Op 16 mrt. 2012, om 17:32 heeft Mike Crowe het volgende geschreven:
> > The magic in rootfs_ipk.bbclass expects to generate debug package
> > names by just appending '-dbg'.

It turns out that the above was a local change so that statement was
incorrect. Nevertheless the change probably makes sense anyway.

> >                                 This breaks for eglibc because whilst
> > the binary package has been renamed to libc6 the debug package is
> > still called libc-dbg. This patch ensures that the debug package is
> > renamed too.
> > ---
> > meta/classes/libc-common.bbclass |    1 +
> > 1 files changed, 1 insertions(+), 0 deletions(-)
> > 
> > diff --git a/meta/classes/libc-common.bbclass 
> > b/meta/classes/libc-common.bbclass
> > index 962f205..d0c1ab4 100644
> > --- a/meta/classes/libc-common.bbclass
> > +++ b/meta/classes/libc-common.bbclass
> > @@ -27,4 +27,5 @@ python populate_packages_prepend () {
> >             bpn = d.getVar('BPN', True)
> >             d.setVar('PKG_'+bpn, 'libc6')
> >             d.setVar('PKG_'+bpn+'-dev', 'libc6-dev')
> > +           d.setVar('PKG_'+bpn+'-dbg', 'libc6-dbg')
> > }
 
On Fri, Mar 16, 2012 at 08:39:00PM +0100, Koen Kooi wrote:
> Missing PR bump

On eglibc?

>                 and broken upgrade path.

I'm having trouble coming up with a way to solve that. I think the
libc6-dbg package needs to RPROVIDE, RREPLACE and RCONFLICT with
libc-dbg but if I add the following:

  d.setVar('RPROVIDES_' + bpn + '-dbg', bpn + '-dbg')
  d.setVar('RCONFLICTS_' + bpn + '-dbg', bpn + '-dbg')
  d.setVar('RREPLACES_' + bpn + '-dbg', bpn + '-dbg')

Then the package names undergo translation and I get:

  Provides: libc6-dbg
  Replaces: libc6-dbg
  Conflicts: libc6-dbg

Which is clearly wrong. :(

How would you recommend I provide backward compatibility for this
change (assuming it is still seen as desirable?)

Thanks for your comments.

Mike.

_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core

Reply via email to