[Buildroot] [PATCH] package/modem-manager: use libqmi and libmbim if they are selected

[Arnout Vandecappelle]

On 04/11/2019 22:05, Thomas Petazzoni wrote:
> Hello,
> 
> Adding our Config.in.legacy expert Arnout.
> 
> On Mon,  4 Nov 2019 09:48:44 -0300
> unixmania at gmail.com wrote:
> 
>> diff --git a/Config.in.legacy b/Config.in.legacy
>> index fcb10b1291..3336ef85d1 100644
>> --- a/Config.in.legacy
>> +++ b/Config.in.legacy
>> @@ -146,6 +146,22 @@ endif
>>  
>>  comment "Legacy options removed in 2019.11"
>>  
>> +config BR2_PACKAGE_MODEM_MANAGER_LIBMBIM
>> +	bool "BR2_PACKAGE_MODEM_MANAGER_LIBMBIM has been removed"
>> +	select BR2_LEGACY
>> +	select BR2_PACKAGE_LIBMBIM
>> +	help
>> +	  Select BR2_PACKAGE_LIBMBIM to enable support for MBIM protocol
>> +	  in ModemManager
>> +
>> +config BR2_PACKAGE_MODEM_MANAGER_LIBQMI
>> +	bool "BR2_PACKAGE_MODEM_MANAGER_LIBQMI has been removed"
>> +	select BR2_LEGACY
>> +	select BR2_PACKAGE_LIBQMI
>> +	help
>> +	  Select BR2_PACKAGE_LIBQMI to enable support for QMI protocol
>> +	  in ModemManager
> 
> I think in this specific case, we don't need Config.in.legacy handling.
> Indeed, a full .config that had BR2_PACKAGE_MODEM_MANAGER_LIBQMI
> enabled will already have BR2_PACKAGE_LIBQMI, and since the .mk file
> now relies on BR2_PACKAGE_LIBQMI to enable libqmi support in
> modem-manager, I think we're good.
> 
> Of course, that won't work for defconfigs, but I believe updating with
> a defconfig is not really guaranteed to give the right result. Arnout ?

 In my point of view, indeed it isn't, but this isn't documented anywhere I think.

 So indeed, I agree that no legacy handling is needed here. It will annoy many
users (because they have to go and disable it in the legacy menu), while being
of no benefit for most of them.

 Regards,
 Arnout


> Other than that, patch looks good to me. No need to resend, let's just
> wait to see what Arnout says about the Config.in.legacy handling.
> 
> Best regards,
> 
> Thomas
>