Just as I thought I have finished the patch, I found
out I was wrong about the following piece:
Werner LEMBERG:
> > Therefore, I suppose that groff applies the ex-
> > isting character translations inversely to get
> > back to some simple characters. Then, hyphen-
> > ation codes can be computed and compared against
> > the hyphenation patterns.
>
> Yes.
Looks like I was wrong here. .hcode allows to assign
hyphenation codes directly to character entities.
Futhermore, it seems to be working on glyph name
level, because its odd parameter can be just any
string. Only if no hyphenation character is assigned
to a glyph, will groff try to convert it back to the
original input character and look for the hyphen-
ation code of the latter.
Therefore, the input character mapping file (like
koi8-r.tmac) is not necessary for UTF-8 input. Just
set hyphenation codes directly:
.hcode x
Anton
