Random vhost for testing about nameko HOT 5 CLOSED

Hey @mattbennett. Why can't you use something like rabbitmqctl delete_vhost test at the end of a test run?

from nameko.

It's definitely preferable to use an API rather than wrapping a command line tool. Also I strongly suspect that rabbitmqctl either directly calls the API itself, or they both invoke the same underlying code.

from nameko.

by "something like" i meant whatever underneath this command line call. what i'm curious about is why we can't create a test vhost and then kill it at the end of the run - seems much tidier than creating random ones.

from nameko.

rabbitmqctl seems to be a wrapper around the erlang interpreter (https://github.com/rabbitmq/rabbitmq-server/blob/master/scripts/rabbitmqctl) so there's no benefit in going to what's "underneath" rabbitmqctl (same problems of the binary perhaps changing between platforms, not being on the path etc.)

As I said I suspect the fact that connections aren't immediately closed is just how things are implemented inside RabbitMQ, and doesn't have anything to do with the management API per se.

I agree that it would be nicer to use and then clean the same vhost between tests, but it doesn't seem to be possible, at least reliably.

from nameko.

Fixed in #288

from nameko.