The content detection mechanism could use some better heuristics.
This particular issue seems to be that mf2py hecked up; given that this page has no mf2 markup on it, there should probably be a more aggressive fallback to the heuristic mechanism based on the determined content of the mf2.
The current content detection mechanism completely hecks up on this article:
https://www.webdistortion.com/2019/05/16/can-we-all-please-stop-using-medium-now/
Instead of the article text I just got a list of commenters:
The content detection mechanism could use some better heuristics.
This particular issue seems to be that
mf2py
hecked up; given that this page has no mf2 markup on it, there should probably be a more aggressive fallback to the heuristic mechanism based on the determined content of the mf2.