Thanks for the bug report. Would you please discuss the practical advantages of the proposed patch? Is the patch is for Japanese users who do not specify a locale via environment variables like LC_ALL? If so, how do they specify their preference to Emacs now, and how would this change affect behavior for them? And how do users specify their coding preference to other programs, such as "cat" and "diff"? I have the impression that Emacs users in Japan are gradually switching from EUC-JP to UTF-8. Is there some way to verify this? I know of website surveys like , which says that since 2010 UTF-8 has grown from 51% to 90% of websites whereas EUC-JP has shrunk from 0.7% to 0.3% worldwide. However, this issue is more about what encodings are used by keyboards and terminal displays in Japan (typically these are programs in windowing systems these days, as GNU/Linux consoles cannot handle enough characters to support Japanese). Is there some way to estimate this? A minor point: a change like this should be noted in the NEWS file, and the commit message should use the usual Emacs style. I'm attaching a copy of your patch with a revised commit message, along with another patch to mention this change in NEWS and to change affected documentation. I will ask for comment about this proposed change on the emacs-devel mailing list.