when test scripts fail to load, no tests may be run and no error is given

[Robert Collins]

Consider this:

I have a test_foo.py.

I run some tests.

I edit the file and run 'testr run someregex'

But I edited test_foo and now it doesn't load. So if I were to list-tests, there would be no tests matching my regex. But testr knows there *were* valid tests.

It would be great if testr could detect this and turn it into an error somehow.