Kim Jong-un looking like he’s about to ask you to come visit his island for the weekend and perhaps pen a wee testimonial. Wonder if he also has a jet standing by at Choteau?

What are the chances this opens some eyes / forces development of empathy? (I asked rhetorically)

Yep, known issue: https://feddit.org/post/13613230

Can confirm, but depending on the VPS, your traffic may only be metered in one direction. Mine only meters egress, not ingress, so it’s not too bad if I want to use my media server.

I’m still putting that together, but it’s basically something like:

UPDATE post set url = thumbnail_url where thumbnail_url like 'https://dubvee.org/pictrs/image/%' and url like 'https://lemm.ee/pictrs/image/%'

For the Kbin users to clear out their avatar/banners, I just did:

UPDATE person set avatar=NULL, banner=NULL where instance_id=(select id from instance where domain='kbin.run')

UPDATE community set icon=NULL, banner=NULL where instance_id=(select id from instance where domain='kbin.run')

…and will update that for lemm.ee

I’m working on a Bun/NodeJS script to download the avatars, icons, and banners for all the lemm.ee users/communities, upload them to my pictrs, and then update the DB with the new values referencing the local images.

Prob gonna have to break that up into two steps so I can download and file the images before lemm.ee goes down and upload/assign them after (so they’re not overwritten by federation).

Not sure if I’m gonna do that for comment / post body images. Basically, I just want the archived communities to look nice.

Depending on how the admin has pict-rs (the “subsystem” it uses for media) configured, post images can create local copies /thumbnails for image posts. Some instances opt to do that, some don’t.

Other images like ones posted a comments, user/community avatars/banners aren’t stored anywhere except their home instance.

But yeah, even the limited caching has proven problematic if questionable/illegal content gets posted.

I’m gonna run some scripts on my instance’s database to look for posts that have lemm.ee pict-rs URLs and local thumbnails. Then update the url value to that of the local thumbnail_url. That’ll at least fix them on my instance where I have a copy.

I won’t be doing that until after lemm.ee shuts down, otherwise the original values will federate back in if there’s any activity on the posts.

For comment images or ones where my instance didn’t create a thumbnail, they’ll just have to be broken I guess - same goes for the community icons, though I suppose I could go ahead and get local versions of those and update the communities afterward. Prob also set those to “Mods only” in the database to keep local users from trying to post there.

For user avatars, I’ll probably just clear those in the database so when looking back at old content from .ee, it won’t try to fetch them and timeout/404.

There were less posts to worry about then, but for community icons and user avatars, that’s the same thing I did when kbin shut down.

All of mine use bind mounts so I can just tar-gz the whole deploy folder for backups and migrations. For volumes that connect to remote shares (SMB, NFS, etc) I use named volumes and let Docker take care of their lifecycle.

If named docker volumes would let me specify the local filesystem location, I’d use them. As-is, I rarely do.

(This is faked in dev tools for a joke, BTW. Just making that clear lol).

Yeah, maybe add “[Satire]” to the title because it reads too much like every other AI-bro press release lol.

In order to improve interoperability with Mastodon and other microblogging platforms, Lemmy now automatically includes a hashtag with new posts.

👎🏻 No way to disable this if I don’t want hashtags cluttering up my posts. And worse, it’s Lemmy putting words in my mouth / speaking for me.

[Re: Image Proxying] The setting works by rewriting links in new posts, comments and other places when they are inserted in the database. This means the setting has no effect on posts created before the setting was activated. And after disabling the setting, existing images will continue to be proxied. It should also be considered experimental.

What an absolutely stupid way to do image proxying. Why not just dynamically re-write image URLs to use the proxy path before serving it via the API?

That way:

1. It works with all content posted any time before/after the setting was activated
2. It lets users decide whether they want to proxy or not
3. Doesn’t break images if the home instance pict-rs is broken (which I’ve been seeing a lot of lately)

If you think “that’s not reliable” or “too hard”, I’ve been doing it successfully exactly that way in Tesseract with it’s image proxy/cache for over 8 months (on the front end…in a cave…with a box of scraps).

I’m not sure if there’s a solution here, but I’d like to urge people to avoid lemmy.ml hosted communities in favor of communities on more reasonable instances.

Did that months ago; defederated completely when they turned into Lemmygrad-lite. At first I missed some more active FOSS communities, but since then, others on different instances have become more active. programming.dev has a lot of communities that overlap with some of the bigger FOSS ones on .ml so maybe check out what they’ve got.

If there’s a community that only exists there, be the change you want to see: create it somewhere else, nurture it, and give it time to grow. You’re not the only one making this complaint about .ml, and you probably won’t be the last.

Related: I genuinely feel that ml being the official or at least de-facto flagship instance is turning people away.

Edit: Oh yeah. Didn’t recognize your username at first, but I was looking at the modlog the other day from my LW account, and saw a bunch of individual community bans from Dessalines and wondered what was up. Figured it was something exactly like this, and it was. Thanks for sharing.