Hello, Achim Gratz <strom...@nexgo.de> writes:
> I looked at the code/documentation and it seems the intention is to do > the reverse lookup, that is look at the buffer encoding and decide what > to give inputenc as argument (when it is specified as "AUTO" in the > header template). I maintain that the other way around (looking at the > header template and deciding the coding system) is probably easier to > understand and it is also what's already been used by Emacs: see > latexenc-find-file-coding-system and latex-inputenc-coding-alist. Emacs also provides `latexenc-coding-system-to-inputenc', so it doesn't look like one way is more legitimate than the other. Note that we can have both ways: with AUTO, let coding-system decide and without AUTO, let inputenc decide. Also, it isn't necessary to change the global coding system to change the coding system used to write the file: see `org-export-coding-system'. Anyway, if you think it's useful to provide a patch, I suggest to dynamically bind `org-export-coding-system' in `org-e-latex-export-to-latex' before `org-export-to-file' is called. You'll have to open "outfile" in order to apply `latexenc-find-file-coding-system'. Do you want to provide such patch? Regards, -- Nicolas Goaziou