On 03/22/2012 04:43 AM, Peter Robinson wrote:
Did someone take meeting notes for those of us that couldn't
Yes, then my system overheated. I recently found where the .swp file
was hiding, so here you are:
Big take-away: The proposal needs work. Please read and modify. If you
don't feel comfortable modifying the text, put something in the Talk tab
and ask others to review. If you have reservations, disagree, or see a
shortcoming, please edit directly or put in the talk section. This is
how we'll get a unified opinion into the document.
The proposal needs to:
Establish why staying the course as a secondary architecture isn't good
1. The things Peter mentioned in fedora-devel about spending all our
time trying to stay caught up.
2. Technology leadership- ARM is the future, Fedora is where
development takes place, it's natural that ARM should be represented as
Provide ARM 101, including niggling issues that we already have internal
consensus on, but have not communicated: LE only. Alignment issue
stance. What armv5tel, v7hl, v8 are, and that v8 is not part of the
proposal. Other foundation stuff.
Provide data based on our F17 experience.
1. How many packages are built in F17? How many of the missing
packages are FTBS in PA? This goes to the ExcldudeArch aspect of the
2. How long do builds take on ARM? How fast will they be on the
server grade hardware? Roughly stated we expect up to a 2.5x speedup
from server hardware vs *trimslice* builds, so we can extrapolate this
Do more to discuss how developers may be affected. Slower builds,
fedora-devel-list tells us, wil result in QE and Security delays.
Chainbuilds will be impacted particularly.
Hardware access is needed for packagers who do not have ARM systems,
1. Redo documentation for QEMU+Virtual Express solution so anybody
with a PA x86 system can use.
2. Contemplate who/how to provide ARM test systems to people with FAS
FESCo specifically asked the ARM team to contact rel-eng,
infrastructure, testing, and kernel groups to solicit feedback.
test: pwhalen (?)
We will revisit the proposal with FESCo next Monday to get incremental
feedback and direction. Please try to have all edits in before this time.
Brendan Conoboy / Red Hat, Inc. / blc(a)redhat.com