[v1] devicetree.bbclass: fix setting S

From: Max Krummenacher <max.krummenacher@toradex.com>

When a user of devicetree.bbclass overrides setting S, e.g. to
git/overlays then having UNPACKDIR set to S results in that
git repo unpacked inside UNPACKDIR. Thus the sources are actually
in git/overlays/git/overlays rather than the expected git/overlays.

Change the logic, so that S by default is set to UNPACKDIR,
but a user can change S to point to were the sources actually
will be.

Fixes: 7a06069a041a ("devicetree.bbclass: switch away from S = WORKDIR")
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
---
 meta/classes-recipe/devicetree.bbclass | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Hello

On Tue, Jun 25, 2024 at 08:30:46PM +0200, max.oss.09@gmail.com wrote:
> From: Max Krummenacher <max.krummenacher@toradex.com>
> 
> When a user of devicetree.bbclass overrides setting S, e.g. to
> git/overlays then having UNPACKDIR set to S results in that
> git repo unpacked inside UNPACKDIR. Thus the sources are actually
> in git/overlays/git/overlays rather than the expected git/overlays.
> 
> Change the logic, so that S by default is set to UNPACKDIR,
> but a user can change S to point to were the sources actually
> will be.
> 
> Fixes: 7a06069a041a ("devicetree.bbclass: switch away from S = WORKDIR")

Gentle ping.

Any comments on this commit?
For me the referenced commit causes a regression. If I understand it
correctly it will do so for any user which uses the devicetree.bbclass
and did set S to anything other than WORKDIR.

I think this should be fixed here rather than any recipe which sets S
also setting UNPACKDIR.

Thanks.
Max
> Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
> ---
>  meta/classes-recipe/devicetree.bbclass | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/meta/classes-recipe/devicetree.bbclass b/meta/classes-recipe/devicetree.bbclass
> index 1806cb62cbf7..e5dbd9a85714 100644
> --- a/meta/classes-recipe/devicetree.bbclass
> +++ b/meta/classes-recipe/devicetree.bbclass
> @@ -40,8 +40,8 @@ PACKAGE_ARCH = "${MACHINE_ARCH}"
>  SYSROOT_DIRS += "/boot/devicetree"
>  FILES:${PN} = "/boot/devicetree/*.dtb /boot/devicetree/*.dtbo"
>  
> -S = "${WORKDIR}/sources"
> -UNPACKDIR = "${S}"
> +UNPACKDIR = "${WORKDIR}/sources"
> +S = "${UNPACKDIR}"
>  B = "${WORKDIR}/build"
>  
>  # Default kernel includes, these represent what are normally used for in-kernel
> -- 
> 2.42.0
> 
>

On Tue, 2024-06-25 at 20:30 +0200, Max Krummenacher via
lists.openembedded.org wrote:
> From: Max Krummenacher <max.krummenacher@toradex.com>
> 
> When a user of devicetree.bbclass overrides setting S, e.g. to
> git/overlays then having UNPACKDIR set to S results in that
> git repo unpacked inside UNPACKDIR. Thus the sources are actually
> in git/overlays/git/overlays rather than the expected git/overlays.
> 
> Change the logic, so that S by default is set to UNPACKDIR,
> but a user can change S to point to were the sources actually
> will be.
> 
> Fixes: 7a06069a041a ("devicetree.bbclass: switch away from S =
> WORKDIR")
> Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
> ---
>  meta/classes-recipe/devicetree.bbclass | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/meta/classes-recipe/devicetree.bbclass b/meta/classes-
> recipe/devicetree.bbclass
> index 1806cb62cbf7..e5dbd9a85714 100644
> --- a/meta/classes-recipe/devicetree.bbclass
> +++ b/meta/classes-recipe/devicetree.bbclass
> @@ -40,8 +40,8 @@ PACKAGE_ARCH = "${MACHINE_ARCH}"
>  SYSROOT_DIRS += "/boot/devicetree"
>  FILES:${PN} = "/boot/devicetree/*.dtb /boot/devicetree/*.dtbo"
>  
> -S = "${WORKDIR}/sources"
> -UNPACKDIR = "${S}"
> +UNPACKDIR = "${WORKDIR}/sources"
> +S = "${UNPACKDIR}"
>  B = "${WORKDIR}/build"
>  
>  # Default kernel includes, these represent what are normally used
> for in-kernel

This doesn't seem to be the right thing to do to me. These settings are
really closely related to SRC_URI so I think they probably shouldn't be
in the class at all but in the recipes alongside SRC_URI?

Cheers,

Richard