Re: drag-and-drop failures

[Dave Love]

"Jan D." <address@hidden> writes:

> I still think it might be an endian problem.  If mozilla for
> instance says that the text is in unicode (UTF-16), Emacs will
> currently assume UTF-16 as little endian.  If this is not so, the
> text will be mangeled.

OK, I see what you mean.  That won't depend on the Emacs locale, for
instance, though.  I'm rather surprised utf-16 would be involved,
rather than utf-8, but it does fit the symptoms.

> However, if it differs between machines,
> that is a stupid design, because there is no way to tell which
> endian the UTF-16 data has.

That should actually be defined, at least if it's an IANA charset, but
I forget what the rules are.  (handa and I went round a couple of
circles getting the utf-16 coding systems right...)

> If you look in lisp/x-dnd.el you can
> find utf-16le there.

OK, I'll see if I can figure out what's happening when I get the
chance, if I can do it at the Lisp level.  Perhaps someone else has
access to other big-endian systems running the same applications and
could test?