Interesting UI bug while I was viewing the Death by user count post. The side bar changed to the Beehaw science community, which I had open in another tab. and it says that einval posted it to the science community, not the general community here.

I didn’t know einval was the mod of [email protected]? /s

I suspect this relates to something in federating with other communities as I subscribed to that community through the einval server.

  • einval@lemmy.einval.netM
    link
    fedilink
    English
    arrow-up
    2
    ·
    3 years ago

    Oh yeah, I’ve seen this too but only when I have multiple Lemmy instances open. When I posted “Death by user count” it told me I posted it to lemmy.einval.net through lemmy.ml’s instance. I had to shift+ctrl+r to get it to show the correct pathway. Its definitely weird that session data is leaking around between logins. You’d think each cookie/session/socket would be totally unique.

    I saw they’ve been working on ditching websockets in favor of a pure REST API. Hopefully this will stop when they release v18.0.

    • lastrogue@lemmy.einval.netOP
      link
      fedilink
      English
      arrow-up
      1
      ·
      3 years ago

      Oh I didn’t realize they were relying on websockets!

      When an update releases, does this instance automatically update? Or do you have to initiate it via ansible, docket pull, or something else?

      • einval@lemmy.einval.netM
        link
        fedilink
        English
        arrow-up
        1
        ·
        3 years ago

        It’s a manual process. In theory an upgrade should be as easy as updating the image tags for Lemmy and Lemmy UI in the docker-compose.yml file, then taking everything down and praying it starts back up again.

        I’m not sure how they’re going to go about telling everyone they need to upgrade. Or I guess I’ll wake up to an unfederated/dead instance with a massive error log at some point. 😂