Bug 8880

Summary: Need better allarch multilib sanity or automated QA tests
Product: [Build System, Metadata & Runtime] OE-Core Reporter: Robert Yang <liezhi.yang>
Component: coreAssignee: Chen Qi <Qi.Chen>
Status: RESOLVED FIXED QA Contact:
Severity: enhancement    
Priority: Medium+ CC: alexandru.c.georgescu, jose.perez.carranza, meta.mr.watcher, meta.watcher, richard.purdie, ross.burton
Version: 2.1   
Target Milestone: 2.3 M1   
Hardware: x86   
OS: Multiple   
Whiteboard:
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: No (bug/feature does not impact docs)

Description Robert Yang 2016-01-06 07:13:00 UTC 
allarch pkg should not depend on TUNE_ARCH or MACHINE_ARCH pkg, we need a way to do the checking automatically.
Comment 1 Ross Burton 2016-01-07 15:35:06 UTC 
As per the thread, I'm not sure this is true.
Comment 2 Richard Purdie 2016-01-07 16:39:27 UTC 
It is ok for an allarch package to depend on a non allarch package as long as its listed in the layer.conf mappings.

I think we need to do something slightly different here, I think we need to extend the allarch sstate QA check to include checking a sample multilib configuration and detect these problems which is how we solved this problem in general for allarch.
Comment 3 Chen Qi 2016-07-22 02:13:43 UTC 
(In reply to comment #2)
> It is ok for an allarch package to depend on a non allarch package as long
> as its listed in the layer.conf mappings.
> 
> I think we need to do something slightly different here, I think we need to
> extend the allarch sstate QA check to include checking a sample multilib
> configuration and detect these problems which is how we solved this problem
> in general for allarch.

Hi Richard,

Below is my understanding about this bug. Do I understand it correctly?

Now that oe-selftest has checked sstate sigs between different MACHINE settings for allarch packages, we should also add a check for sstate sigs for allarch packages between different multilib settings.

So we should add a testcase to oe-selftest to make sure sstate sigs for allarch packages don't change when we change the setting of multilib. Right?

Best Regards,
Chen Qi
Comment 4 Richard Purdie 2016-07-22 09:16:16 UTC 
Hi Chen,

(In reply to comment #3)
> Below is my understanding about this bug. Do I understand it correctly?
> 
> Now that oe-selftest has checked sstate sigs between different MACHINE
> settings for allarch packages, we should also add a check for sstate sigs
> for allarch packages between different multilib settings.
> 
> So we should add a testcase to oe-selftest to make sure sstate sigs for
> allarch packages don't change when we change the setting of multilib. Right?

Yes, that is correct. This bug is about adding a similar sstate sigs test to oe-selftest which checks that allarch packages don't change between different multilib settings.

Cheers,

Richard