[OE-core] [PATCH] eglibc-dbg: provides libc6-dbg

[Phil Blundell]

On Mon, 2014-01-27 at 11:20 +0800, Kang Kai wrote:
> On 2014年01月26日 03:30, Phil Blundell wrote:
> > On Wed, 2014-01-22 at 17:06 +0800, Kang Kai wrote:
> >> Use glib-dbg does work. But if I use it in a recipe, when uclibc is
> >> used, I have to edit the recipe.
> > Isn't that still true of "libc6-dbg"?  Worse, it isn't even guaranteed
> > that the eglibc-debug package will be named libc6-dbg on all targets; it
> > would probably end up as libc6.1-dbg on alpha-linux or ia64-linux for
> > example.  Encouraging people to hardcode "libc6" doesn't seem very
> > wholesome.
> 
> I think this is the way to deal the "RPROVIDES" and please ref to
> libc6-dev in eglibc-package.inc:
> 
> RPROVIDES_${PN}-dev = "glibc-dev libc6-dev virtual-libc-dev"

I'm not quite sure I understand what you're saying here.  But, a few
observations:

- the RPROVIDES that you mentioned is somewhat bogus since, as
previously mentioned, the autonamer is not guaranteed to come up with
"libc6" for eglibc on all platforms

- uclibc doesn't include eglibc-package.inc anyway so it won't be
affected by any RPROVIDES in that file

- adding that RPROVIDES for uclibc would be especially bogus since there
are no existing platforms where uclibc does get autonamed as "libc6" 

So, I still don't really understand how adding this extra RPROVIDES does
much to help with uclibc compatibility.

p.