Weird encoding issue on categories page

Product Support
11

so I downgraded our package to use nokogiri 1.16 again. What I dont get. When ever I bump a gem to reduce duplicated packaging, I check if there had been related changes in main there haven’t been any. Unless I missed something

        "description": "Witaj w polskiej sekcji społeczności openSUSE!",
        "description_text": "Witaj w polskiej sekcji społeczności openSUSE!",
        "description_excerpt": "Witaj w polskiej sekcji społeczności openSUSE!",

as you can see we have the correct text twice, only when it runs through PrettyText.excerpt it is broken. How is that handled in main?

@hendersj I am already preparing a main package so we can test that with a copy of the DB.

I guess it was handled in DEV: Update nokogiri to 1.18.1 (#30554) · discourse/discourse@affe26f · GitHub

but I wonder … in lib/retrieve_title.rb

doc = Nokogiri.HTML5(html, encoding:)

shouldn’t this be:

doc = Nokogiri.HTML5(html, encoding: Encoding::UTF_8)