[Aboriginal] Regression in tarball naming?

Rob Landley rob at landley.net
Mon Oct 17 06:16:44 PDT 2011

On 10/16/2011 04:05 PM, David Seikel wrote:
> I just ended up with a "lfs-bootstrap.tar.gz-i486" file after running
> the lfs-bootstrap to test out the latest versions of things.  I thought
> we sorted that issue out before?

Not a regression, I believe it's always done that.  I blogged about it
last month, but didn't actually address it:


As I said, "This is a design conflict between the control image being
target agnostic and the host system needing to distinguish between files
uploaded into the same directory."

The thing is, if I rename the tarball then I should really rename the
directory the tarball extracts to so they match, which means either the
build has to use the $HOST name in setup-chroot or I need to rename the
directory before tarring it up.

Not a huge deal, but just icky enough I haven't done it yet, since the
problem it fixes is also minor.  The upload function sends it with a
host-specific name, nothing else in the native build cares about that at
the moment.



More information about the Aboriginal mailing list