[meta-oe][PATCH] glm: Fix packaging for header-only recipe

[meta-oe][PATCH] glm: Fix packaging for header-only recipe

[Tom Hochstein]

Set ALLOW_EMPTY for the main package to "1" for this header-only recipe. This
allows the SDK to be created properly when there is an RDEPENDS on glm.

Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
---
 meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb b/meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb
index e2f4dbebc..019e6257d 100644
--- a/meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb
+++ b/meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb
@@ -34,6 +34,7 @@ do_install() {
 
 }
 
-RDEPENDS_${PN}-dev = ""
+# This is a header-only library, so the main package will be empty.
+ALLOW_EMPTY_${PN} = "1"
 
 BBCLASSEXTEND = "native"
-- 
2.17.1

Re: [OE-core] [meta-oe][PATCH] glm: Fix packaging for header-only recipe

[Khem Raj]

[-- Attachment #1: Type: text/plain, Size: 1032 bytes --]

On Thu, Jan 28, 2021 at 5:44 PM Tom Hochstein <tom.hochstein@nxp.com> wrote:

> Set ALLOW_EMPTY for the main package to "1" for this header-only recipe.
> This
> allows the SDK to be created properly when there is an RDEPENDS on glm.


Runtime dependency on a header only package seems a bit iffy it’s perhaps
better to depend on glm-dev perhaps ?

>
>
> Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
> ---
>  meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb
> b/meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb
> index e2f4dbebc..019e6257d 100644
> --- a/meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb
> +++ b/meta-oe/recipes-graphics/glm/glm_0.9.9.6.bb
> @@ -34,6 +34,7 @@ do_install() {
>
>  }
>
> -RDEPENDS_${PN}-dev = ""
> +# This is a header-only library, so the main package will be empty.
> +ALLOW_EMPTY_${PN} = "1"
>
>  BBCLASSEXTEND = "native"
> --
> 2.17.1
>
>
> 
>
>

[-- Attachment #2: Type: text/html, Size: 2091 bytes --]

Re: [OE-core] [meta-oe][PATCH] glm: Fix packaging for header-only recipe

[Tom Hochstein]

On Thu, Jan 28, 2021 at 10:57 PM Khem Raj <mailto:raj.khem@gmail.com> wrote:
> 
> On Thu, Jan 28, 2021 at 5:44 PM Tom Hochstein <mailto:tom.hochstein@nxp.com> wrote:
> > Set ALLOW_EMPTY for the main package to "1" for this header-only recipe. This
> > allows the SDK to be created properly when there is an RDEPENDS on glm.
> 
> Runtime dependency on a header only package seems a bit iffy it's perhaps better to depend on glm-dev perhaps ?

Well, it's all iffy in my mind since these are just different workarounds for the root problem that header-only packages are not included in the SDK. If you think including the headers in the image is the better workaround, then I will switch to that.

Tom

Re: [OE-core] [meta-oe][PATCH] glm: Fix packaging for header-only recipe

[Otavio Salvador]

Em sex., 29 de jan. de 2021 às 13:06, Tom Hochstein
<tom.hochstein@nxp.com> escreveu:
> On Thu, Jan 28, 2021 at 10:57 PM Khem Raj <mailto:raj.khem@gmail.com> wrote:
> > On Thu, Jan 28, 2021 at 5:44 PM Tom Hochstein <mailto:tom.hochstein@nxp.com> wrote:
> > > Set ALLOW_EMPTY for the main package to "1" for this header-only recipe. This
> > > allows the SDK to be created properly when there is an RDEPENDS on glm.
> >
> > Runtime dependency on a header only package seems a bit iffy it's perhaps better to depend on glm-dev perhaps ?
>
> Well, it's all iffy in my mind since these are just different workarounds for the root problem that header-only packages are not included in the SDK. If you think including the headers in the image is the better workaround, then I will switch to that.

The allow empty seems to be the right workaround here as it allow
people to consistently use the `dev-pkgs` filter to populate the SDK.

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9 9981-7854          Mobile: +1 (347) 903-9750