Patrick Ohly
Patrick Ohly
> While there is a mention of 'unique behavior' for Error, I don't think it's a good idea to implement it in the sink as 'is always logged'. I'm surprised...
Regarding my original question: I just noticed that there is a "if present" in "the err argument should be used to attach the actual error that triggered this log line,...
We could add a `logger.ErrorMsg("message", ...)`. But I am not sure whether that is really a big improvement over `logger.Error(nil, "message", ...)`.
Let's reopen for discussion. @thockin: what do you think about adding `ErrorMsg` as an alias for `Error(nil`?
Let's give wojas a chance to comment first - see https://github.com/wojas/genericr/issues/2
I agree that `DeferCleanups` solves the problem.
From the documentation: > When a SpecTimeout expires the current node is interrupted (i.e. it's context is cancelled) and Ginkgo proceeds to run the clean up nodes subject to their...
I updated ginkgo to a87ec827fd11 in https://github.com/kubernetes/kubernetes/pull/112923. I ran a similar "broken" test locally with `-focus` set so that only that one spec runs. ``` $ _output/bin/ginkgo -v -junit-report=junit.xml -progress...
See https://prow.k8s.io/view/gs/kubernetes-jenkins/pr-logs/pull/112923/pull-kubernetes-e2e-kind/1579548220873248768 for test results as shown by the Kubernetes CI.
Perhaps we can add a `ginkgo.RegisterFailCallback` similar to `gomega.RegisterFailHandler`? Then `ginkgo.Fail` could invoke such a callback before proceeding with handling the failure. I had considered `ginkgo.RegisterFailHandler`, but such a handler...