accept tracker connections

[laur89]

Note this change is on top of #172 -- merge that beforehand.
merged

[laur89]

Note no I/O is done here at the moment, we only accept the connection. Any ideas how to proceed, if at all?

[anthonyraymond]

The macro vision is:

• When a peer connect though bittorrent protocol => send a choke message

But "connect through bittorrent protocol" is easier said than done 😆 .

it's something related to : https://github.com/mpetazzoni/ttorrent/blob/2c8bb62f9f69b74c014358d446554df96ec91c7c/ttorrent-client/src/main/java/com/turn/ttorrent/client/network/HandshakeReceiver.java#L148

Instead of adding peers to whatever store you'll just choke them (maybe close connection?).

This feature require a deep comprehension of the peer section of BitTorrent RFC (and also to investigate the source code of libtorrent and other OpenSource clients to understand how they individualy do their job).

[laur89]

Ye this is going to need quite a bit of research/know-how.
Would you be willing to merge this as-is, considering it's reportedly solved problem to one user. And possibly/hopefully it improves existing logic somewhat.

[WZrocks]

Ye this is going to need quite a bit of research/know-how. Would you be willing to merge this as-is, considering it's reportedly solved problem to one user. And possibly/hopefully it improves existing logic somewhat.

Two users. :-) I have been waiting for this merge, and updated Docker image, since the opening of this PR.

[anthonyraymond]

Hello there, i've not forgotten about you.

But i'm lacking time to review that. I try to give it a look ASAP

[WZrocks]

Do you have any idea if and when you'll have the time to look at this, @anthonyraymond? I don't want to rush you, I'm already very grateful for what you did with this project so far, I'm just looking for an ETA.

[anthonyraymond]

Hello,
i'm still unsure if i want to merge this as-is or not. Not sure if that makes its more unsafe or not.

An analogy for this PR is:

• before nobody won't ever answer the call when the phone rings.
• Now someone will, but he is clearly not speaking the language that you expect the people to speak. You'll know it's not the person it is supposed to be on the other end.

[WZrocks]

That is a valid point. Would it be possible to make it optional in config? @laur89 @anthonyraymond

[rursache]

@anthonyraymond i think it's best to just add a toggle in settings. for me it works great but maybe others can have issues, indeed. can't wait to see this merged and published <3

[WZrocks]

I don't think the config option will come anytime soon either, right? 😔

[rursache]

@anthonyraymond can we please merge this as well? been using it since then and it's all good. maybe add a config toggle in the settings just in case?

[anthonyraymond]

@rursache i don't think this is ever going to be added as is. It's sounds a bit dangerous for most people.

I'll include a real and complete implementation in the next version (built from scratch), but for now i prefer not to include it as is

[rursache]

@anthonyraymond
I'll include a real and complete implementation in the next version (built from scratch), but for now i prefer not to include it as is

thank you, looking forward to your implementation!

[rursache]

@anthonyraymond
any update on merging this one? as a flag, setting, hidden, whatever

i've been using it for 2 years already, works fine. most trackers refuse joal with the normal release

[Neurology0443]

Hello :)

After a period of testing I noticed a higher RAM usage that keeps increasing over time when merging this PR (see here #268 ).

Can someone confirm this behavior on their side ?

[rursache]

@Neurology0443 saw it as well but had a similar issue before the PR as well. for years i simply had a hourly cronjob to restart the contaier and that's fine so far. or it's probably easier to limit ram usage via docker compose and let it handle this

[Neurology0443]

Thanks for your answer 👍

I tried with the following docker-compose env. variable:

    environment:
      - "JAVA_TOOL_OPTIONS=-Xms256m -Xmx512m"

From the docker logs it seems to pick-up the variable, however it's ignored:

CPU %     MEM USAGE / LIMIT     MEM %     NET I/O           BLOCK I/O         PIDS
0.31%     780.3MiB / 7.649GiB   9.96%     248MB / 238MB     3.28MB / 19.8MB   59

---

Yeah, a hourly cronjob is more of a workarround than a real solution, but I guess that's how it is... Invidious has a similar issue anyway :/

Thanks !