Patrick Ohly
Patrick Ohly
/assign @nojnhuh
> Is the issue the testing.T type vs. the ktesting.TContext type, or is it not using the most local one, whichever type that happens to be? The latter. The problem...
> The problem is that tCtx doesn't shadow t, so the outer testing.T can be (and was, as seen in this PR) used by accident The opposite happened in https://github.com/kubernetes/kubernetes/pull/133964#pullrequestreview-3538955084
I agree. Closing in favor of https://github.com/kubernetes/kubernetes/pull/135608 which uses the helper function approach and fixes the same mistakes.
/sig node /sig scheduling /wg device-management
Description updated.
Sorry for the confusion. Yes, this is in 1.33, only 5027 isn't. I've reopened the doc placeholder PR.
For taints applied by the admin we probably should literally do what the admin says: if the `DeviceTaintRule` applies to all devices of a driver regardless of the pool, it...
This will remain in alpha for 1.34. No KEP update is planned for this cycle, but there might be some minor code PRs.