Re: [PATCH] gstreamer1.0-plugins-bad: add missing wayland-native dependency

[Richard Purdie <richard.purdie@linuxfoundation.org>]

On Wed, 2016-07-27 at 18:40 -0700, Khem Raj wrote:
> 
> > On Jul 27, 2016, at 12:51 PM, Burton, Ross <ross.burton@intel.com>
> > wrote:
> > Why is this needed when wayland already depends on wayland
> > > -native?
> > Because builds from sstate won't pull in all the build dependencies
> > transitively, just enough to do the required tasks. The
> > requirements are the wayland runtime libraries (wayland) and the
> > native tooling (waylaid-native).
> Is it so that populate_sysroot tasks wont do the above two jobs ? if
> yes then why
> If you are saying that behavior is different with or without sstate
> reuse if  I do
> bitbake gstreamer1.0-plugins-bad, then I think there is a problem. 
>  The entries changed here will
> get added to DEPENDS in the end. So Martin has a point.

The way the sstate code is coded, it assumes that if a target recipe X
is available from sstate, any -native/-cross dependencies of that
target are not needed and not installed.

In general this makes sense as quilt-native isn't needed if we aren't
patching, we don't need a compiler/toolchain if we're not compiling and
so on.

What it does mean is that any recipe needing a -native recipe to build 
should list it in DEPENDS directly, not rely on other dependencies to
pull it in for them. This applies to pkgconfig-native, intltool-native
and also to wayland-native.

I'm open to other ideas on how this could work but right now, its the
only way I can see to make sstate use semi-optimal, rather tahn
requiring *everything* come from sstate all the time.

Cheers,

Richard