With the release of Ruby 2.0 the default encoding for files is UTF-8. Thus you no longer need to prepend the magic encoding string to every .rb source file
However, if you have a project that has to be run under Ruby 1.9.x, adding those magic strings to every file can be really cumbersome. Here’s the trick:
On Linux systems
find . -name "*\.rb" | xargs sed -i "1s/^/#encoding: UTF-8\n\n/"
On OS X systems, the BSD sed version doesn’t behave the same when comes to replace line feeds. Strange, but yes. Read on:
The best solution for me was install GNU sed first:
brew install gnu-sed
And now run:
find . -name "*\.rb" | xargs gsed -i "1s/^/#encoding: UTF-8\n\n/"