2007/7/24, Kenichi Handa : > > Is this patch correct? Most particularly, is it correct to use > > buffer-file-coding-system for a URL? I have doubts about that. > > I doubts too. I'm not the expert of URL (or URI) encoding, > but, as far as I remember, non-ASCII characters in URL must > be at first encoded by UTF-8 then by %-encoding. So, for > instance, �+" (U+00E0) must be encoded to "%C3%80". Oh, you are right. buffer-file-coding-system should be replaced by 'utf-8. It happens to be the same value here. -- William