The ofx files have the correct charset inside their headers and
trying to convert the file to StringIO using a particular encoding
is subject to failures when the file isn't exactly an ascii file.
For example a file encoded with CP1252 with accents on letters like
é or è will not be parser as utf-8 and will fail to load.
Also since OfxParse is supposed to receive a file handle, it is
correctly reading the file header and choosing the appropriate charset
to read the rest of the file. For this reason, pass the bytes as a
ByteIO that doesn't care about the encoding.
Loïc Faure-Lacroix
1059bf6b3e