Re: lessen user alarm when coding system is questioned

[Eli Zaretskii]

> From: address@hidden (Werner LEMBERG)
> Date: Sat, 22 Dec 2001 09:27:58 +0100 (CET)
> 
> > ISTR that at one point we set it up to display that character along
> > with the message.  Did that feature go away?  In any case, it would
> > be a good feature to have.  Does anyone know HOW we would find that
> > character?
> 
> I can imagine it is a descendant of the routine which checks the
> buffer for valid encodings supported by the current coding system.
> This function has to find the offending location anyway, so the
> information is already there.

I think your assumption is wrong.  The function which prints that
message doesn't try to encode the text.  Instead, it computes the list
of possible ways to encode the text (by looking at certain tables
which specify, for each character, how can it be encoded), then try to
match that list with the default coding systems defined by the user's
environment and preferred encodings.  If these two don't match, Emacs
pops up the question.

So the information about the offending characters isn't there to
begin with.