• 1 Post
  • 45 Comments
Joined 2 years ago
cake
Cake day: April 24th, 2023

help-circle






  • Ope sorry, right now I just have the serve config doing a redirect of port 22, however when I try to SSH in I get rejected by tailscale ACL. Says there’s no user named git.

    If I followed the steps for the vanilla docker setup I’d add a git user to the host and softlink the host authorized_keys file to the gogs container’s version, as well as add a shim script to forward the command into the container using the docker exec command, but I’d rather not do that by mucking about in the sidecar if there’s a better way. The tailscale universal docker mod for linuxserver.io says they have ssh access for their containers but as far as I can tell it just pops in the --ssh flag in tailscale up.











  • I recently fell into this rabbit hole myself! Though I decided against hosting the blog myself(because I don’t want to do anything stupid lol) Nowadays for sites like that you can install a static site generator to automatically build the site based off of markdown files. I personally use Hugo but I hear good things about Jekyll too.

    The way mine is set up I make a post or a page on my machine and push it to my backend github repo. Github detects a change and rebuilds the site with the new content using a github action, then uploads the whole public folder to my host at neocities.




  • I’m just curious as to how you think these are insurmountable problems while every instance in existence today is already managing to navigate these issues.

    The only thing the author is suggesting is to pool the resources that are going to waste copying media posts around the fediverse into a new backend (that means it’s not directly user accessible and presumably subject to the same restrictions as posts right now) so that the cost of media hosting is more distributed between all the fediverse instances instead of having the big ones hogging all the bandwidth of the small ones with memes because some users decided to subscribe to a community on say Lemmy.world.


  • Did you even read the article? It’s not like all the users just get unrestricted access to storage to treat like a google drive, this is a backend thing. This guy is trying to find a solution to all the wasted bandwidth and storage space from sending copies to all the other instances they’re federated with, which is a legimate issue that the instance admins are already dealing with on a daily basis. This will let them pool resources to help lower costs for smaller instances.

    As to the CSAM thing I can only imagine it would be easier for one instance to purge fifty images/restore from a backup and everyone else just have to redo their thumbnails as opposed to all the instances having to purge and restore but that’s just me.