test: Re-enable scenario consistency checks after disabling them

[james-garner-canonical]

This PR corrects one of the testing tests to correctly re-enable consistency checks after turning them off. They default to on, and we run our own tests with them on.

Before this PR, they would be turned off, as we check if os.getenv(...), which will be True when the variable is set to the string 'false'.

[tonyandrewmeyer]

Interesting that nothing else had to be fixed. I guess we normally check consistency more directly, or have things consistent.

[dwilding]

Would this be easier to follow in future if test_rubbish_event_raises became two separate tests? One test that checks for AttributeError on the non-pebble-ready events, and another that checks for AttributeError on kazoo_pebble_ready (then consistency checks could be set & unset unconditionally)

[tonyandrewmeyer]

Would this be easier to follow in future if test_rubbish_event_raises became two separate tests? One test that checks for AttributeError on the non-pebble-ready events, and another that checks for AttributeError on kazoo_pebble_ready (then consistency checks could be set & unset unconditionally)

It'd probably be easier to use monkeypatch.setenv then too, which seems like it would be the typical way to do this and ensure the cleanup. But good with me either way.

[james-garner-canonical]

I've refactored to use two tests as you suggested, @dwilding. Would you mind having another quick look before merge?

[dwilding]

Looks really clear to me, thanks!