For techies and why we need this stress test:

Minibits started to have issues with keeping properly incremented recovery indexes (simple counters per each mint keyset) when zaps were introduced.

They kept flowing from nostr relays in parallel, racing the counter that needs to be strictly ordered otherwise blinded ecash sent to the mint will fail and cause the wallet to brick.

Few weeks ago Minibits implemented pessimistic lock on counters. It has improved the situation, however waiting for locks to release blocks singlethread js processing making wallet unresponsive and there were troubles on how to solve edge cases like force-releasing the lock. Simply said, they are great as last resort safety net, but should not play prime.

Enter new synchronous queue that orders coming zaps into strict order a process them one by one, offloading not critical but time consuming tasks to asynchronous event handlers...

This is what we were testing, now released in #minibits over the air update.

So, all heavy zap collectors, head to Settings to get an update!

Reply to this note

Please Login to reply.

Discussion

No replies yet.