Failure of encoding detection on Python2 causes hard to debug termination

in the case that someone moves the default coding line down a few lines,
on python2 ascii is detected instead of utf-8, therefore coverage fails flat when decoding

i was only able to find the problem by adding a comment in coverages html reporting at the end of the html_file method right before the html = html.decode(encoding) line

The encoding declaration must be on the first or second line of the file. This is according to PEP 263: http://legacy.python.org/dev/peps/pep-0263/

Can you provide a reproducible test case that demonstrates the problem? I'll reopen the ticket then.

Thanks.

The author contacted me on IRC to say, "the problem was mostly that the error left no way to figure what file it broke on."

This was fixed in 529fef3d32ab. Encoding errors now don't break the HTML report at all.