Ruby 1.9 - Invalid multibyte character (utf-8)

I have a ruby file with only these two lines:

# encoding: utf-8
puts "—"

When I run it with ruby test_enc.rb it fails with:

test_enc.rb:2: invalid multibyte char (UTF-8)
test_enc.rb:2: unterminated string meets end of file

I don't know how to properly specify the character code of — (emdash), but vim tells me it is 151, Hex 97, Octal 227. It fails the same way with other characters like ã as well, so I doubt it is related specifically to that character. I am running on Windows XP and the version of ruby I'm using is:

ruby 1.9.1p430 (2010-08-16 revision 28998) [i386-mingw32]

I feel like there is something very obvious I am missing here. Any ideas?

EDIT: Learned a valuable lesson about assumptions today - specifically assuming your editor IS using UTF-8 without actually checking it. Oops!

Thanks for the quick and accurate replies all!

EDIT AGAIN: The 'setting up vim properly for utf-8' grew too big and wasn't really relevant to this ques开发者_如何学JAVAtion, so it is now a separate question.

Given that Ruby is explicitly calling your attention to UTF-8, I strongly suspect that you haven't actually written out a UTF-8 file to start with. Make sure that Vim (or whatever text editor you're using to create the file) is really set to write out UTF-8.

Note that in UTF-8, any non-ASCII character will be represented by multiple bytes, not a single byte as you've described from the Vim diagnostics. I'd recommend using a binary file editor (or dump, or whatever) to really show what's in the text file though. Something that doesn't already have some preconceived notion of the encoding - something that isn't even trying to think of it as a text file.

Notepad lets you write out a file in UTF-8, so you might want to try that just to see what happens. (I don't have Ruby installed myself, otherwise I'd try it for you.)

Your file is in latin1. Ruby is right.

emdash would be encoded on two bytes not one in UTF-8.