Re: Replace Youtube-DL with HyperVideo

[rain1]

On 2019-08-07 10:02, Hartmut Goebel wrote:
> Am 05.08.19 um 20:56 schrieb Raghav Gururajan:
> > I came across this project called HyperVideo 
> > (https://libregit.org/heck
> > yel/hypervideo). It is a fork of Youtube-DL, without non-free JS and
> > SWF bytecode.
>
> Huh? youtube-dl does not include any JS or SWF bytecode - so there is
> nothing to remove.
>
> And I just checked the diff, and there are not blobs removed (which is
> what I would have expected). The only things changed are:
>
>   * Signature decryption for youtube signatures is removed.
>       o The JS bytecode interpreter is removed - which AFAIU is not a
>         complete interpreter, but only tried to fetch the calculated 
> URL
>         from the code.
>       o The SWF bytecode interpreter is removed - which also AFAIU is
>         not a complete interpreter.
>   * The IqiyiSDKInterpreter removed in youtube_dl/extractor/iqiyi.py -
>     theses seem to use some hard-coded authentication tokens for login
>     at http://kylin.iqiyi.com/
>
> I can't see why this would require to replace youttube-dl by 
> hypervideo.

Hello

According to my analysis youtube-dl is libre software.

there was some confusion at some point because it was thought that it 
has a "javascript interpreter" and "executes non-free javascript 
downloaded from the youtube website". I put these in quotes because they 
are not really true. The "interpreter" is actually just a regex that 
strips out a certain snippet and an simple arithmetic expression 
evaluator, it doesn't handle much of javascript and it's very much below 
turing complete. It uses this to compute a key that's needed to access 
certain videos.

Here is my post about it on gnu-linux-libre 
https://lists.nongnu.org/archive/html/gnu-linux-libre/2017-07/msg00000.html

I think the fork will divide our efforts and reduce our ability to 
cooperate with each other as a free software community. I don't 
recommend it.