In 8eb6e2ed7b8187262b58a08370926a59369346ef, charset handling was
refactored to make use of the ianaindex package. A number of charsets
present in the ianaindex package were removed from being explicitly
mentioned in go-message, leaving only a "quirks table." While nearly all
of the removed charsets are handled by the ianaindex package, support
for three charsets which were previously handled by go-message was
dropped: cp1250, cp1251, and cp1252.
Was this intentional? Guessing not since the commit didn't mention it. I
double checked all of the other encodings which were removed in that
commit, these are the only three that are not handled by ianaindex.