In the mlem community, I clicked on this post:

https://lemmy.ml/post/1223230

I was shown the title and post abouta protest taking place at Reddit HQ and in Ireland. You can see in the screenshot, the post doesn’t really match the community and you can see it was posted to the Reddit community. The comments were all about the mlem app.

There was another post I saw earlier today about someone being shown another user’s page when refreshing. Very interesting.

  • qprimedEnglish
    2·
    1 year ago

    jerboa also exhibts the same tendency sometimes and does not clear the target layer while composing the new layer. putting it down to growing pains. as more dev hands and eyeballs turn to Lemmy and the various clients, things should improve dramatically. if your post was not re: the actual mlem app, but instead the web instance, then I am not sure… the web user interface should be pretty smooth.

    edit: typos

      • qprimedEnglish
        2·
        1 year ago

        the lemmy.ml instance has been under pretty intense network pressure for a while now and was consistently serving up errors on both the website and apps for a while. during that time there were also reports of ActivityPub updates not getting through to other instances. depending on how activityPub handles queues and retries, there may be unsynched communities between instances. not a massive issue and something that should get sorted out as lemmy’s use of the actvitypub protocol gets refined.

        in this specific issue (serving up wrong data via the web interface), it’s also possible that the database backend on lemmy.ml is not quite coherent due to local service outages.

        so we may have several types of issues going in here…

        1. possible bugs to track down in lemmy’s activitypub usage
        2. possible fixups on the lemmy.ml db backend using a sysadmin stick and poking it to correct any issues there

        the server running lemmy.ml instance was just upgraded, so that may help with active issues. if there are instance DB issues, those may need to wait until the someone gets time to take a look. obviously any “fragility” issues will hopefully get sorted out over time.

        @nutomic and @dessalines have a ton going on right now.