continuum - P2P replicated document storage service

	Commit message (Collapse)	Author	Age	Files	Lines
*	dissemination trait	tv	24 hours	1	-0/+2
\|
*	topology trait	tv	24 hours	1	-0/+2
\|
*	handshakes counted	tv	34 hours	1	-0/+56
\|
*	working links recorded	tv	35 hours	1	-0/+2
\|
*	worker sends off the pass	tv	36 hours	1	-1/+6
\|
*	suspend detected	tv	36 hours	1	-0/+12
\|
*	one grace wait per peer	tv	36 hours	1	-0/+4
\|
*	asks by asker	tv	43 hours	1	-0/+5
\|
*	round trips in trees	tv	2 days	1	-0/+5
\|
*	push by default	tv	2 days	1	-4/+0
\|
*	push changes experiment	tv	3 days	1	-0/+4
\|
*	announce only to those behind	tv	3 days	1	-2/+2
\|
*	asked again counter	tv	3 days	1	-0/+5
\|
*	hop delays	tv	3 days	1	-0/+3
\|
*	slow round back off	tv	3 days	2	-0/+5
\|
*	hello again when heard from	tv	3 days	1	-0/+52
\| \| \| \| \| \|	While we wait for a peer's HELLO_REPLY, anything else that comes from it makes our HELLO go again at once, at most once a second, instead of when the backoff says.
*	no heal while both move on	tv	3 days	1	-3/+49
\| \| \| \| \| \| \|	The heal clock now also restarts when our heads and the peer's have both changed since the last STATUS that showed them differing. While writes go on, two neighbours's heads differ because of changes still on their way, and that alone no longer leads to a heal.
*	overdue sync asked after from then	tv	3 days	1	-45/+88
\|
*	unanswered sync asked after	tv	3 days	1	-6/+28
\| \| \| \| \|	An overdue sync answer is now followed by a PING too, and the peer stays reached once it answers the PING.
*	unanswered members not reached	tv	4 days	2	-3/+88
\|
*	announce what trees miss	tv	4 days	1	-0/+3
\|
*	replace eager links by trees	tv	4 days	1	-10/+0
\|
*	add writer trees	tv	4 days	2	-3/+8
\|
*	measure 2-hop neighbourhood	tv	5 days	2	-0/+7
\|
*	introduce eager links	tv	5 days	1	-0/+10
\|
*	skip neighbours with our heads	tv	5 days	1	-0/+5
\|
*	retry new peers before probing	tv	5 days	1	-5/+37
\|
*	probe list members	tv	5 days	1	-3/+8
\|
*	add stream heartbeat	tv	5 days	1	-0/+3
\|
*	dashboard: double-click opens peer dashboard	tv	5 days	1	-0/+3
\|
*	api: add x-continuum-heads to writes	tv	6 days	1	-6/+10
\|
*	record node ID and API user for every write	tv	6 days	1	-1/+24
\|
*	large messages over slow links are deltas	tv	6 days	1	-0/+29
\|
*	introduce per-transport sync reply timeout	tv	6 days	2	-11/+14
\|
*	peer worker forget records on change	tv	6 days	1	-0/+3
\|
*	peer worker sleeps until deadline, not fixed time	tv	6 days	2	-18/+60
\|
*	bootstrap: replace pull by progress notify	tv	6 days	2	-0/+6
\|
*	document: allow writes during rejoin	tv	6 days	3	-110/+207
\|
*	LWW everywhere, mostly	tv	6 days	2	-0/+31
\|
*	remake: copy stale view when writes stop	tv	6 days	1	-0/+13
\| \| \| \|	This ensures that stale reads show every write that was answered.
*	state: drop epoch_tx	tv	6 days	2	-8/+7
\| \| \| \|	The stamp is the only lock-free epoch label.
*	document: reject writes during rejoin	tv	6 days	2	-5/+23
\|
*	use Automerge as Document instead of AutoCommit	tv	7 days	2	-14/+36
\| \| \| \|	This allows a lot of operations to drop mut.
*	move SyncSession to document	tv	7 days	3	-560/+2
\|
*	move all automerge references to document	tv	7 days	2	-20/+1
\|
*	sync messages are bytes outside the session	tv	7 days	2	-13/+25
\| \| \| \| \| \| \| \|	SyncPayload is now an opaque byte container in transport. SyncSession::generate returns the encoded payload, state::decode_sync(&payload) decodes it off the document lock (as before), and receive takes the decoded message, which is a type only the session can open. transport/mod.rs no longer imports Automerge.
*	commits go through the replica	tv	7 days	1	-13/+4
\|
*	rename AutoMerge to Document	tv	7 days	4	-36/+36
\|
*	detect local clock jumps	tv	7 days	1	-2/+2
\|
*	remake needs sane clock	tv	7 days	1	-0/+4
\|