nativesdk-buildtools-perl-dummy.bb: disable ICECC

Submitted by Alex Franco on March 22, 2016, 4:37 p.m. | Patch ID: 118479

Details

Message ID 1458664665-20517-1-git-send-email-alejandro.franco@linux.intel.com
State New
Headers show

Commit Message

Alex Franco March 22, 2016, 4:37 p.m.
Disable ICECC, as this recipe was added for compatibility with
the extensible SDK (eSDK), and therefore does not need to be
able to be built with the distributed icecream compiler.

[YOCTO #8934]

Signed-off-by: Alex Franco <alejandro.franco@linux.intel.com>
---
 meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb | 1 +
 1 file changed, 1 insertion(+)

Patch hide | download patch | download mbox

diff --git a/meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb b/meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb
index 90784a9..b488b03 100644
--- a/meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb
+++ b/meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb
@@ -18,6 +18,7 @@  PERLPACKAGES = "nativesdk-perl \
                 nativesdk-perl-module-file-path"
 
 ALLOW_EMPTY_${PN} = "1"
+ICECC_DISABLED = "1"
 
 python populate_packages_prepend() {
     d.appendVar(d.expand('RPROVIDES_${PN}'), '${PERLPACKAGES}')

Comments

Alex Franco March 22, 2016, 4:40 p.m.
Fix originally submitted on bugzilla [1] by Andreas Schultz, I get good
results with it and I do think that ICECC should be disabled for this
recipe...

Alex Franco

[1] https://bugzilla.yoctoproject.org/show_bug.cgi?id=8934

On 03/22/2016 10:37 AM, Alex Franco wrote:
> Disable ICECC, as this recipe was added for compatibility with
> the extensible SDK (eSDK), and therefore does not need to be
> able to be built with the distributed icecream compiler.
>
> [YOCTO #8934]
>
> Signed-off-by: Alex Franco <alejandro.franco@linux.intel.com>
> ---
>  meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb b/meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb
> index 90784a9..b488b03 100644
> --- a/meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb
> +++ b/meta/recipes-core/meta/nativesdk-buildtools-perl-dummy.bb
> @@ -18,6 +18,7 @@ PERLPACKAGES = "nativesdk-perl \
>                  nativesdk-perl-module-file-path"
>  
>  ALLOW_EMPTY_${PN} = "1"
> +ICECC_DISABLED = "1"
>  
>  python populate_packages_prepend() {
>      d.appendVar(d.expand('RPROVIDES_${PN}'), '${PERLPACKAGES}')
Ross Burton March 22, 2016, 4:41 p.m.
On 22 March 2016 at 16:40, Alex Franco <alejandro.franco@linux.intel.com>
wrote:

> Fix originally submitted on bugzilla [1] by Andreas Schultz, I get good
> results with it and I do think that ICECC should be disabled for this
> recipe...
>

Do we understand why this recipe breaks the icecc class?

Ross
Ross Burton March 22, 2016, 4:45 p.m.
On 22 March 2016 at 16:41, Burton, Ross <ross.burton@intel.com> wrote:

> Do we understand why this recipe breaks the icecc class?
>

So I'm guessing it's because the architecture of the recipe is "special" -
it's sort-of allarch and sort of nativesdk.

I wonder if changing the icecc_is_allarch() test to check for inherit of
allarch would fix this without working around the problem.

Ross
Alex Franco March 22, 2016, 6:21 p.m.
Ross, I suppose you mean " icc_is_allarch()" ?

Ok , so this recipe does inherit allarch but it is sort of nativesdk,
and when the test icc_is_allarch() checks for the value of
'PACKAGE_ARCH', it gets "buildtools-dummy-${SDKPKGSUFFIX}" (which should
expand to 'buildtools-dummy-nativesdk', but doesn't) instead of the
expected "all".

I am now modifying this test to check for an inherit of allarch, and
playing a bit with this. I'll keep you posted.

Other stuff to consider:

- Expand the value of PACKAGE_ARCH if necessary? I am doing it now as
part of my tests but not sure I will be committing this:

     package_arch = d.getVar("PACKAGE_ARCH", False)
     if '${' in package_arch:
         package_arch = d.expand(package_arch)
    ...

- All of these functions in the icecc.bbclass that incude "icc" in their
names, should probably be modified to say "icecc" instead. As it stands
now, it can cause confusion as some people might think we're talking
about the intel c compiler rather than the icecream compiler - I wonder
if there's a bug for this already

Alex Franco

On 03/22/2016 10:45 AM, Burton, Ross wrote:
>
> On 22 March 2016 at 16:41, Burton, Ross <ross.burton@intel.com
> <mailto:ross.burton@intel.com>> wrote:
>
>     Do we understand why this recipe breaks the icecc class?
>
>
> So I'm guessing it's because the architecture of the recipe is
> "special" - it's sort-of allarch and sort of nativesdk.
>
> I wonder if changing the icecc_is_allarch() test to check for inherit
> of allarch would fix this without working around the problem.
>
> Ross 
>