I wrote:
> I think that changing pg_config's behavior would be enough to resolve
> the complaints you listed, but perhaps I'm missing some fine points.
Meanwhile I've gone ahead and pushed my v1 patch (plus Munro's
recommendation about _fullpath error handling), so we can see if
the buildfarm blows up. The question of whether we can sometimes
skip replacement of symlinks seems like material for a second patch
in any case.
regards, tom lane