Re: [PATCH] rename SOC_FAMILY to TARGET_SOC_FAMILY and make it global

[Tom Rini <tom_rini@mentor.com>]

Denys Dmytriyenko wrote:
> On Wed, Sep 08, 2010 at 07:58:39AM +0200, Frans Meulenbroeks wrote:
>> 2010/9/7 Maupin, Chase <chase.maupin@ti.com>:
>>
>>>> Forgot about this, but a variable should go with documentation.
>>>> This already has been raised with the orginal patch, but was ignored
>>>> by the submitters.
>>> Frans, Are you counting me as one of the original submitters?  I mainly 
>>> use the SOC_FAMILY, I didn't create it.  That being said, if you have a 
>>> recommended section of the usermanual that this should be documented in 
>>> I'll try to add something.
>> Hi Chase,
>> I see you indeed as the submitter. The commit carries your name and
>> there is no reference in the commit message of another author.
>>
>> Wrt the manual: actually I thought we had a section for this, but as
>> it stands i cannot find it. Guess I saw it somewhere else.
>> I'm not really sure where it would fit. I guess there are others who
>> have a better view on the manual and its structure.
> 
> Frans,
> 
> I guess then you haven't read the original discussion closely enough. Chase 
> only made a change to add existing SOC_FAMILY override into COMPATIBLE_MACHINE 
> variable... I keep repeating myself - SOC_FAMILY feature is about a year old. 
> It was introduced as part of a large merge from Arago. Here is the original 
> commit and some description:
> 
> http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=b788499815bfc1974d4277c0a261b9a7c24aa816

I suppose you could save Chase some time, Denys, and update the manual 
instead :)  But to me, it's perfectly reasonable for Frans to ask that 
when someone renames a not previously documented variable to go ahead 
and document it too (since they must have some knowledge of how it 
works, and some correct knowledge being written down is better than none).

-- 
Tom Rini
Mentor Graphics Corporation