When testing #19 I found that the tests were mistakenly ignoring some
whitespace changes. I've enabled whitespace checks and had to update the
tests because my editor likes newlines at the end of files and removes
whitespace at the end of lines. To make the test files easier to manage,
I've modified some scenarios to still test for the feature but also work
with the whitespace modifications that editors would perform.
This is based off of pull request #10 from @athieriot.
Thanks for the idea! I've implemented it as a flag and an environment
variable that can be enabled. I've also added another test to handle
the environment variable.
This lets me run tests in a terminal. If I mess up and the terminal is
being read by `cat` or similar, this input is used instead and now my
tests will automatically fail properly.
