Video do not update nor show up in channel page

[MatteoAlberghini]

Official Instance

• The bug is reproducible on the official hosted instance, or is API-related.

Describe the bug

There is an issue where no matter what channel you go to, no videos are showing up

Also, if you have scripts running to auto-pull videos they won't work

See libretube issue: libre-tube/LibreTube#8443

Potential fix from PipePipe (debug branch): InfinityLoop1308/PipePipe#2309 (comment)

To Reproduce

1. Go to any channel
2. No videos are showing

Expected behavior

Load all videos from youtube API

Logs/Errors

No response

Browser, and OS with Version.

Fedora 44 KDE, Firefox

Additional context

No response

[syboxez]

NewPipe is also working on a fix.

TeamNewPipe/NewPipeExtractor#1491

[t6fb3m59]

I made a fork that uses the latest version of NewPipeExtractor. It fixes this issue alongside probably others, including the live videos in recommended bug. It's not properly tested so let me know if anything breaks. There are a few performance changes to NewPipeExtractor that I did not copy on my fork, but it's probably fine for self hosted instances.
image: logicalkarma/piped:latest

I also forked the frontend to address some Firefox specific issues for video loading times (optional).
image: logicalkarma/piped-frontend:latest

[Alifoss]

I made a fork that uses the latest version of NewPipeExtractor. It fixes this issue alongside probably others, including the live videos in recommended bug. It's not properly tested so let me know if anything breaks. There are a few performance changes to NewPipeExtractor that I did not copy on my fork, but it's probably fine for self hosted instances. image: logicalkarma/piped:latest

Thank u, i can use piped again. Won't u consider making a pr to piped-backend?

[t6fb3m59]

I would, but I am not so certain it would help. I ended up making my own fork of NewPipeExtractor so I can use some changes made in Piped's fork of NewPipeExtractor and the upstream version. That is the fix. The NewPipeExtractor from Piped isn't really maintained much and last time it broke it took a while for fixes to be made.

In short, this is at least a 2 repo issue, the fork is not very serious, at least for now, and I am not confident in the maintainer answering if I start making pull requests. So for now it's neither a serious fork, nor for submitting PRs to upstream. In the future I may pick one of these paths because the gradual breakages have been driving me a bit mad.

[MatteoAlberghini]

I appreciate both images, the firefox fixes are very useful as well :) thank you

[nyarau]

I made a fork that uses the latest version of NewPipeExtractor. It fixes this issue alongside probably others, including the live videos in recommended bug. It's not properly tested so let me know if anything breaks. There are a few performance changes to NewPipeExtractor that I did not copy on my fork, but it's probably fine for self hosted instances. image: logicalkarma/piped:latest

Works like a charm!

[10203]

I would, but I am not so certain it would help. I ended up making my own fork of NewPipeExtractor so I can use some changes made in Piped's fork of NewPipeExtractor and the upstream version. That is the fix. The NewPipeExtractor from Piped isn't really maintained much and last time it broke it took a while for fixes to be made.

In short, this is at least a 2 repo issue, the fork is not very serious, at least for now, and I am not confident in the maintainer answering if I start making pull requests. So for now it's neither a serious fork, nor for submitting PRs to upstream. In the future I may pick one of these paths because the gradual breakages have been driving me a bit mad.

I need to include a small disclaimer, every piece of code written in their fork is written by Claude. I'm just putting the information out there, nothing else.

[t6fb3m59]

That is somewhat true. Keep that in mind, however none of the changes written by claude on the latest tag are very significant as far as I can remember, but rather solve simple issues with little surface for bugs. The biggest one is the NPE update, which is human written, but I had claude work the git repo and sign off on it. Most of the code you see on the latest image is likely from a human on another repo cherry picked by claude because I asked it to.

On that note, I have been playing with SABR by stitching what I could find as a temporary solution at the very least. It's currently on the experimental branch and is essentially vibecoded (since I am running out of free time to work on this for now) and requires the experimental tag for proxy, frontend and backend which is now available in the docker repo. I only tested if it could play a video on desktop chromium. I will likely improve it bit by bit and eventually commit to actually working on it seriously (hopefully around July). I do not recommend running experimental just yet. That said, it will probably be fine if you expect breakages and are ok with changing the images back if disappointed. I mostly stand by the broad strokes of how the SABR implementation is done.

If anyone finds issues with these forks and would like to submit them, feel free to visit the issues on the frontend fork. They are very much not finished, but they are something.

[raycheung]

I ended up making my own fork...

While it's really great of you, @t6fb3m59, to provide a working fork for the breaking issue, I highly recommend to submit a quality fix MR to the original project repository(ies) before considering forking. This is to keep the open-source community active and consolidated, also as an encouragement to the project owner to support. I believe if you raise the relevant MR, they would be merged here.

[t6fb3m59]

I have added my current take on this matter in the fork's readme. Basically, I will figure out later what to do with these changes, for now they stay as is with best-effort gradual improvements. One key note I did not mention is that if anyone wants to try filing merge requests of those changes here, they are more than welcome, just drop my name somewhere in there and make sure you properly flag LLM generated code if there is any.

I finally got SABR to a somewhat usable state and there is no way that I could push that here in the near future. The code quality on the experimental branch is lower and even for lively repos shipping SABR has been in discussion for 1-2 years with little to no code written for them despite Youtube systematically tearing down the DASH protocol frontends currently use for decent quality playback. I am not saying they are doing something wrong, it's just that getting a lot of people to agree on an architecture always takes more time and SABR forces drastic changes. For those curious, only Freetube currently has a functional SABR implementation among them as far as I know.

I will stop discussing my fork here since I am pretty sure this is beyond the scope of the issue. The fix for this is nothing special, the Piped NPE fork needs to be updated from upstream. If you want to contribute, report issues or start discussions, feel free to visit the Piped-LK fork. If somebody still wants to convince me to do MRs here, please move the discussion there.