> -----Original Message-----
> From: Bruce Ashfield <[email protected]>
> Sent: den 16 januari 2024 16:22
> To: Peter Kjellerstedt <[email protected]>
> Cc: [email protected]
> Subject: Re: [meta-virtualization] [PATCH] containerd-opencontainers: Use
> RDEPENDS:${PN} instead of RDEPENDS:${BPN}
>
> On Tue, Jan 16, 2024 at 6:35 AM Peter Kjellerstedt
> <[email protected]> wrote:
> >
> > > -----Original Message-----
> > > From: Bruce Ashfield <[email protected]>
> > > Sent: den 15 januari 2024 18:10
> > > To: Peter Kjellerstedt <[email protected]>
> > > Cc: [email protected]
> > > Subject: Re: [meta-virtualization] [PATCH] containerd-opencontainers: Use
> > > RDEPENDS:${PN} instead of RDEPENDS:${BPN}
> > >
> > > On Mon, Jan 15, 2024 at 12:00 PM Peter Kjellerstedt
> > > <[email protected]> wrote:
> > > >
> > > > This avoids the following warning:
> > > >
> > > > WARNING: .../containerd-opencontainers_git.bb: Variable key
> > > > RDEPENDS:${PN} (...) replaces original key
> > > > RDEPENDS:containerd-opencontainers
> > > > ( ${VIRTUAL-RUNTIME_container_runtime}).
> > > >
> > > > if, e.g., RDEPENDS:${PN}:append = " ..." is used in a bbappend or
> > > > bbclass.
> > >
> > > All that the above commit message does is explain the syntax
> > > of what you are fixing .. but not why.
> > >
> > > We had the BPN rdepends on purpose, showing why that
> > > original purpose is not needed ... should be in the commit
> > > message.
> >
> > I have no idea why it was RDEPENDS:${BPN} rather than RDEPENDS:${PN},
> > I only assumed it was a typo. There is nothing in the original commit
> > that indicates that it was done on purpose.
>
> I did a bit more digging, and at the time there were x32 multilibs at play,
> and
> I remember that solving some sort of inter package dependency .. but I
> also can't find the details.
I have no experience with x32, so not something I'd have considered.
I only looked at the recipe and couldn't find any references to
BBCLASSEXTEND or similar that would explain why ${BPN} would make a
difference compared to ${PN}.
> > > My answer right now would be .. .don't use PN in your bbappend
> > > and you can avoid the warning.
> >
> > The problem in my case is that our addition to RDEPENDS:${PN} is done in
> > a common bbclass, so I cannot control it per recipe.
>
> Ah, so you have a class that is doing a generic RDEPENDS addition
> to all packages and that triggers the double variable key warning.
>
> I'll tweak the commit message a bit to explain the history and trigger
> of the warning and queue it up.
Thank you. Can you also backport it to Nanbield (and possibly also other
branches), once it has made it into master?
//Peter
> Bruce
>
> >
> > //Peter
> >
> > >
> > > Bruce
> > >
> > > > Signed-off-by: Peter Kjellerstedt <[email protected]>
> > > > ---
> > > > recipes-containers/containerd/containerd-opencontainers_git.bb | 2 +-
> > > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git
> > > > a/recipes-containers/containerd/containerd-opencontainers_git.bb
> > > > b/recipes-containers/containerd/containerd-opencontainers_git.bb
> > > > index 5c1e127..1dadf28 100644
> > > > --- a/recipes-containers/containerd/containerd-opencontainers_git.bb
> > > > +++ b/recipes-containers/containerd/containerd-opencontainers_git.bb
> > > > @@ -95,6 +95,6 @@ INSANE_SKIP:${PN} += "ldflags already-stripped"
> > > >
> > > > COMPATIBLE_HOST = "^(?!(qemu)?mips).*"
> > > >
> > > > -RDEPENDS:${BPN} += " ${VIRTUAL-RUNTIME_container_runtime}"
> > > > +RDEPENDS:${PN} += " ${VIRTUAL-RUNTIME_container_runtime}"
> > > >
> > > > CVE_PRODUCT = "containerd"
> > > >
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#8545):
https://lists.yoctoproject.org/g/meta-virtualization/message/8545
Mute This Topic: https://lists.yoctoproject.org/mt/103743235/21656
Group Owner: [email protected]
Unsubscribe: https://lists.yoctoproject.org/g/meta-virtualization/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
