Output hidden when CLI test fails

By default, the CLI test infrastructure does not echo the output of CLI tests to stdout. This means that if a CLI test fails, it can be difficult to figure out what the issue is. One option is to add debug = true to the cliRun function and re-run the test, but that isn't obvious if your not familiar with the cliRun command.

Instead, we should just always echo the CLI output and rely on test tools or IDEs to capture the output and only show it when a test fails or more information. Note that we conditionally dump the CLI test streams if there is an error, since in some cases the stream is already lost by the time an error is found (e.g. unexpected exit code).