Attempt to make the output of failed command output expectations more helpful

[Krisell]

Hi all,

I am trying to put together a PR to improve the output of failed expectsOutput or expectsOutputContains when testing artisan commands. However I am not completely satisfied with the PR yet and would like some help, especially with the Mockery-parts.

Please see the draft PR here, along with some questions: Krisell#2

---

For context – here's the planned PR description

Currently, when testing artisan commands and using expectsOutput or expectsOutputContains, the error message on a failed expectation is not as helpful as it could be.

Before:

1) Tests\Feature\TestExpectedOutput::email_sent_test
Output "Email sent to 17 users." was not printed.

After:

1) Tests\Feature\TestExpectedOutput::email_sent_test
Output "Email sent to 17 users." was not printed.

Actual output:
"Email sent to 23 users."

Previously, the developer didn't know if no output was generated at all or almost the correct output was generated. If there is indeed no output, that is also mentioned with this PR:

1) Tests\Feature\TestExpectedOutput::email_sent_test
Output "Email sent to 17 users." was not printed.

No output was printed.

Of course, if there are multiple lines of output, all are displayed.

1) Tests\Feature\TestExpectedOutput::email_sent_test
Output "User with id 3 was updated" was not printed.

Actual output:
"Starting process ..."
"User with id 5 was updated."
"All done!"