This community got removed from lemmy.world (again) - Divisions by zero
lemmy.dbzer0.com
This is another post that alerted me of this. https://lemmy.world/post/13287681 [https://lemmy.world/post/13287681] And here is the modlog: https://lemmy.world/modlog?page=1&actionType=ModRemoveCommunity [https://lemmy.world/modlog?page=1&actionType=ModRemoveCommunity]

This has happened once before and they reversed it. But they said this last time too:

The discussions that have happened in various threads on Lemmy make it very clear that removing the communites before we announced our intent to remove them is not the level of transparency the community expects, and that as stewards of this community we need to be extremely transparent before we do this again in the future as well as make sure that we get feedback around what the planned changes are, because lemmy.world is yours as much as it is ours.

https://lemmy.world/post/3234363

  • Skull giver@popplesburger.hilciferous.nlEnglish
    0·
    8 months ago

    Lemmy makes local copies of everything when federation occurs. It’s 100% on their server. The only exceptions are images posted as part of the comments, those are loaded directly. Then again, that adds the ability to add tracking pixels, so that’s not exactly great for a piracy community either.

    Image loading example

    I turned off all the logging for this proof of concept but this could’ve been a transparent PNG pixel that tracks every bit of information your browser will give it.

    • nutsack@lemmy.worldEnglish
      0·
      8 months ago

      i really wish there were a way to disable images with some of these fancy lemmy clients for android. I’m not interested in any of them

      • nutsack@lemmy.worldEnglish
        0·
        8 months ago

        you could safeguard against this on the client side by not loading images from untrusted sources. irc clients did this

      • Skull giver@popplesburger.hilciferous.nlEnglish
        0·
        8 months ago

        I’m not sure, but anything doing Markdown parsing and allowing images to be embedded is vulnerable to this. I kind of doubt that the devs don’t know about this.

        The alternative would be to download every image on the server and cache it until users start requesting the image files, rewriting the Markdown to link to the new image location. I can think of a few reasons why that’s not implemented.

        Proxying all comments was implemented in the backend at some point, I’m not sure why this feature was removed again. I can’t find much in the repo history, you could ask the devs why the feature got removed if you’re curious.

      • Skull giver@popplesburger.hilciferous.nlEnglish
        0·
        8 months ago

        Your client asks my server for the image, my server does a basic IP location lookup based on a free internet database I downloaded last year and turns it into an image on the fly.