[email protected] you're right, and in the ideal case, NodeBB would not send summary either, as it is automatically generated.
Right now, we are in a situation where Mastodon will utilize summary for non-note objects, but ignore content. This is why we send both, currently.
When Mastodon updates their implementation to consume preview, then we will update ours to no longer send summary.
[email protected] but to answer your initial query, I feel that summary is always going to be subservient to content. It's either a truncation or derivative of content, so if you have the capabilities to parse content, just use that in its entirety and pass over summary
(Until T&S standardize CWs, but that's a different story)
May as well ping [email protected] while I am at it 
Hi [email protected], I was just curious why [email protected] was a user and not a group on NodeBB, so I checked out its AP representation on BrowserStack.
This stood out:
...
"generator": {
"type": "Application",
"implements": {
"href": "https://datatracker.ietf.org/doc/html/rfc9421",
"name": "RFC-9421: HTTP Message Signatures"
}
},
...
I thought "how would you be able to represent multiple implementations if implements is an object?
It looks like it is supposed to be an array, according to FEP 844e.
As an aside, I guess not all WP actors are groups? I feel like we've had this conversation before though hehe.
After years of lofi being constantly suggested to me — and to be fair, it's not all bad... I'm really enjoying jazzy but not too jazzy and funky but not too funky
Super stoked that Mastodon is rolling this out after many months of testing.
That even a modicum of effort was put in to address the social failings of quote posting (as implemented on X/Twitter) is already a huge win for online public discourse.
Ben-Stokes you'll want to use NodeBB-plugin-sso-oauth2-multiple or nodebb-plugin-session-sharing 
[email protected] that's surprising, isn't aguppe just a standard 1b12 community?
What integration did you have to add?
I mean... I could stand up a replacement service using NodeBB... How hard could it be?
famous last words
yasas turning on "I'm Under Attack" mode will likely block search engine crawlers as well.
It is meant to be used as a last resort.
eeeee /world hanging (but the rest of the site being accessible) sounds like an issue with the site, not with CF...
Other pages work ok?
eeeee we have "I'm under attack" enabled on CloudFlare.
It's in response to AI crawlers, or potentially a DDoS. Not sure.
Either way it's taking down this site when the traffic floods in.
Hi [email protected], please link out to the original, not this AI summarized copy.
https://community.nodebb.org/topic/18844/backfilling-conversations-two-major-approaches
[email protected] I agree on the one hand, but simple and secure are hard to have together.
I'm not saying the Mastodon migration system can't be improved however...
This past June, I put together a write-up about two major approaches to backfilling conversations. The ability to properly backfill conversations means we will be able to make major inroads toward solving the feeling that the fediverse is quiet.
I, alongside several other members of the SWICG Forums and Threaded Discussions Task Force (ForumWG) have been working toward building implementor support for Conversational Contexts — the ability to explicitly classify a set of objects as belonging to a conversation, whether that be a topic, reply tree, or similar.
I am happy to report that we have made some wonderful inroads this past few months!
This marks a major milestone in the adoption of conversational contexts. With Mastodon on board backfill will be possible with the majority of the microblogiverse. With Lemmy and Piefed on board, backfill will be possible with the majority of the threadiverse.
Remember that [email protected] was an early adopter of conversational contexts, and we have been able to backfill from WordPress blogs for quite awhile now (so that's the blogiverse too) 
I for one, am eagerly awaiting the next version of all of these softwares!!
D1re_W0lf those rules do not seem to help, I have enabled them as well.
yasas we have been seeing the same behaviour starting perhaps a week ago. The bursts seem to happen for maybe half a day and then disappear.
Since we wanted the site to start working quickly, we opted for an allow-list approach. We turned on "I'm under attack" mode, which has some unfortunate side effects (namely, causing federation to stop).
These are our rules:
(any(http.request.headers["accept"][*] eq "application/ld+json; profile=\"https://www.w3.org/ns/activitystreams\"")) or (any(http.request.headers["accept"][*] eq "application/activity+json")) or (http.request.method eq "POST" and starts_with(http.request.uri.path, "/inbox")) or (starts_with(http.request.uri.path, "/assets")) or (starts_with(http.request.uri.path, "/.well-known"))
This lets (respectively):
KirillEvo indeed. I have added firewall rules to allow access to ActivityPub, docs, and the main contact form.
Frits-Lyneborg might be they're AI crawlers... I am not a security expert.
All I know is the traffic volume overwhelmed our little server, the IP addresses were all unique, and the user-agents all generic. Our forum and docs have been crawled by AI crawlers before, none have really stopped legitimate traffic from going through.
[email protected] you don't have to associate unfollows with the original follow. Some software do, some don't and just rely on the actor sent in the object.
Being very strict about it makes it harder to implement and test 
It's up to you.