Ar an naoiú lá de mí Feabhra, scríobh David Kastrup: It’s a coding system name that that’s mentioned several times in the XEmacs 21.4 source tree. But wherever it’s mentioned, it’s commented out. It’s really surprising that you say that XEmacs is automatically setting this--can you give us details of the file in question and any pertinent ~/.xemacs/init.el settings? For example, I can manually do: M-: (setq buffer-file-coding-system 'iso-latin-1) RET and the assignment will succeed, and the coding system indicator in my mode line will disappear, because this coding system doesn’t exist; but whenever I try to save or even M-x describe-coding-system RET RET, it gives `No such coding system: iso-latin-1', so it’s pretty evident this setting is wrong. Or are you inheriting your settings from GNU Emacs? In that case, I have some coding system warnings that I have a mind to add to XEmacs; the intention is that they should be run after evaluating the initialisation file, and they should give warnings about inconsistent user settings. A check along the lines of (unless (coding-system-p (find-coding-system (default-value buffer-file-coding-system))) (display-warning 'error (concat "The default value of buffer-file-coding-system _must_ be " "an existing coding system. It's not, at the moment, and " "you _will_ lose data as a result!"))) after ~/.xemacs/init.el might better deal with this case. -- “Ah come on now Ted, a Volkswagen with a mind of its own, driving all over the place and going mad, if that’s not scary I don’t know what is.”

Ar an naoiú lá de mí Feabhra, scríobh David Kastrup: Thoughts, people? Should we put in the alias? I honestly think “yes” if it makes David’s life easier. -- “Ah come on now Ted, a Volkswagen with a mind of its own, driving all over the place and going mad, if that’s not scary I don’t know what is.”

Aidan Kehoe <kehoea(a)parhasard.net&gt; writes: I've been bitten by the same problem again, this time with mule-utf-8-unix. Really, I should think you have to decide for one of four courses: a) add aliases for everything for which you have a different name everywhere b) add alias handling local to the desktop package that translates when loading and saving to the GNU Emacs equivalents. Or that c) simply removes the buffer coding system from what gets saved and restored. d) tell people that they might just as well forget sharing a desktop file between Emacs and XEmacs. In case c), you need to provide a mechanism by which the desktop package will use a file name different from the file name Emacs uses. But it does not make sense to wait for a new bug report every time and then add a single alias. Version d) would be a permanent and probably the cleanest as well as easiest solution with the obvious (and in my opinion tolerable) drawback that you can't share desktops when switching editors. Version b) would be considerable work, version c) is not optimal, but a file that does not get identified correctly when reloading without the desktop package could be expected to do the same with it. Version a) would turn your own choice of encoding names into a mockery since then it would become quite hard to enforce _your_ names as long as all the aliases stay around. It would also fail to bring across the message to third party package authors that they better use the XEmacs encoding names or else. Anyway, I am not the only person that is going to run into trouble with that, I believe, and so you really should not wait for the next bug report before agreeing on a strategy. -- David Kastrup, Kriemhildstr. 15, 44793 Bochum

David Kastrup <dak(a)gnu.org&gt;, on Wed, 09 Feb 2005 20:42:12 +0100 you wrote: [snip] The coding system ought to be named iso-8859-1-unix. Like Aidan, I wonder where the setting of iso-latin-1-unix came from. Where did you get the RedHat build? Rawhide? -- Jerry James http://www.ittc.ku.edu/~james/