Adds a NopLogger to the backend/log package.

[joeblubaugh]

This is useful in testing and occasionally in other scenarios for plugin authors.

What this PR does / why we need it:
Adds a NopLogger to the backend/log package. This is useful in testing and occasionally in other scenarios for plugin authors.

Which issue(s) this PR fixes:

Fixes #708

[CLAassistant]

All committers have signed the CLA.

[joeblubaugh]

This is a speculative PR. If #708 isn't desired, we can just close.

[andresmgot]

This is useful in testing and occasionally in other scenarios for plugin authors.

Can you elaborate the use case when it's useful? For example, in tests, it will only print logs if the test fail, which can be helpful to identify the root cause of the error.

[joeblubaugh]

Fair point about the tests; I hadn't considered that.

In my case, I implemented this NopLogger locally because I have some code that does complex data transformations over several hundred lines and many functions. In one context, failures are errors and need to be logged. I want to use the logger to ensure that the issue is logged where it occurs and the logs can be monitored. In another case, the errors don't need to be logged, because the transformation is used in a validation step. These errors should not be logged, but errors are returned to the user in a "bad request" form.

[andresmgot]

I want to use the logger to ensure that the issue is logged where it occurs and the logs can be monitored

From this, it seems that you may want to change the logger implementation to assert things on execution? In any case that's not something you can do with this NopLogger.

Since we are using the Hashicorp logger under the hoods, I think that if we add the Off level that they support, it should cover the use case that you want, suppressing any logging for a certain test. Is that correct?

[joeblubaugh]

Yes, I think adding support for Off would cover my use case.

I don't want to assert in the logger - just pass a different implementation to functions in some cases.

[wbrowne]

Closing this due to inactivity. However, please feel free to continue the conversation here if this is still relevant, and we can re-open if necessary. Thank you!

[joeblubaugh]

Sorry, I think I took this message as a handoff rather than a request for me to do it:

Since we are using the Hashicorp logger under the hoods, I think that if we add the Off level that they support, it should cover the use case that you want, suppressing any logging for a certain test. Is that correct?

[joeblubaugh]

This was addressed in a very similar way in #876, so no need to reopen.