I'm not sure how there can be "one true sane thing". There's more to strings than Unicode, and more to Unicode than utf8. In fact the os does pretty sane here -

I'm not sure how there can be "one true sane thing". There's more to strings than Unicode, and more to Unicode than utf8. In fact the os does pretty sane here - not checking for utf16 or utf32 isn't really great on Linux either (haven't checked how/if nim handles locales on Linux). Afaik, eg python3 does the right thing on Windows and Linux w/print(). It's just one of the many little things were one may whish for "one, simple, sane way" - but have to deal with the realities. Same goes for things like handling paths, line endings for text files, resource forks (or not) for files, changing file meta-data...