[vlc-devel] Use case for the youtube-dl filter?

[Rémi Denis-Courmont]

Le torstaina 24. syyskuuta 2020, 10.38.31 EEST Pierre Ynard via vlc-devel a 
écrit :
> > Antoine had the foresight to add Lua parsers 13 years ago with the
> > other Pierre's help, before YoutubeDL was even a thing. But I think
> > it's about time to admit that that plan failed. We simply don't have
> > the people to contribute and maintain those parsers, nor the automated
> > test cases.
> 
> I wouldn't presume of what the plan was, and what the aim and ambition
> were, to say that it failed. It's especially not a failure if other
> people build upon the concept we put forward to create a next generation
> that's even better.

It's a success in terms of allowing playlists to be written in a simple and 
memory-safe language.

But it's a failure in terms of how many actual parsers have actually been 
contributed and maintained. Or at least it is in how many are actually 
delivered with VLC and not left in a dark corner of the VideoLAN extensions 
website. It surely did not help that the feature was only available in 0.9 
development tree for a year or so. In the mean time, YoutubeDL came to be.

And then what you call release management was a complete oversight.

> As for VLC still, personally I have had the resources to contribute
> and maintain scripts for several major sites that I use: youtube,
> dailymotion, vimeo, soundcloud, vocaroo. Those work and are not a
> failure. Then there's a release management issue for in-tree scripts;
> and indeed that's a factor towards an external youtube-dl.

The Lua playlist framework is a partial failure, not the specific Lua scripts. 
But TBH, even the scripts have a severe problem with updates delivery to final 
users - in spite of your maintenance efforts.

> I think the plan was also for an external lua script ecosystem. I don't
> know that this part of the plan failed.

Compare the number of supported services on VLC side and in mpv (via 
YoutubeDL).

-- 
Rémi Denis-Courmont
http://www.remlab.net/