Yes, "Open in FIrefox" works fine, but that's not the problem being discussed here, which is for those of us who have Firefox configured to "Always Ask" (in my case for PDFs).

Thanks for the tip about "Copy Download Link". That indeed gives the URL as finally resolved, not the one with all the search metadata that the search engine gives out.

However, I'm afraid this still doesn't solve the problem I described when trying to copy the internal URLs within a download (at least not without copy-pasting the ends of the internal URLs onto the copied download link, which is still an unwarranted regression). But thanks for trying.

[In case anyone is wondering why some of us need to use the "Always Ask" setting: The Firefox PDF viewer is OK for some PDFs, but when you're looking at a PDF containing complex vector graphics (e.g. graphs with thousands of plotted points, which is my bread and butter), as you scroll back and forth or zoom in and out, you spend all your time watching the "loading..." wheel for pages you were just looking at a few seconds ago. Whereas my local PDF viewer handles all this like a breeze. Also printing PDFs from the Firefox viewer often produces gobbledygook fonts.