continuum - P2P replicated document storage service

	Commit message (Collapse)	Author	Age	Files	Lines
*	get swarm RTT from paths	tv	3 days	2	-133/+91
\|
*	don't sample lost PONGs	tv	3 days	1	-29/+23
\|
*	keepalive takes turns	tv	3 days	3	-3/+19
\|
*	move keepalive and firewall to paths	tv	3 days	6	-62/+196
\|
*	measure paths	tv	3 days	1	-1/+4
\|
*	rename PING/PONG to STATUS/STATUS_REPLY	tv	3 days	9	-70/+72
\|
*	build topology per transport	tv	4 days	1	-4/+128
\|
*	use libp2p mDNS	tv	4 days	1	-0/+11
\|
*	add support for several transports	tv	4 days	1	-1/+7
\|
*	admit nodeId and addresses in the API	tv	4 days	1	-1/+8
\|
*	add identity layer	tv	4 days	1	-2/+6
\|
*	introduce addresses	tv	4 days	4	-16/+100
\|
*	introduce node keys	tv	4 days	1	-6/+6
\|
*	talk about "transport" unless mycelium is meant	tv	4 days	9	-32/+37
\|
*	add smaller topology test	tv	4 days	1	-19/+35
\|
*	reject and warn about too large delta blobs	tv	4 days	2	-0/+9
\|
*	add delta blobs	tv	4 days	3	-19/+312
\|
*	make clippy happy again	tv	4 days	5	-35/+35
\|
*	warn about dropped blob requests	tv	4 days	1	-0/+28
\|
*	add sync and relay stats	tv	4 days	3	-1/+44
\|
*	introduce bounded blob downloads	tv	4 days	1	-13/+41
\|
*	warn if swarm key is readable by others	tv	4 days	1	-0/+22
\|
*	add farewell ping	tv	4 days	2	-2/+22
\|
*	decode sync messages lazily	tv	4 days	1	-1/+3
\| \| \| \| \| \| \|	Sync messages are now being decoded only where they are applied: after membership has been checked, in the sync lane's blocking task. Debug now shows only the size instead of the whole message.
*	keepalive sends head samples alongside digest	tv	4 days	1	-21/+22
\| \| \| \| \| \|	When there are more than 32 heads that can be send in a single keepalive message, the digest is accompanied by 32 heads, a different window each time, so after a few keepalives every head has been listed.
*	add pull queue	tv	4 days	3	-34/+84
\|
*	send sync messages off worker	tv	4 days	1	-1/+23
\|
*	read stamps sync and lock free	tv	4 days	4	-13/+20
\|
*	add sync lanes	tv	4 days	3	-2/+70
\|
*	compute member pair scores cheaply	tv	4 days	1	-22/+65
\|
*	unreachability doesn't expire	tv	4 days	2	-9/+3
\|
*	forget records only once	tv	5 days	2	-9/+37
\|
*	fix race between alive check and forgetting	tv	5 days	2	-8/+23
\|
*	forgotten nodes cancel their own records	tv	5 days	1	-15/+14
\|
*	add session IDs	tv	5 days	1	-15/+23
\| \| \| \| \|	This prevents the regenerate flag from being lost across reset_sync during an in-flight send.
*	number handshakes	tv	5 days	2	-7/+32
\| \| \| \|	This prevents late hellos from resetting the session.
*	keep silent strangers for a day	tv	5 days	2	-3/+40
\|
*	send bounded peer lists	tv	5 days	2	-7/+38
\|
*	wrap unix timestamps in newtype	tv	5 days	3	-3/+3
\|
*	allow humand-friendly durations and sizes in CLI	tv	5 days	2	-2/+2
\|
*	SyncSessions owns Automerge's sync state	tv	5 days	2	-80/+12
\|
*	unify keepalive	tv	5 days	1	-15/+23
\|
*	split peer module	tv	5 days	9	-915/+995
\|
*	unify peers snapshot logic	tv	5 days	1	-11/+3
\|
*	unify lock logic	tv	5 days	10	-70/+31
\|
*	move protocol paths to dedicated module	tv	5 days	3	-7/+7
\|
*	refresh peers after stall	tv	5 days	2	-1/+19
\|
*	add random swarms test	tv	5 days	1	-0/+62
\|
*	add clock module	tv	5 days	3	-6/+11
\| \| \| \| \|	The Clock module is the system clock that every unix timestamp new goes through. It comes with an offset that only tests can set.
*	avoid stalled source	tv	5 days	1	-2/+24
\| \| \| \| \| \| \| \| \| \| \| \| \|	A source given up for not answering our sync messages could become the source again straight away. Whoever asks first after the reset takes the role, and that's often the same peer, which would then stall for another delivery limit. Now, a stalled source isn't chosen again for one more delivery_limit. If it's the only candidate, catching up waits that long and then tries it again. Before, a retry with that peer couldn't have helped any sooner anyway: a source is only given up after a whole delivery limit without a sync message, and that clock restarts when it is chosen again.