Michael Bell

It is very similar to the original “let’s wait for a 403 approach” already demonstrated to be incorrect

Is it feasible to ask for n + 1 objects when issuing the query? Then you don’t need a count per se

I am fairly certain this is simply correct behavior. The SERVER SIDE of this will not support paging for CalendarView. You have to "fake it", which is a little tricky...

That is jnfact exactly what I meant by the tricky comment I made :;

This is in fact a problem and I assume @brettwooldridge did it for the same reasons as Jetty - the modularity stuff doesn't play nice with 1.7.x

Here's the TLDR - works fine against local k8s - kind or minikube - Fails against all our QA cluster - Works fine inside an ubuntu container in the cluster...

Confirmed that recompiling the linkerd-cli with ``` cd cli; GODEBUG=netdns=cgo GOOS=darwin GOARCH=amd64 go build ``` works. However this is a pita and not very convenient. Perhaps publishing another variant of...

Yes there are many references to this issue and yes it affected kubectl for a while too. It also affects "lens", a popular ui, and at least a dozen tools...

https://github.com/golang/go/issues/12524 Abd I figured it out by recompiling . This issue hits oh I’d guess thirty percent of go tools I’ve used in the mac but maybe I’m blessed :;

A colleague recommends https://github.com/greenboxal/dns-heaven as a workaround but frankly that feels weird