OK, you’ve found a new one 
…
… one of the problematic URLs (over here) was that capturing webarchives from discourse forums uses e.g. https://discourse.devontechnologies.com/latest instead of the actual thread’s (or post’s) URL.
This Script: Create webarchive from selection with correct URL fixes this “error” (be aware, it uses UI scripting).
Although I’m still sure it’s really an error in case of discourse forums, I meanwhile think that not including what comes after a # in a URL maybe does make sense (e.g. headings may change - I guess a URL that includes a # will not be redirected in this case, so one would end up with nothing.)
But given that one knows that a URL that’s not accessible may be not accessible due to a # heading that changed, I still think it’s far more convient to be able to directly jump to a record’s online source.
(Wrote the script when I was very disappointed to find out that I couldn’t open a record’s URL to it’s place in a very very long Markdown document. But again, there’s probably a reason why they decided to do it this way - however that doesn’t mean we shouldn’t customize it)