Andy Piper <andy(a)xemacs.org> writes:
>Neither can I. The code looks like it's able to handle
placing
>glyphs on a detached or string extent without problems.
Isn't the argument string-or-buffer rather than just buffer?
The problem is that set-extent-*-glyph functions end up doing:
set_extent_glyph_1 (Lisp_Object extent_obj, Lisp_Object glyph, int endp,
Lisp_Object layout_obj)
{
EXTENT extent = decode_extent (extent_obj, DE_MUST_HAVE_BUFFER);
^^^^^^^^^^^^^^^^^^^
I don't see why this function requires the extent to be attached to a
buffer. It should be perfectly legal to attach it later.