> The man page says that \(bu has unicode U+2022, but nroff puts > out U+00b7. Is this a "bug" (... or did I once again got the > documentation wrong ;-)?
bash$ groff -Tutf8 | xxd \(bu 0000000: c2b7 0a0a 0a0a 0a0a 0a0a 0a0a 0a0a 0a0a ................ Indeed. I see this as a bug. I find in /usr/share/groff/1.22.2/tmac/tty.tmac the following lines: .ie c\[pc] \ . tr \[bu]\[pc] .el \ . if c\[md] \ . tr \[bu]\[md] If I remove them I get (correctly) bash$ groff -Tutf8 | xxd \(bu 0000000: e280 a20a 0a0a 0a0a 0a0a 0a0a 0a0a 0a0a ................ I don't know why these lines are there. Remapping the bullet character if it doesn't exist might be legitimate, but in this case it gets remapped on the condition that *another* character exists, not that bullet *doesn't* exist. That's wrong.
