For example, if someone forked Lemmy, and made an LGBT or POC instance with a different word filter how would federation work with instances with incompatible filters?
For example, if someone forked Lemmy, and made an LGBT or POC instance with a different word filter how would federation work with instances with incompatible filters?
The slur filter has nothing to do with federation, so it would work just fine (assuming both instances add each other to their allowlist/don’t block each other).
However, say I write a word in instance A.
Instance A appears to be very open and accepting of the use of commonly blocked words as the n-word, for whatever reason.
Instance A is also federated with instance B, which blocks the n-word.
What happens If a post containing the n-word appears on B homepage?
If I recall correctly, in the github issue about the words filter it was said the ones blocked are replaced when accepting the post/comment, and not while retrieving it to be viewed.
This means that if no other checks happen, the n-word would pop up in B homepage without any problems, right?
Or have you already tought about it and fixed this case?
(I know B could just block A, but what if the filter is changed after they accepted to be federated together, to block a new word for example? This is the situation I’m imagining)
The current implementation will replace any slurs in posts/comments from other instances with removed.
Ok, that’s good then.
What happens in the other direction? I mean, If a post with a more strict filter appears on a less filtered server.
Do the words remain censored?
Yes, in that case it is filtered by the API right after it is submitted. Of course we can think about changes to the slur filter, so it works better with federation.
Thanks for your explanation, I was curious!