Like Couchy, Leopard needs ReconstructSome implementations s.t. consumers can avoid reconstructing unnecessary data. Currently, it falls back to regular Reconstruct. User's like Celestia currently have to recompute all the 128 shards when only single one is needed.
Like Couchy, Leopard needs
ReconstructSome
implementations s.t. consumers can avoid reconstructing unnecessary data. Currently, it falls back to regular Reconstruct. User's like Celestia currently have to recompute all the 128 shards when only single one is needed.