Looks like I was wrong before: whilst it’s true that Lemmy isn’t always great at sending an ‘Accept’ back, that’s not the issue here. I was thinking about it from the perspective of a Person following a Group, not a Group following a Group.
It’ll be interesting to see how the Lemmy devs respond. It’s not really relevant for PieFed, because I’m not sure that there’s communities there that have enough activity to warrant syncing to a category, but I thought about what would happen if you tried, and it would fail. The follower would be checked to see if they are a Person, and the request would go no further if not (there’s no handy error returned, though).
I’m not sure it would be possible to enable, with the way the database is currently set up. We have a ‘community_member’ table, that looks like:
pyfedi=# \d community_member
Table "public.community_member"
Column | Type | Collation | Nullable | Default
------------------+-----------------------------+-----------+----------+---------
user_id | integer | | not null |
community_id | integer | | not null |
is_moderator | boolean | | |
is_owner | boolean | | |
is_banned | boolean | | |
notify_new_posts | boolean | | |
created_at | timestamp without time zone | | |
Indexes:
"community_member_pkey" PRIMARY KEY, btree (user_id, community_id)
"ix_community_member_is_banned" btree (is_banned)
Foreign-key constraints:
"community_member_community_id_fkey" FOREIGN KEY (community_id) REFERENCES community(id)
"community_member_user_id_fkey" FOREIGN KEY (user_id) REFERENCES "user"(id)
So if the ‘user_id’ was actually a Group instead of a Person, then it would have a join to the wrong table.
@rglullis@communick.news No, I don’t there’s any overlap between PieFed and Fediverser either. The potential of Fediverser seems like it got cut off at the knees by how widely defederated alien.top is.