Word classifying may break on strings containing ':'

[Artoria2e5]

# functions.coffee.old:52, undented
[word, tag] = r.split(':')

Should be:

# functions.coffee.new:52, undented
[word..., tag] = r.split(':')