Global settings to solve invalid multibyte char (US-ASCII) in Rails 3

Question

My problem is simillar to Rails 3 invalid multibyte char (US-ASCII)

Solution, presented there does works, but i find this a little goofy to include comment # encoding: UTF-8 to every file, that uses non ASCII characters.

In config/application.rb, i have following option:

# Configure the default encoding used in templates for Ruby 1.9.
config.encoding = "utf-8"

But, it does't helps.

Maybe, some another global configuration option is presented exactly for such cases ?

Answer 1

There is no way to default Ruby 1.9 to use UTF-8 for all source files (and yes, it's a huge PITA).

The config.encoding setting is telling Rails something, nothing to do with Ruby.

The best solution I've come up with relies on something that I haven't seen mentioned. There's another way to get Ruby 1.9 to see a file as UTF-8, instad of using that magic comment you can include the UTF-8 BOM (Byte Order Marker) in your file. Ruby 1.9 will do the right thing.

Now to the cool part - I use gvim, and it's actually very simple to configure gvim to put a BOM into every file that should have one. It's as simple as adding this to my .vimrc:

set bomb

Not sure if you're using gvim, or if other editors can do this. So obviously YMMV :)

Answer 2

You can set the environment variable RUBYOPT=-Ku when running the application, and it will make ruby (even 1.9) assume utf-8 encoding for all .rb source files.

For a rails app though, I would rather recommend not using that trick, and instead put your non-ascii parts in the i18n files and referencing them through I18n.t.

It's a bit more tedious to set up if you've already got your application up and running, but should you ever need to translate it into another language you're already halfway there.