Re: [PATCH] gstreamer: make pulseaudio support dependent on DISTRO_FEATURES

[Paul Eggleton <paul.eggleton@linux.intel.com>]

On Wednesday 26 September 2012 15:30:13 Martin Jansa wrote:
> On Wed, Sep 26, 2012 at 02:24:07PM +0100, Paul Eggleton wrote:
> > On Wednesday 26 September 2012 14:32:57 Martin Jansa wrote:
> > > On Wed, Sep 26, 2012 at 01:31:10PM +0100, Paul Eggleton wrote:
> > > > On Wednesday 26 September 2012 13:38:19 Martin Jansa wrote:
> > > > > On Wed, Sep 26, 2012 at 12:34:05PM +0100, Paul Eggleton wrote:
> > > > > > This should be no change to the previous situation unless you
> > > > > > explicitly have pulseaudio in DISTRO_FEATURES_BACKFILL_CONSIDERED
> > > > > > (currently).
> > > > > 
> > > > > Why not use PACKAGECONFIG for this?
> > > > 
> > > > Only because I didn't see the need complicate it just to be able to
> > > > configure it on a per-recipe basis - you almost certainly want
> > > > pulseaudio
> > > > either on or off everywhere. If you want it for stylistic reasons,
> > > > sure,
> > > > I guess... it is a bit ugly though.
> > > 
> > > Well it looked to me like ideal use-case for PACKAGECONFIG.
> > 
> > It does encapsulate the DEPENDS and EXTRA_OECONF into one piece, yes, but
> > then when you make the PACKAGECONFIG option set from DISTRO_FEATURES
> > anyway it gets ugly, and here I'm not convinced it adds anything useful.
> 
> Yes, but that's the case for almost every other PACKAGECONFIG use in
> oe-core too AFAIK and I was asked to convert similar commit to use
> PACKAGECONFIG before so I was expecting that the ultimate goal is to
> replace all such DISTRO_FEATURE tests with PACKAGECONFIG where possible,
> not to add more of them.

Perhaps. Well, I just noticed I need to correct the commit message anyway 
(it's a change to gst-plugins-good, not gstreamer) so given that I'll send a 
v2 using PACKAGECONFIG.

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre