[Buildroot] [PATCH 1/1] package/meson: Fix linking of shared/static libs with static libs

[Adam Duskett <aduskett@gmail.com>]

Hello;

I asked for an update on this pull request on github here:
https://github.com/mesonbuild/meson/pull/3939#issuecomment-492355588

And it seems that the general consensus will be to fix the link_whole behavior,
but strip out or rework a lot of this patch. Mainly because many other projects
already have workarounds, such as gstreamer, mesa, and libglib.

On Tue, May 21, 2019 at 12:54 PM Peter Seiderer <ps.report@gmx.net> wrote:
>
> Hello Thomas,
>
> On Mon, 20 May 2019 20:45:17 +0200, Thomas Petazzoni <thomas.petazzoni@bootlin.com> wrote:
>
> > Hello Peter,
> >
> > Thanks for your feedback!
> >
> > On Mon, 20 May 2019 20:26:15 +0200
> > Peter Seiderer <ps.report@gmx.net> wrote:
> >
> > > I can confirm the patch fixes the gstreamer static build/static libs
> > > problem (see [1])....
> > >
> > > But as it is a (not yet) upstream committed bugfix, maybe it is better to
> > > wait for a upstream committ and/or next meson version bump?
> >
> > That would of course be the best approach...
> >
> > > On the other side the gstreamer folks where bored with waiting for the
> > > meson fix to be accepted and committed an workaround for the problem
> > > already (see [2])...
> > >
> > > Conclusion: Take the patch in case there is already a buildroot
> > > in-tree package which needs the fix (or commit the patch with
> > > in series with a in-tree user), otherwise wait for meson
> > > upstream...
> >
> > ... however, if it starts holding off bumping too many packages, I
> > guess we'll have to take this patch. Having to add a !BR2_STATIC_LIBS
> > dependency to libglib2 would be really horrible, considering the number
> > of reverse dependencies that libglib2 has.
>
> As far as I know only gstreamer itself is affected (the only package
> linking a static library using an internal static helper library)?
> But the first other package affected by this bug justifies the
> patch applied ;-)
>
> >
> > At least, if the problem is acknowledged upstream, we know it will be
> > fixed at some point. Perhaps not with the exact patch we will have, but
> > at least it will be fixed, which means we won't have to carry this
> > patch forever.
>
> Yes...
>
> Regards,
> Peter
>
> >
> > Thomas
>