I have a shell command I'm running from Emacs that has a Δ character in its output. How can I configure the *Shell Command Output* buffer to display UTF-8 by default?

edit: There could also be something else weird going on. It displays Î8 μs Î8 μs when it should be Δ15 μs

  • Running the command with C-x Return c utf-8 works, so it should be possible to write a wrapper.
    – choroba
    Jul 2, 2020 at 17:14
  • I'm running it via "Shell command on region" with M+|
    – Aido
    Jul 2, 2020 at 17:20
  • Yes, prefixing it with the encoding specifier works for me.
    – choroba
    Jul 2, 2020 at 17:30
  • Oh, do you just say utf-8 <cmd>?
    – Aido
    Jul 2, 2020 at 17:32
  • You say C-x RET c utf-8 <cmd>.
    – NickD
    Jul 3, 2020 at 18:24

1 Answer 1


You can use the per-command setting of the encoding that @choroba mentions in his comment (C-x RET c utf-8) which will only affect the next command, but you can also set UTF-8 to be the default encoding everywhere. Add

(set-language-environment "UTF-8")

to your init file and restart emacs. Or for the current session, click on "Options/Multilingual Environment/Set Language Environment" and select "UTF-8".

Unless you have very specific needs (CJK characters with special encodings, Microsoft Code Page stuff etc), I believe that UTF-8 is safe: I have never regretted setting it globally, but I cannot guarantee that you will not run into problems. IME, the problems tend to be few and far-between and have to do with the special encodings I mentioned (and in those cases, you can use the one-off setting of the encoding to deal with the problem).

But you might want to wait for other opinions before switching over wholesale, and in the meantime deal with the problem one case at a time, using the C-x RET c utf-8 method for each command.

  • I'm worried that this will change how files are saved and mess up other programs. Am I just being paranoid?
    – Aido
    Jul 2, 2020 at 18:33
  • Please incorporate "the encoding that @choroba mentions in his comment" into your answer, instead of just referring to it. Comments can be deleted at any time.
    – Drew
    Jul 2, 2020 at 19:10
  • It’s wise to be cautious, but the chances are quite good that all your other programs already read and write in UTF–8, or that they can be configured to do so. You should examine their help files to see what they do, but unless you have specifically configured them to use some other encoding, I would be greatly surprised if they didn’t already support UTF–8.
    – db48x
    Mar 29, 2021 at 23:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.