HTMLErrorFormatter isn't cleanly handling unicode characters in .myt and
their compiled .py files. When your .myt raises an exception, and there's
unicode characters nearby the originating line of the exception,
HTMLErrorFormatter will crap out with a UnicodeError when attempting to
print out those source code lines.
les pylonautes français
<br />encoding is: <% m.output_encoding %>
Myghty doesn't know what kind of unicode it's going to encounter in
whatever random .myt/.py raises the original exception, so I'm having it
use the 'replace' errors handler during the translations.
Beware: I'm not sure how else the Request.apply_escapes method is used,
so mind my patch to it (maybe give it an errors keyword option instead).