@tuxcrafting @solidsanek Because you know where the deleted post or account is, so you don't have to send deletes to every instance like Mastodon does

@matrix @solidsanek ohhhh i see
yeah that makes sense
but does mastodon even use that data in determining where to send the deletes?

@tuxcrafting @solidsanek I don't know if it does, but Gargron said it should.

@dirb @matrix @tuxcrafting This. I got the answer earlier and that's the only reason

@dirb @tuxcrafting @solidsanek That's the main reason, because muh harassment.

@matrix @solidsanek @tuxcrafting But imagine random people seeing posts I made publicly on the internet

@dirb @tuxcrafting @solidsanek The horror. Someone might call me out on my bullshit.

@dirb @matrix @solidsanek @tuxcrafting
> unless they track every domain that fetches that post but I don't think they do that
pleroma does that now actually, it turned out to be not too expensive on storage. but yeah, mastodon doesn't

@rin @matrix @solidsanek @tuxcrafting but I doubt that's still enough to guarantee a delete, you still should need to send deletes to those domains until you guarantee they deleted it, it could turn out they were down or just busy the first time you seen one

@dirb @rin @matrix @solidsanek @tuxcrafting and refusing deletes would become a way to cause load

@shpuld @matrix @rin @solidsanek @tuxcrafting that's why I wonder if it's worth the effort

@dirb @matrix @solidsanek @tuxcrafting I mean, nothing could truly guarantee a delete when there is more than one node involved. however the problem with instances being down/busy is solved by using a retry queue, which will retry deliveries after a certain time period if received a bad status code/timed out. it's database-backed and enabled by default in develop