defense.gouv.fr
In some pages the html tags are not closed, python 2.6 cannot handle it.
They are listed in the "supportedsites.html" page.
Allows a continue with a warning when an extractor cannot retrieve a description.
Only appears sometimes, nearly identical to length 86.
Fix orf.at extractor by adding file coding mark
Make find_xpath_attr also accept numbers in the value
Fix MIT extractor for Python 2.6
Like in "<meta charset='utf-8'/>"
The HTML for the MIT page does not parse cleanly for Python 2.6 due to script tags within an actual script element. The offending piece is inside a comment block, so removing all such comment blocks fixes the parsing.
Conflicts: youtube_dl/utils.py
Allow changes to run under Python 3
Causes DeprecationWarning: http://www.python.org/dev/peps/pep-0352/