Hi, I think that unicode_start should not mess with the user’s console font. It should only put the console into Unicode mode and be done with it. The problem with the current approach is that the default Unicode fonts considered by the script are often not what the user wants to see. Every font should be fine to display enough characters to be readable, anyway, and users who need a larger subset of Unicode characters can run setfont manually. It would also make sense because unicode_stop never changes font. The obvious counter-argument is backwards compatibility, but then, most users will not call unicode_start themselves but leave it to their distribution’s init scripts. What do you think? -- Michael Schutte