I am catching errors from hypercore.clear() method regularly. If the error message is to be trusted "No node found" could be a non-problem because clear() should be a node delete. I would like to create a test case to reproduce this, it involves replication between two hypercores involving a specific write pattern and could take awhile to pin down.
Additionally, supporting --limit involves replacing the default random-access-file storage provider with something that can partition data across multiple files, my fork of multi-random-access is a start towards this.
I am catching errors from
hypercore.clear()method regularly. If the error message is to be trusted "No node found" could be a non-problem becauseclear()should be a node delete. I would like to create a test case to reproduce this, it involves replication between two hypercores involving a specific write pattern and could take awhile to pin down.https://github.com/rhodey/radiowitness/blob/fde9f3915c4e88761e19107fe991018630f4c980/lib/js/rw-peer/lib/ws-mirror.js#L77
Additionally, supporting
--limitinvolves replacing the default random-access-file storage provider with something that can partition data across multiple files, my fork of multi-random-access is a start towards this.