On 06/08/2012 12:51 PM, Steve McIntyre wrote:
I've seen the log from  and I'm disappointed, to say the
made an agreement about this several weeks ago  in a conference
call. Various of the people on the call claimed to be Red Hat / Fedora
ARM and toolchain folks who were both interested in working out a
reasonable answer to the linker path debate and empowered to implement
whatever was agreed. We had a productive discussion and quickly came
to an amicable agreement.
I know you're disappointed, but as I mentioned to you at Connect, this
is about the bigger picture. It was important that we clean up the
linker path discrepency before someone did an "LTS" style release, so
that we don't have many years of incompatibility. It was, therefore,
vital that Canonical ship with the right path, and it would have been
ideal if we could too. As I mentioned, our tools guys have been looking
at this but I totally agree with what Peter and others have said, the
unfortunate reality of needing a hack to effectively avoid a mass
rebuild means this may well take longer time than we have for 17.
It's not a failure if we ship 17 and then fix this linker path in an
update, then have it resolved in 18. We still get to a point within 6
months where the "LTS" Canonical 5 year release is compatible with us
and SuSE, and whoever else. If we do a multi-year "LTS" style release it
will serve us well to have figured this out now. There is no loss in the
longer term, just perhaps some pragmatism needed this week.