Ilpo> Loading a file with '-*- coding: utf-8 -*-' in the first Ilpo> line works OK, characters like ö and ä look OK. Yup. This will be extended to the second line in 21.4.8 so that coding cookies don't intefere with shebangs. I could be convinced to go farther than that for similar formats where you can't reasonably have a comment in the first or second line. Ilpo> Loading a file with Ilpo> Local variables: Ilpo> mode: text Ilpo> coding: utf-8 Ilpo> End: Yup. This has never worked and isn't going to work in 21.4. It's not a matter of tweaking the local variables code; you have to read the buffer once raw, recognize the coding system, then re-read the file with the correct coding system. Very low-level code. Ilpo> As a workaround: is there a way to convert the buffer to Ilpo> look right? (other than putting the -*- coding: utf-8 -*- Ilpo> in the first line, closing it and opening it again) M-: (decode-coding-system (point-min) (point-max) 'utf-8) RET You could also try (set-coding-priority-list '(utf-8)) in your init file, or (set-coding-priority-list '(iso-8-1 utf-8)), to get autorecognition. Ilpo> A third thing: xml-mode should notice the coding system from Ilpo> the encoding-attribute. Not in 21.4 for the reasons given above. Maybe in a package. -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN My nostalgia for Icon makes me forget about any of the bad things. I don't have much nostalgia for Perl, so its faults I remember. Scott Gilbert c.l.py