On Tue, Nov 25, 2025 at 09:31:20 +0000, Daniel P. Berrangé wrote: > On Tue, Nov 25, 2025 at 09:31:30AM +0100, Peter Krempa via Devel wrote: > > On Tue, Nov 25, 2025 at 09:21:18 +0100, Jiri Denemark wrote: > > > On Mon, Nov 24, 2025 at 11:14:00 +0100, Peter Krempa wrote: > > > > From: Peter Krempa <[email protected]> > > > > > > > > Add the test data based on 'v10.2.0-rc1-11-g5a5b06d2f6'. > > > > > > > > Note that the data was collected on a Raspberry Pi 5, thus the CPU has > > > > changed to what we had collected before. > > > > > > > > Signed-off-by: Peter Krempa <[email protected]> > > > > > > Is this marked as RFC because you generated it on a different host? > > > > It's a combination of factors ... > > > > > I don't think the CPU itself should be a problem, we're changing that for > > > other archs as well since you don't want to keep an ancient machine > > > prepared somewhere just to keep using the same CPU for regenerating > > > capabilities data. > > > > ... I'm willing to commit and maintain the aarch64 caps as well as long > > as they are on hardware I have permanent access to. So if we're okay > > with rpi5 based capabilities I can do them along with the x86 basd ones. > > If using rpi5 means you can do this regularly, that's a good tradeoff. > I don't think enterprise class hardware matters for the sake of the > unit tests.
I'm willing to add them to my periodic updates along with x86_64 as well as update on request from others in order to stay consistent. In this instance, since it doesn't seem to be needed I'll send the update for 10.2 once it's released, to avoid some churn.
