hashmap 466ce986a6 Use round robin for peer selection in body sync ... We select a peer to ask a block randomly. Peer's send channel has capacity 10. If we need too many blocks we limit number of blocks to asks as a number of peers * 10, which means that there is some probability (pretty high) that we will overflow send buffer capacity. This fix freezes a peer list (which gives also some performance boost) and create a cycle iteraror to equally distribute requests among the peers. There is a risk that a peer may be disconnected while we are sending a request to the chanel, but stricltly speaking it was possible in the old code too, perhaps with a lower probability. Fixes #1748		2018-10-16 12:14:17 +02:00
..
common	[T4] Rename all shiftsize / cuckoo_size to edge_bits and change value for T4 (#1752)	2018-10-15 16:14:23 -07:00
grin	Use round robin for peer selection in body sync	2018-10-16 12:14:17 +02:00
mining	[T4] Rename all shiftsize / cuckoo_size to edge_bits and change value for T4 (#1752)	2018-10-15 16:14:23 -07:00
webwallet	Serve up web wallet application (#1367)	2018-08-17 18:51:50 +01:00
lib.rs	Add a peers_preferred option in Grin (#1401)	2018-08-23 15:16:04 -04:00