I have a hunch it’s some Nostr client using forcing compression and that is incompatible with Nostream.

Reply to this note

Please Login to reply.

Discussion

What debug level would you suggest to track it down? There is an ENV variable, so DEBUG=DEBUG? Or true false, what does it expect?

The error is clear there, there’s no need to track down. There’s a client forcefully sending compressed websocket frames and Nostream relays don’t like it. Nostream negotiates what websocket extensions and features are available, and compression is not. But there’s a Nostr client that just sends compressed websocket frames anyways.

Okay, I see. But maybe it is possible to find out which npub is causing it or which client it is?

In any case nostream should not die in such a case. But I suppose this happens inside the ws library, so difficult to fix for your, right?

#[3]​ Can you have a look at this nostream issue, please? It seems a request sent by Damus is causing nostream to crash:

https://github.com/Cameri/nostream/issues/256

How do you know this is Damus? Are you using a relay no one else is using to test this?

I see my own IP in the nginx logs and had no other client running at that time!