Looks like Mastodon is finally getting some built in functionality to fetch missing replies:

github.com/mastodon/mastodon/p…

🥳

#mastoadmin #fedifetcher

Martin Vermeer FCD reshared this.

in reply to Michael

This entry was edited (9 months ago)

reshared this

in reply to Michael

@uastronomer Well, that someone doesn't understand the purpose of robots.txt then. It’s not meant to control user-initiated connections (direct or indirect) to your webserver.

(Though I will admit it gets a little ambiguous these days when folks ask AI chatbots for specific details and the AI chatbot tries to access a webserver in realtime [like some of them do] -- because that IS user-initiated, but the AI might then also ingest whatever it finds into its DB, and some folks don't want that.)

in reply to pieceofthepie

yeah, I find the same with quite a lot of missing replies.

I’m personally still not convinced it should adhere to robots.txt. At least not to the user-agent: * part, as it really doesn’t remotely behave or act like a bot. FediFetcher honouring this block is to me akin to Mastodon honouring it, which would be nonsense (and be an effective defederation from a large number of instances).

But the community has spoken…

I actually wanted to do an analysis to see how prevalent blanket disallows are, and if any instances with blanket disallows make exceptions for FediFetcher for quite some time. I’ll see if I get around to it eventually…

This entry was edited (9 months ago)