write_charset_8bit() claims to return the character set that should be
used when writing 8bit characters (usage bears this thinking out; this
is only used when we want to indicate a charset for 8-bit characters).
However, in practice what that means is it will return whatever
get_charset() returns, unless it returns NULL ... in that case, it will
return "x-unknown".
I don't think that get_charset() can return NULL, because I don't
think that nl_langinfo() can. (As an aside, nl_langinfo() can
return an empty string, but I don't think nl_langinfo(CODESET)
will, at least in practice.)
This seems wrong to me. I guess the question I'm asking is: if the
locale specifies US-ASCII is the character set but we detect 8-bit
characters, what should be putting as the character set? US-ASCII is
definitely wrong (that's what we do now). x-unknown seems slightly less
wrong, but I'm not in love with it. "Aborting with an error" is a
possible answer, but I'm not in love with that idea either.
This thread might help:
http://lists.gnu.org/archive/html/nmh-workers/2006-08/msg00000.html
David
_______________________________________________
Nmh-workers mailing list
Nmh-workers(_at_)nongnu(_dot_)org
https://lists.nongnu.org/mailman/listinfo/nmh-workers