Align Spec::Be, BeClose failure message to other messages

[jgaskins]

The failure message previously had its to_s method called on it, but other expectations use inspect. This was creating failure messages like this:

       Expected 2022-03-28 22:05:06.966811000 UTC to be GreaterThan 2022-03-28 22:05:06 UTC

[beta-ziliani]

Thanks! I'm adding it to 1.5 just because it might break some existing workflows and I don't want to rush it in 1.4

[asterite]

This was creating failure messages like this

Sorry that I don't have time to try this, but what's the new failure message?

[jgaskins]

Sorry that I don't have time to try this, but what's the new failure message?

New error message is

       Expected 2022-03-29 12:45:07.764260000 UTC to be GreaterThan 2022-03-29 12:45:07.764260000 UTC

We can see from this one that the problem is because the first is not greater than the second because they're equal. The problem with the live one is that 2022-03-28 22:05:06.966811000 UTC is greater than 2022-03-28 22:05:06 UTC, but the error message does not communicate it.

[jgaskins]

Thanks! I'm adding it to 1.5 just because it might break some existing workflows and I don't want to rush it in 1.4

@beta-ziliani This makes sense. I never ran into this until yesterday because most of the be_close expectations I've used have been numeric, where to_s and inspect look the same. And, honestly, if it weren't for the fact that macOS only has microsecond precision on the system clock I doubt I would've noticed it on timestamps, either.

I have it monkeypatched in my own spec_helper, so I can wait on it being added to the stdlib. 👍🏼

[asterite]

Thanks!

As a side note, "GreaterThan" could be changed to "greater than" at some point. I think before moving things to an enum it was "greater_than", which was slightly better though of course not ideal.

[HertzDevil]

It was :">" before the refactor. An exhaustive case or an overridden Relation#to_s will do.