Marius Bakke skriver: > I have a hunch that this has to do with the grafting code affecting > origins (with gexps?), but have not confirmed this. I have now confirmed this with a small shell script: