Some feeds with broken encodings fail to parse

Ever since 1a484ec3f58ff5e7a5745f231b68cb64ba65929d, which removed handling for feeds with broken encodings, a surprising amount of my feeds are constantly broken, because they claim UTF-8, but then contain ISO-8859-xx characters in some headlines or image descriptions. LibXML thus fails to parse them.

Is there any way to get that functionality back? For now I have reverted the commit in question, to get my feeds working again.

write a plugin and/or ask publisher to fix their broken content.

I doubt I will manage to find anyone who cares there anytime soon.