Re: [OE-core] [PATCH 1/1] base.bbclass: Fix dangling NATIVELSBSTRING

[Robert Yang <liezhi.yang@windriver.com>]

Hi RP,

On 2/14/23 16:44, Richard Purdie wrote:
> On Tue, 2023-02-14 at 00:28 -0800, Robert Yang wrote:
>> Fixed:
>> $ rm -fr tmp; bitbake quilt-native -n
>> Build Configuration:
>> [snip]
>> NATIVELSBSTRING      = "ubuntu-18.04"
>> [snip]
>>
>> And when run bitbake again:
>> $ bitbake quilt-native -n
>> Build Configuration:
>> NATIVELSBSTRING      = "universal"
>>
>> It has been changed from ubuntu-18.04 to universal on the same host and build
>> directory, this is because it is overridded by NATIVELSBSTRING. This patch
>> makes it print the correct value.
>>
>> Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
>> ---
>>   meta/classes-global/base.bbclass | 7 ++++++-
>>   1 file changed, 6 insertions(+), 1 deletion(-)
>>
>> diff --git a/meta/classes-global/base.bbclass b/meta/classes-global/base.bbclass
>> index 64e805c947..cf42d9d4d2 100644
>> --- a/meta/classes-global/base.bbclass
>> +++ b/meta/classes-global/base.bbclass
>> @@ -226,7 +226,12 @@ BUILDCFG_FUNCS[type] = "list"
>>   def buildcfg_vars(d):
>>       statusvars = oe.data.typed_value('BUILDCFG_VARS', d)
>>       for var in statusvars:
>> -        value = d.getVar(var)
>> +        # NATIVELSBSTRING var may have been overridden with "universal", so
>> +        # get actual host distribution id and version
>> +        if var == 'NATIVELSBSTRING':
>> +            value = lsb_distro_identifier(d)
>> +        else:
>> +            value = d.getVar(var)
>>           if value is not None:
>>               yield '%-20s = "%s"' % (var, value)
> 
> The uninative code runs "late" and this is an known issue, there is an
> open bug for it if I remember correctly.
> 
> I don't really want to put hacks into the BUILDCFG display code to work
> around it since it is just going to make it less clear what is going on
> and is also likely to break if we evern change the way NATIVELSBSTRING
> is set. If you really don't like this, we should fix it properly some
> other way rather than work around it.

I'd like to drop the patch since it doesn't affect the build.

// Robert

> 
> Cheers,
> 
> Richard