never again as the output wouldn't seen
> > changed next time :(
> >
> > We'll have to fix this, its nasty.
> >
> > Cc: Joshua
>
> Its not a hashequiv issue directly although the QA warnings not being
> cached is am unrelated problem.
>
> It seems if you build igt-gpu-tools with /usr/lib64/ it works (as on
> the autobuilder where multilib is enabled), if you build with
> /usr/lib/
> it gives the QA warning despite actually having the right RDEPENDS
> anyway.
I see the same failure locally even when using the autobuilder
configuration so it's probably not related to multilib.
I guess it's happening because procps was split up recently to have the
lib installed in a separate package and that didn't result in a rebuild
of igt-gpu-tools?
Thanks,
Anuj
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#147548):
https://lists.openembedded.org/g/openembedded-core/message/147548
Mute This Topic: https://lists.openembedded.org/mt/79972258/21656
Group Owner: openembedded-core+ow...@lists.openembedded.org
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
