Dominik Charousset
Dominik Charousset
> I presume the broker websocket interface makes addressing this a bit mood I'm not so sure about that. If we replace subscribers with WebSocket clients, it'll run into the...
> Hmmm... The above change didn't actually stop TLS 1.0 or 1.1 connections. That change only affects the zeek-to-zeek (binary) connections. I'll look into the WS part. @timwoj, @ckreibich: do...
@MMChrisHinshaw your patch actually should affect both binary and WebSocket connections. This is the output for me locally: ``` $ bash test.bash Testing tls1 Result 1 Testing tls1_1 Result 1...
I've merged the PR, so TLS 1.2 should be enforced as minimum going forward. However, I could not verify this. The CentOS 8 docker image seems to be dead/EOL, so...
> Could this be lifted into an API so that de-serialization can be re-used outside of json_client? Sure, I'll take care of it.
Nice, thanks for doing all that digging. 👍
> I can somewhat reproduce this on a bigger machine with the original reproducer Thanks, Arne. That's good to know! I'll ping you when I have a patch ready to...
@awelzel can you try `topic/neverlord/gh-352`? The branch contains the Broker-side fix for #366 plus a CAF-side patch to address frequent `max_element` calls. I've written a small benchmark that pushes 1k...
Thank you for taking care of playing this back to the reporter. 🙂 Let me try to reproduce this bug with a stress test. I'll ping you if I have...
Same as #1796: > Closing the PR, because we are currently finalizing 1.0, i.e., won't add new features at this point. After 1.0 is out, we will revisit this.