Failing test cases often start out with: ``` incorrect output, diff: --- expected +++ actual ``` which makes it difficult to figure out where the expected output might live. That information can be derived from examining the tests, of course, but it'd be much nicer if it were provided in the diff to see straightaway. We do this by introducing a new check, one which takes a filename of expected output, which then enables us to display the failing file, e.g.: ``` incorrect output, diff: --- /home/froydnj/src/binaryen.git/test/passes/code-folding.txt +++ actual ``` which is arguably nicer. Having this new check also enables reducing some boilerplate `open(...).read()` calls in various places. There are still a few places using `fail_if_not_identical`, usually because `.strip()` is used on the expected output.