[OE-core] [PATCHv2] glibc: Separate out AArch64 multilib loader symlink creation

[Mike Crowe]

On Thursday 21 February 2019 at 13:01:27 +0000, Richard Purdie wrote:
> On Wed, 2019-02-20 at 16:47 +0000, Mike Crowe wrote:
> > On Wednesday 23 January 2019 at 16:27:59 +0000, Mike Crowe wrote:
> > > 
> > > Does the lack of response mean that both approaches are horrendous?
> > > If so, do you have any better ideas?
> > 
> > There doesn't seem to be much interest in this one.
> > 
> > The only other idea I've had is to create the symlink at rootfs-creation
> > time. However, that seems to be detaching it too much from glibc, which
> > would cause confusion.
> > 
> > I can apply one of my fixes locally (we currently have some far uglier
> > hacks than these that I'd like to remove), but I'd really rather not have
> > to diverge from upstream this much.
> 
> Sorry for the lack of feedback.
> 
> I'm fairly strongly against a separate aarch64 only recipe for this.
> Realistically we therefore have to fix this in the existing
> class/code/recipe.
> 
> The changes you sent previously didn't look that horrific, did you find
> a working version of that?
> 
> I'm wondering if we can come up with something which works and
> hopefully we can spot some way to then clean it up to be a bit nicer...
> 
> The delay was that I was going to try and come up with something but
> that hasn't happened yet. If you have a horrific non-separate recipe
> version that is complete, we can see if we can do anything to improve
> it.

I had suspected that was the reason for the delay.

I shall knock my modifications to glibc-package.inc into shape and send a
tested patch. I've already improved it a little compared to the version I
sent previously.

Thanks for the response.

Mike.