I submitted the following in an issue in the upstream repo:

It works in a lot of cases, but I've just noticed while working with quoted-printable text that it doesn't do that very well. For example, the string "=E2=80=94" should decode to "—", but this extension shows "â��". Also, "=E2=80=99" should decode to "’", but it gave "â��" again. This seems to be a problem with Unicode support.

It doesn't handle the "=" linebreaks correctly, either. The decoded text includes a lot of equal signs in them.

The quoted_printable_decode function is on line 689 of js/D3.js.

The linebreak problem can probably be solved by stripping whitespace from the end of the strings, or altering the regex to ignore whitespace between "=" and the end of the line.

The Unicode problem is probably just a matter of handling the decoded characters correctly.