calling all MST/sync 1.1 sickos: what is happening behind the scenes when I get a repo commit from a sync 1.1 relay with a prevData CID that does NOT match the previous commit's data CID? #atdev

oooh. your own relay? (do you know if it’s running in lenient mode? (idk if lenient mode allows bad sync1.1 commits or just allows non-sync1.1 (bluesky sync1.1 relays run in lenient mode rn)))

no, this was coming from relay1.us-east.bsky.network just something I noticed last night... did:plc:e4elbtctnfqocyfcml6h2lf7 was deleting listitems every few seconds and in each commit prevData was not matching the data CID of the previous commit

I was running this tool (github.com/bluesky-soci...) and getting "repo state desynced" which kicked off a whole resync process each time not seeing the issue anymore, and also not seeing the listitem deletes anymore either (good thinking on lenient mode, I hadn't considered that)

oh interesting. if it’s not the upstream repo being sketchy, it would be worrying. (hope there’s a ratelimit somewhere in nexus for repo re-syncs at least by the official release time)