[OE-core] [PATCH] pkg-config: allow kernel to be build with esdk

[Saul Wold]

On Fri, 2017-06-23 at 15:47 +0100, Richard Purdie wrote:
> On Wed, 2017-06-14 at 07:03 -0700, Saul Wold wrote:
> > 
> > When the kernel's menuconfig target is called while using the esdk
> > or
> > an esdk-based container, the pkg-config info that is found is not
> > correct.
> > The pkg-config info is for the target, but we need the eSDK's
> > information in order to build the host based menuconfig.
> > 
> > The new pkg-config script checks both that it's in SDK and being
> > called from the check-lxdialog script in order to limit the scope
> > of
> > when the pkg-config automagically switches to pkg-config-native.
> > 
> > This script also replaces the default pkg-config-native script.
> > 
> > [YOCTO #11155]
> > 
> > Signed-off-by: Saul Wold <sgw at linux.intel.com>
> > ---
> 
> To be honest, I'm scared of adding this kind of complexity to the
> system. I appreciate the problem you're trying to solve but I worry
> adding this kind of change is complex, invasive and introduces too
> much
> risk.
> 
> I have some specific worries with this implementation too, in
> particular that it changes on target pkg-config, it really should
> only
> change nativesdk-pkgconfig? If we have risk, we want to minimise that
> risk and adding this wrapper everywhere doesn't seem to do that...
> 
There is no nativesdk-pkgconfig, but I had some confusion around the
pkg-config-native and it's usage of a pkg-config binary.  I was getting
some recursion initially and my feeble brain thought it was the target
version :-(.

I have rebuilt this patch and will send a V2 that ONLY affects the pkg-
config-native, basically expands the script to check if it's being
called as -native or from the Kernel's check-lxdialog.

I am also going to change the script such that it checks it's base name
of pkg-config and only then add the parent pid name checking. 

I hope that will mitigate the risk.

Sau!


> Cheers,
> 
> Richard
> 
>