[poky] RDEPENDS = "foo" in core layer

[Frans Meulenbroeks]

2010/12/22 Richard Purdie <richard.purdie at linuxfoundation.org>:
> On Wed, 2010-12-22 at 16:04 +0100, Koen Kooi wrote:
>> Op 22 dec 2010, om 15:49 heeft Richard Purdie het volgende geschreven:
>>
>> > On Wed, 2010-12-22 at 14:49 +0100, Koen Kooi wrote:
>> >> The core layer has tons of RDEPENDS = "foo", which should actually be
>> >> RDEPENDS_${PN}. What's the policy on those?
>> >
>> > They should be fixed. Patches very welcome...
>>
>> A quick sed gives me:
>>
>>  76 files changed, 77 insertions(+), 77 deletions(-)
>>
>> Do you want that as a single patch and if not, how do you want to have
>> it split and are PR bumps needed?
>
> One patch is fine. One tricky part is that this is a change in behaviour
> as an RDEPENDS is applied to every RDEPENDS_XXX, not just XXX=${PN} and
> that means we really should bump PR :/.
>
> Is there also a similar problem with RRECOMMENDS?
>
> Cheers,
>
> Richard
>

Yes.

This was discussed on the oe mailing list in oct
See http://thread.gmane.org/gmane.comp.handhelds.openembedded/37858/focus=37903
Martin Jansa did a lot of  work on this.
If you read the thread you'll see that there is a case in which the
_${PN} should not be added.
Not checked if this specific exception is also present in poky.

Frans