[PATCH] multilib_header: avoid sstate checksum issues for -nativesdk recipes

[PATCH] multilib_header: avoid sstate checksum issues for -nativesdk recipes

[Joshua Lock]

Much as with -native recipes, as addressed in commit
b15730caf0d4c40271796887505507f2501958bb, arch specific variables
like MIPSPKGSFX_ABI were affecting -nativesdk sstate checksums for
recipes like nativesdk-glibc-initial.

Disable multilib_header for nativesdk as we don't use multilibs in
this scenario.

[YOCTO #10320]

Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
---
 meta/classes/multilib_header.bbclass | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/meta/classes/multilib_header.bbclass b/meta/classes/multilib_header.bbclass
index 5ee0a2d..304c28e 100644
--- a/meta/classes/multilib_header.bbclass
+++ b/meta/classes/multilib_header.bbclass
@@ -52,3 +52,8 @@ oe_multilib_header() {
 oe_multilib_header_class-native () {
 	return
 }
+
+# Nor do we need multilib headers for nativesdk builds.
+oe_multilib_header_class-nativesdk () {
+	return
+}
-- 
2.7.4

Re: [PATCH] multilib_header: avoid sstate checksum issues for -nativesdk recipes

[akuster808]

On 09/30/2016 02:09 AM, Joshua Lock wrote:
> Much as with -native recipes, as addressed in commit
> b15730caf0d4c40271796887505507f2501958bb, arch specific variables
> like MIPSPKGSFX_ABI were affecting -nativesdk sstate checksums for
> recipes like nativesdk-glibc-initial.
Since commit

b15730caf0d4c40271796887505507f2501958bb is in Jethro and Krogoth, do both of those branch need this fix?

- armin

>
> Disable multilib_header for nativesdk as we don't use multilibs in
> this scenario.
>
> [YOCTO #10320]
>
> Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
> ---
>   meta/classes/multilib_header.bbclass | 5 +++++
>   1 file changed, 5 insertions(+)
>
> diff --git a/meta/classes/multilib_header.bbclass b/meta/classes/multilib_header.bbclass
> index 5ee0a2d..304c28e 100644
> --- a/meta/classes/multilib_header.bbclass
> +++ b/meta/classes/multilib_header.bbclass
> @@ -52,3 +52,8 @@ oe_multilib_header() {
>   oe_multilib_header_class-native () {
>   	return
>   }
> +
> +# Nor do we need multilib headers for nativesdk builds.
> +oe_multilib_header_class-nativesdk () {
> +	return
> +}

Re: [PATCH] multilib_header: avoid sstate checksum issues for -nativesdk recipes

[Joshua Lock]

On Fri, 2016-09-30 at 08:29 -0700, akuster808 wrote:
> 
> On 09/30/2016 02:09 AM, Joshua Lock wrote:
> > 
> > Much as with -native recipes, as addressed in commit
> > b15730caf0d4c40271796887505507f2501958bb, arch specific variables
> > like MIPSPKGSFX_ABI were affecting -nativesdk sstate checksums for
> > recipes like nativesdk-glibc-initial.
> Since commit
> 
> b15730caf0d4c40271796887505507f2501958bb is in Jethro and Krogoth, do
> both of those branch need this fix?

This patch improves sstate object reuse for nativesdk recipes when
switching between a mips and non-mips MACHINE. 
It's no doubt useful to have in krogoth and jethro too.

Joshua

> - armin
> 
> > 
> > 
> > Disable multilib_header for nativesdk as we don't use multilibs in
> > this scenario.
> > 
> > [YOCTO #10320]
> > 
> > Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
> > ---
> >   meta/classes/multilib_header.bbclass | 5 +++++
> >   1 file changed, 5 insertions(+)
> > 
> > diff --git a/meta/classes/multilib_header.bbclass
> > b/meta/classes/multilib_header.bbclass
> > index 5ee0a2d..304c28e 100644
> > --- a/meta/classes/multilib_header.bbclass
> > +++ b/meta/classes/multilib_header.bbclass
> > @@ -52,3 +52,8 @@ oe_multilib_header() {
> >   oe_multilib_header_class-native () {
> >   	return
> >   }
> > +
> > +# Nor do we need multilib headers for nativesdk builds.
> > +oe_multilib_header_class-nativesdk () {
> > +	return
> > +}
>