XPost: alt.comp.software.firefox   
   From: ldo@nz.invalid   
      
   On Sun, 1 Feb 2026 17:44:30 -0500, Maria Sophia wrote:   
      
   > 1. What a proof-of-origin check is   
   >    A. It is a server-side test that tries to confirm that a request   
   >       really came from a real web browser, not from a script or a   
   >       downloader.   
      
   Also you have officially-sanctioned mobile apps that are allowed to   
   play YouTube videos, that go through the same sort of check.   
      
   Users of youtube-dl/yt-dlp will have seen messages about “Android   
   player” or “IOS player”. I think YouTube may be offering different   
   quality options to different clients, so the downloader offers the   
   option to masquerade as any of them, to try to maximize the choices   
   available to the user.   
      
   >    C. yt-dlp must therefore call an external JS engine to run the   
   >       code and produce the correct proof-of-origin values.   
      
   In previous times, it would use a library like PhantomJS, which   
   actually behaves like a full-function web browser (as far as the   
   server side is concerned), but has no GUI that a human user can see,   
   and is totally controlled from a client program.   
      
   Nowadays this has been phased out in favour of the JavaScript-engine   
   approach. Not sure why: I suspect the PhantomJS-style approach was   
   complex to maintain and keep up to date.   
      
   > That then begs the question of why doesn't everyone see the same errors.   
      
   Or even on different streams of the same video. There was one I tried   
   to download recently where I think the video came down OK, but the   
   audio hit the dreaded 403 Forbidden error (or was it the other way   
   round? I didn’t bother to check before deleting the .part file). This   
   happened consistently on multiple attempts before I gave up.   
      
   That one worked when I tried again about a week later, after an update   
   to yt-dlp.   
      
   > 2. What this means   
   >    A. yt-dlp can appear to work normally even when no JS engine is   
   >       installed.   
   >    B. This does not mean the new JS requirement is gone. It only   
   >       means our requests have not triggered the new challenge   
   >       recently.   
   >    C. The yt-dlp developers have stated that a JS runtime will   
   >       become necessary for reliable long-term YouTube extraction.   
      
   Yeah. And so the whack-a-mole game continues ...   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)