It's a limitation of PDFKit. We don't have a solution yet.
I have reopened this ticket considering we'll be looking for a new source capturing tool.
So not expecting that we will fix this, but maybe we can diagnose the problem a bit more, and make sure whatever new tool we choose has a solution?
This source is another example of a blank page... https://wikirate.org/Source_000105498
But when you download it, it does show up properly.