My replies via Mastodon to Lemmy posts don’t get distributed as expected. For example:

It seems my reply only shows in these Lemmy servers:

  • lemmy.ml (the server of the group to which the post was made)
  • lemmy.world (the server of the post’s author)
  • ttrpg.network (the server of the comment’s author)

From some other lemmy servers, my comment is not present:

I expected that my reply would show on any other Lemmy server with subscriptions to !privacy@lemmy.ml. Does that make sense? I’m hoping to help troubleshoot federation like this as I’m super excited about ActivityPub and what it means for the internet! :)

  • Rimu@piefed.social
    6·
    11 months ago

    While both Lemmy and Mastodon use ActivityPub, the way they send ActivityPub messages to each other differs quite a bit. There is not a simple or quick solution to this. However recently I read about Lemmy having improved integration with https://a.gup.pe/, a service that adds Groups support to Mastodon so that might be something you want to try out.

    Mastodon having so many users makes it a tempting target for integration but federation with software that uses Groups properly (e.g. PeerTube, Friendica) is likely to be easier to achieve.

    • Oliver Lowe@lemmy.sdf.orgOPEnglish
      31·
      11 months ago

      Gotcha. I had a feeling something around how Mastodon doesn’t support ActivityPub Groups (yet?) would be where things are going on. Congrats on piefed, by the way. I’ll start studying the codebase now as I’m keen to understand how server-to-server communication works more deeply than I do now. Sending Announce(?) and fetching stuff from other servers…

      When I look at the ActivityPub Note object (via curl -H 'Accept: application/activity+json https://hachyderm.io/@otl/111887721960075860) I see:

      {
    "@context": [
        "https://www.w3.org/ns/activitystreams",
        {
            "ostatus": "http://ostatus.org#",
            "atomUri": "ostatus:atomUri",
            "inReplyToAtomUri": "ostatus:inReplyToAtomUri",
            "conversation": "ostatus:conversation",
            "sensitive": "as:sensitive",
            "toot": "http://joinmastodon.org/ns#",
            "votersCount": "toot:votersCount"
        }
    ],
    "id": "https://hachyderm.io/users/otl/statuses/111887721960075860",
    "type": "Note",
    "summary": null,
    "inReplyTo": "https://ttrpg.network/comment/4965852",
    "published": "2024-02-07T01:59:08Z",
    "url": "https://hachyderm.io/@otl/111887721960075860",
    "attributedTo": "https://hachyderm.io/users/otl",
    "to": [
        "https://www.w3.org/ns/activitystreams#Public"
    ],
    "cc": [
        "https://hachyderm.io/users/otl/followers",
        "https://ttrpg.network/u/Neato",
        "https://lemmy.world/u/ForgottenFlux"
    ],
    "sensitive": false,
    "atomUri": "https://hachyderm.io/users/otl/statuses/111887721960075860",
    "inReplyToAtomUri": "https://ttrpg.network/comment/4965852",
    "conversation": "tag:hachyderm.io,2024-02-06:objectId=123754186:objectType=Conversation",
    "content": "<p><span class=\"h-card\" translate=\"no\"><a href=\"https://ttrpg.network/u/Neato\" class=\"u-url mention\">@<span>Neato</span></a></span> <span class=\"h-card\" translate=\"no\"><a href=\"https://lemmy.world/u/ForgottenFlux\" class=\"u-url mention\">@<span>ForgottenFlux</span></a></span>  I totally get how you feel. One use-case I think of is machine-generated image alt-text. These are often not added to images. But with image-to-text ML models, visually-impaired people could hear a descriptions of images that before were never annotated.</p>",
    "contentMap": {
        "en": "<p><span class=\"h-card\" translate=\"no\"><a href=\"https://ttrpg.network/u/Neato\" class=\"u-url mention\">@<span>Neato</span></a></span> <span class=\"h-card\" translate=\"no\"><a href=\"https://lemmy.world/u/ForgottenFlux\" class=\"u-url mention\">@<span>ForgottenFlux</span></a></span>  I totally get how you feel. One use-case I think of is machine-generated image alt-text. These are often not added to images. But with image-to-text ML models, visually-impaired people could hear a descriptions of images that before were never annotated.</p>"
    },
    "attachment": [],
    "tag": [
        {
            "type": "Mention",
            "href": "https://ttrpg.network/u/Neato",
            "name": "@Neato@ttrpg.network"
        },
        {
            "type": "Mention",
            "href": "https://lemmy.world/u/ForgottenFlux",
            "name": "@ForgottenFlux@lemmy.world"
        }
    ],
    "replies": {
        "id": "https://hachyderm.io/users/otl/statuses/111887721960075860/replies",
        "type": "Collection",
        "first": {
            "type": "CollectionPage",
            "next": "https://hachyderm.io/users/otl/statuses/111887721960075860/replies?only_other_accounts=true&page=true",
            "partOf": "https://hachyderm.io/users/otl/statuses/111887721960075860/replies",
            "items": []
        }
    }
}

      So I’m assuming an Announce was posted to the shared inboxes at lemmy.ml, lemmy.world and ttrpg.network… hmm… I better start reading!