[oe] `PR = "r0"`: Add or not to add?

[Frans Meulenbroeks]

2011/9/1 Philip Balister <philip at balister.org>

> On 08/31/2011 06:43 AM, Otavio Salvador wrote:
>
>> On Wed, Aug 31, 2011 at 10:30, Koen Kooi<koen at dominion.thruhere.**net<koen at dominion.thruhere.net>>
>>  wrote:
>>
>>> Besides, this is also on the policy
>>>>
>>>
>>> The OE classic policy maybe, not on the meta-oe policy[1].
>>>
>>
>> ETOOMANYPOLICIES
>>
>>  so or you change the policy or you follow it. Otherwise makes no sense to
>>>> have a policy if the person who merge the patches do not follow it.
>>>>
>>>
>>> - From http://wiki.openembedded.org/**index.php/Category:Policy<http://wiki.openembedded.org/index.php/Category:Policy>I read:
>>>
>>> http://wiki.openembedded.org/**index.php/Commit_Patch_**
>>> Message_Guidelines<http://wiki.openembedded.org/index.php/Commit_Patch_Message_Guidelines>
>>> http://wiki.openembedded.org/**index.php/Commit_Policy<http://wiki.openembedded.org/index.php/Commit_Policy>
>>> http://wiki.openembedded.org/**index.php/Commit_log_example<http://wiki.openembedded.org/index.php/Commit_log_example>
>>> http://wiki.openembedded.org/**index.php/Styleguide<http://wiki.openembedded.org/index.php/Styleguide>
>>> http://wiki.openembedded.org/**index.php/Versioning_Policy<http://wiki.openembedded.org/index.php/Versioning_Policy>
>>>
>>> And none of those say PR = r0 is wanted behaviour. OTOH it doesn't say
>>> it's
>>> unwanted either.
>>>
>>
>> So until it is explicit we might keep PR = r0 as it seems most people
>> prefer this way.
>>
>
> As long as we are all chiming in, I would like to see all of us decide once
> and for all what to do.
>
> I also prefer the explicitly setting PR = "r0". Can we work out the
> troublesome cases and see if we can get them fixed? Also, what is the
> oe-core view on this?
>
> Let's work out what the technical answer is and go from there


Good plan.

BTW not sure if it is feasible but ideally a 2nd assingment to a var should
give an error (or at least a warning). This avoids that e.g. a recipe and a
.inc file set the var (or that a var is set twice within the same recipe, in
the past I have seen this). If a variable is intended to be reassigned, weak
binding should be used. ( ?= or so).

Frans