Re: blank-mode.el

[Vinicius Jose Latorre]

Robert J. Chassell wrote:
>    >     (setq default-input-method 'ucs)
>
>    Humm, the command above didn't work in my system,
>    that is, it continues to display hollow box at end of line.
>
> That is a problem.  It means that different machines have different
> possibilities.  I fear that only ASCII is common to them all.  In that
> case, the end of line should be a dollar sign (octal 44, decimal 36,
> hexadecimal 24).  I don't like it.  A dollar sign is heavy, ?\u21b5 is
> good -- but it is not within the ASCII range and of those glyphs that
> are, the dollar sign is the best for this purpose.
>
> Perhaps you could provide an option for ?\u21b5.  Certainly, the
> documentation should mention it.
>   

I found the problem.

The problem only happens for characters code
above #x0FFFF (?\u21B5 == #x53BB5).

blank-mode uses the same mechanism of visws.el
(from Miles Bader <address@hidden>) to map
character code representation using
buffer-display-table variable.

This mechanism also embeds the face id
into the character code, it seems that this is the only
way to have face property associated with these
characters without interference of font-lock.

Humm, ok, one way to solve this it is not embed face id
for characters code above #x0FFFF, but these characters
will not have any face property, that is,
if font-lock is enabled, it'll use the face that font-lock
designates to that region of text.