freamon @ freamon @endlesstalk.org

aka freamon@lemmy.world, freamon@feddit.nl, and any username from lemmon.we

---

Read more

Posts

40

Comments

382

Joined

2 yr. ago

Just posting to nag you about this: can !tails@lemmon.website be added as an exception to the automod?

Lemmy doesn't seem to get much recognition in the wider Fediverse - it tends to get bundled as part of 'other apps'. Mastodon is much bigger, so better integration with Lemmy probably gets deprioritised below their own issues and feature requests (e.g. I was reading today that Markdown support is often requested, but the base version still doesn't have it)

I don't think it's technically impossible - all the information that another site needs to properly interpret some activity is in the JSON that's sent. I get the sense that it might be unrealistic to expect Mastodon to make the necessary changes though. It seems more of a political issue than a technical one.

It's partly an issue of keys. Every fediverse actor has a private key and a public key. When my instance sends this to fediverse@lemmy.world, it's signed by my private key, and lemmy.world uses my public key to verify it. When fediverse@lemmy.world sends this comment out, it uses it's own private key to sign it. It can't just re-transmit my comment, because it doesn't have my private key. All it can do is Announce that I've made the comment (and sign the Announce).

Mastodon treats Announces as Boosts, so every post/comment is interpreted as a thing that fediverse@lemmy.world has boosted, so you get all these un-connected posts appearing. I think it's mostly up to Mastodon to remedy.

It works better if a Mastodon actor posts into a Lemmy community, then you get the mix like you imagine. e.g.: https://mastodon.world/@Flash/112095241193510662 (this particular post was crowbarred into Lemmy via !tails@lemmon.website, but it would be the same if the author had done it.)

One of the reasons it's inconsistent is that Lemmy tries to balance the media it serves locally vs. the media it lets remote hosts serve. Also it's a bit naïve about image conversion. So if you link to an actual GIF at giphy.com (for example), it works consistently across the most the most platforms if Lemmy leaves it alone. If it doesn't, it'll bring it in and convert it to a WebM file, which not all clients know what to do with. Even if they do though, looping isn't always on by default for video, so the effect of a GIF that relies on looping might be nullified.

It's probably best to focus on the clients with the biggest user-base, rather than try to target them all. For lemmy.world/c/gifs, the most popular posts have been uploaded to imgur.com - which convert most GIFs to MP4s - and the post's author has linked to the inline version at i.imgur.com.

No settings page (as far as I'm aware), but you can use the API to get everything (posts, comments, etc):

step 1: get login token -

    
curl --request POST \
     --url https://lemmy.ml/api/v3/user/login \
     --header 'accept: application/json' \
     --header 'content-type: application/json' \
     --data '
{
  "username_or_email": "2br02b",
  "password": "YOUR-PASSWORD"
}
'

  

step 2: use login token (big long string starting with 'ey') to get data -

    
curl --request GET \
     --url 'https://lemmy.ml/api/v3/user?username=2br02b&page=1' \
     --header 'accept: application/json' \
     --header 'authorization: Bearer YOUR-JWT'

  

Increment page number until you have everything. source: https://lemmy.readme.io/reference/get_user

Voyager has a setting for "No subscribed in All/Local" that does this. It's better on than off, obviously, but it doesn't turn All or Local into some kind of goldmine.

I get the sense that, unless you're willing to do it yourself, feature requests for Lemmy don't have much chance of being realised.

Bit mad, as in a bit strange - e.g. "it's a bit mad to take a single word out of its obvious context for such a desperately cheap shot".

Wait, what? A user posts a thing to a server, and that thing isn't then duplicated to 50 other servers ... yeah, I don't see how that can work.

(I'm just kidding - your site looks neat.)

(sorry to intrude). This has had a few cross-posts on Lemmy - this one technically wasn't, but it seems Lemmy has picked up on it. If you want some ALT text, and to reply to author of this image, the link that goes to !tails@lemmon.website provides that.

Yeah it was pretty nifty in lots of ways. Loads of emulators, of course, and doing stuff like compiling apps directly on the device itself was neat.

Frustrating in lots of other ways too, though. I don't think the Pyra ever really got off the ground, unfortunately.

This was always an impressive emulator - it was originally made for the Open Pandora (an ARM-based mini-computer not much bigger than a DS). It's always been free for that - it's just, you know, you'd have to own an Open Pandora (I do!)

Not quite, no. I know what it isn't at least.

I'll keep going - I'm sure the article's author is someone who genuinely uploaded some confidential info and then became really involved with privacy/GDPR etc, and not someone who was always been really involved with privacy/GDPR issues and now has a story to fit.

I'm gonna find this guy's image ...

https://monero.town/pictrs/image/00000000-0000-0000-0000-000000000000.jpeg ... nope
https://monero.town/pictrs/image/00000000-0000-0000-0000-000000000001.jpeg ... nope
https://monero.town/pictrs/image/00000000-0000-0000-0000-000000000002.jpeg ... nope
https://monero.town/pictrs/image/00000000-0000-0000-0000-000000000003.jpeg ... nope

Mmm, I'm sure it won't take long. Just have to remember to do it all again for .jpg, .webp, and .png.

Anyway, I'll let you know when I get it.

Informing a human that their work has already been done for them by a bot is oddly appropriate for this post.

It's been censored out, but it looks like it's heading for Shit Creek, appropriately enough.

Not that I'm aware of, just communities that seem like they might fit, but don't quite:

!communitypromo@lemmy.ca
!communityspotlight@lemmy.world
!newcommunities@lemmy.world
!newcommunities@discuss.online

I'm sure you could make a more meta-type post in one of these though.

A quick Google suggests you're correct - big sites have implemented their own versions, but it looks like everyone has spent the last decade arguing with other about a CommonMark standard.

I think it's because spoiler tags in lemmy have been custom-made for some reason, whereas all the other stuff is standard markdown. Voyager is a web app, so it can maybe only render whatever the engine it relies on can render.

Edit: Turns out I'm 2 for 2 on making incorrect statements in this comment.

This issue is about the fact that if you delete your account, lemmy will delete all the text in any posts and comments you've made. It won't go through those comments, read any URLs you've uploaded pictures too, and delete them if they've been hosted/cached locally.

Putting the lemmy devs response to one side for a moment: what's the concern here? The URLs for images in picts-rs are a random hex string - if you don't know the URL, you can't find it, and even if you do, you wouldn't be able to connect it to someone unless the info was literally in the image itself.