Is this being worked on/looked at? I ask since the post mentioned discussion was happening with the .id admins but I haven’t seen any answers about whether that’s a priority (it seems Lemdro.id is the one that needs to federate with k.bin).
Just kinda bummed since I use kbin the vast majority of this time.
Edit: ah cool, quick turnarounds are sexy. Thanks @cole@lemdro.id!
Edit: A fix has been deployed, and federation should work now. Unfortunately, kbin.social is taking its sweet time federating with us, it is probably overloaded. Lemdro.id is already accessible from other kbin instances such as fedia.io (https://fedia.io/m/android@lemdro.id)
Hello! Admin here at lemdro.id. This is the result of several problems in the lemmy default reference nginx config. I am working on resolving this right now and should have it fixed within the next 30 minutes!
Yo, I’m a Kbin.social member and can report we’re federating now
Great news! Likewise, kbin.social doesn’t seem to be updating content from lemmy.world communities either (I’m seeing some a week late).
Not sure if this is related, but I tried replying to a comment from a user on kbin, and my instance wouldn’t send the reply. They still have federated posts and top-level comments though, even mine. Maybe their queue is just overloaded.
Once you’ve finished updating it, can you share the changes you made to your Nginx config? If the default references have issues with them, I’d like to make sure mine doesn’t have the same problems.
sure! take a look at this issue: https://github.com/LemmyNet/lemmy-ansible/issues/106
It is likely yours is fixed if you deployed recently
You can test an instance like this.
Find a local user on the instance and get their page (https://instance.tld/u/username).
In linux:
curl -X GET -H ‘Accept: application/activity+json,application/json’ https://userlink
If the result is html they have not patched, if it is json, they have patched.
curl -X GET -H ‘Accept: application/activity+json’ https://userlink
Will return json all the time on all instances.
I patched my kbin instance to send just the latter until it’s resolved everywhere.
Thanks! I use my own deployment that uses Caddy, and I based it on their example Caddyfile. But it looks like I might be good. Their reference Caddyfile accepts
application/*with an asterisk, which seems to sidestep this issue entirely!It took a few tries for me to get here, but I was able to sub from a kbin instance just now