If you experience broken CSS (or the top nav bar being broken, which is actually a JS issues because of course we need JavaScript to re-implement basic HTML features), I recommend you to check the headers of your .html files. It could be that wget retrieved some content from the cloudfront.net CDN in gzip format and forgot to extract them.
run grep -r s.gz * in your directory base, that's the most likely culprit, if you get any hits, that's the most likely culprit. Simply run gunzip -r * in the cloudfront.net directory then and something like find ./ -type f -exec sed -i -e 's/s.gz/s/g' {} \; in the archive root to fix repair every instance.
I think the --compression= command flag can help with this stuff but seriously somehow each option does something different than what you'd expect so running a fix-up script might unironically be the most pragmatic solution for now.
As for the forum: …no, I don't think I fetched any XML, but upon further inspection, I think that's for multi-page forum threads? I frankly never bothered to check beyond loading up the first page of a forum thread so I never noticed the rest being borked. I think the fix is probably also JavaScript related to load those afterwards but as of now I'm mostly guessing.
I took a look at your github link and frankly, my visceral recoil from seeing some dockerized node.js-bullshit already makes me incredibly biased against whatever the flying fuck those industry-bootlickers have cooked up now.
I suppose, since the WD API is a strong case of WikiDon't, this might be somewhat useful if you actually want the retrieve the per-page revision history and changelogs and whatnot; somewhat superfluous for our size.
Also, like, wget is a preinstalled utility on pretty much every linux distro ever and has an unblemished 30-year history of doing exactly what it's supposed to do, with extensive documentation and interoperability and customization options.
That options is so fundamentally broken they expect you to run it containerized from the get-go (where it will slurp up unreasonable amounts of resources) and I furthermore based on Node.js and adsfdadfdgeafh I don't even know where to start there. Was there even one week this year that disgrace upon human achievement did not have some critical eco-system-wide CVE and/or package-based worm? No kinkshaming, but I'm really not into the whole pain and humiliation that brings with it.
I haven't grabbed the sandbox yet, but might, frankly.
Oh yeah, also since I'm currently in the middle of moving (for the second time this year), that means I technically have off-site backups as of now. Yeah, we're playing in the big leagues.
