Konrad authored Nov 20, 2018 and Robert committed Dec 20, 2018

OpenGraph clients could not properly parse our content because
`content.get_absolute_url` didn’t return anything and the canonical url
of our content therefore always referred to `http_origin`. This
basically broke everything as only the canonical URL is used to create
the OpenGraph dataset.

Images were broken as well, as the way the image was stored was changed
in the backend without updating this template. We now have dedicated
Facebook and Twitter meta image fields.