[OE-core] [CONSOLIDATED PULL (V2) 00/35] Updates, Multilib Patches and other fixes

[Marko Lindqvist]

On 9 January 2013 12:29, Martin Jansa <martin.jansa at gmail.com> wrote:
> On Wed, Jan 09, 2013 at 12:00:20PM +0200, Marko Lindqvist wrote:
>> On 9 January 2013 11:42, Ross Burton <ross.burton at intel.com> wrote:
>> > On Wednesday, 9 January 2013 at 09:33, Marko Lindqvist wrote:
>> >> Exactly, except that it would be provided by bitbake, and not
>> >> constructed by each recipe itself - less recipe writing work + you
>> >> could count on it to always mean same thing.
>> >
>> > Everything but the last dot, or the first two components, or what?  For anything GNOMEy you want the first two elements as some packages use nano releases for development snapshots (1.2.3.4).
>>
>>  Ok, do we have any counter-examples where you'd want, say, three out
>> of four parts? One out of n? Of course there will always be need to
>> use variables of their own in some recipes, but I'm after something
>> that would be usable in most cases.
>
> I don't remember where but 1st of 4 is also used.
> What about stuff like "1.2.3.4+gitAUTOINC"?
>
> GNOME is quite consistent, but it still looks better defined in
> recipe/bbclass then having "something" defined by bitbake in all recipes
> and usable only in some.

 Given the examples, yes. Only if external tools would benefit (=
could use bitbake defined, but not recipe constructed var), defining
it bitbake side would make sense.


 - ML