that is odd. i wonder if it's about the specific heredoc that we're using there:

>-

i know there are several variants and i wonder if the docs are using the wrong one