Re: set-buffer-file-coding-system doesn't

Jan> But -- that leaves my main question open -- is the following Jan> sequence supposed to work in XEmacs and produce an unibyte file, Jan> just as it does in GNU Emacs, or not? Stephen> -- open the file with iso-2022 characters Stephen> -- set the coding system to the one you know can represent all Stephen> characters in the buffer Stephen> -- write the buffer to a file Stephen> Yes, it should, if the coding system is unibyte, of course. Stephen> I'm seeing problems with this recently in 21.4, too, though. Stephen> I thought it was latin-unity related. Ie, if I turn off Stephen> latin-unity it works fine, and latin-unity is the only thing Stephen> that should be different in XEmacs Stephen> 21.4 from a few months ago. Interesting -- I've just tried and even if I run xemacs -q I can reproduce the above problem and I do not get a unibyte result (I'm specifying the ISO-8859-2 encoding, so I should get a unibyte file). So, unless just having latin-unity sitting somewhere on my hard drive makes this happen, it is not the culprit. Stephen> Are you currently normally using latin-unity in XEmacs? Or Stephen> did you just try it again because of this problem? I do not normally use latin-unity, I've just tried it once to solve the problem. Stephen> Are you using Mule-UCS by any chance? If so, in XEmacs, GNU Stephen> Emacs, or both? I do not use Mule-UCS, either, I tried to use it, but I couldn't really even find out what it does (the manual being "terse"...). --J.