continuum - P2P replicated document storage service

	Commit message (Collapse)	Author	Age	Files	Lines
*	broken pushes refused	tv	10 days	1	-11/+31
\|
*	pushed have is the senders heads	tv	10 days	1	-10/+39
\|
*	pushed heads as tips	tv	10 days	1	-6/+60
\|
*	push changes experiment	tv	10 days	1	-0/+123
\|
*	announce only to those behind	tv	10 days	2	-0/+48
\|
*	asked known elsewhere	tv	10 days	1	-7/+36
\|
*	predecessors go along	tv	10 days	1	-3/+60
\|
*	covered neighbours have it	tv	10 days	2	-9/+52
\|
*	new changes per message	tv	10 days	1	-3/+5
\|
*	delta from held heads	tv	10 days	2	-25/+79
\|
*	offer again ungenerated	tv	10 days	1	-0/+86
\|
*	stale delta offers	tv	10 days	1	-0/+60
\|
*	delta known to sender	tv	10 days	1	-5/+7
\|
*	offered deltas lapse	tv	10 days	1	-3/+38
\|
*	fix: level means the peer has our heads	tv	10 days	1	-6/+13
\|
*	tests count the changes sent	tv	10 days	1	-5/+7
\|
*	no reset for heads still on their way	tv	10 days	1	-1/+44
\|
*	benchmark what sync bytes are	tv	10 days	1	-0/+12
\|
*	changes from a peer never go back	tv	10 days	1	-0/+48
\|
*	older delivered messages leave us level	tv	10 days	1	-35/+73
\|
*	resend only what is needed	tv	10 days	1	-26/+61
\|
*	delivered means has all up to our heads	tv	10 days	1	-13/+59
\|
*	any message taken as delivered	tv	10 days	1	-26/+82
\|
*	tree messages need no answer	tv	10 days	1	-0/+194
\|
*	add tree tests	tv	12 days	1	-0/+35
\|
*	fixup 700, 701, 702	tv	12 days	2	-1/+38
\|
*	writers from authors	tv	12 days	1	-0/+22
\|
*	tree changes had already	tv	12 days	1	-0/+7
\|
*	replace eager links by trees	tv	12 days	1	-7/+0
\|
*	introduce eager links	tv	12 days	1	-0/+7
\|
*	document: put user in commit message trailer	tv	12 days	1	-5/+70
\|
*	api: /admin/changes read ops, not diffs	tv	13 days	1	-0/+86
\| \| \| \| \| \| \| \| \| \| \| \|	The new store::changed_by reads each change's own operations instead of diffing the document before and after it. Each object's path is found once per change, as it was right after that change. So the cost now grows with the size of the change, not the size of the map it touched. - It names paths the same way the diff did: a map entry by its key, and list and text edits by the list or text. - It relies on import_obj, which Automerge marks as hidden from its docs, to turn an operation's object ID back into an object.
*	api: drop $float as JSON number is OK	tv	13 days	2	-27/+40
\|
*	api: add type casts	tv	13 days	2	-0/+196
\|
*	add remake benchmarks	tv	13 days	2	-3/+3
\|
*	api: types are sticky	tv	13 days	2	-54/+216
\|
*	api: add application/vnd.continuum.typed+json	tv	13 days	2	-131/+308
\|
*	api: add application/vnd.continuum.* scalar types	tv	13 days	3	-45/+474
\|
*	add document benchmarks	tv	13 days	2	-0/+172
\|
*	document: create new compound values at once	tv	13 days	1	-11/+29
\| \| \| \| \|	This speeds up large writes as new maps and lists are created in one batch operation instead of one operation per value.
*	document: introduce patch log	tv	13 days	3	-17/+143
\|
*	api: add support for counters	tv	13 days	1	-0/+57
\|
*	api: allow reading past versions	tv	13 days	3	-0/+253
\|
*	api: introduce /admin/changes	tv	13 days	1	-1/+1
\|
*	record node ID and API user for every write	tv	13 days	1	-10/+44
\|
*	api: PUT changes only what differs	tv	13 days	1	-23/+148
\|
*	large messages over slow links are deltas	tv	13 days	1	-7/+16
\|
*	document: allow writes during rejoin	tv	13 days	2	-13/+28
\|
*	document: reject writes during rejoin	tv	13 days	2	-0/+11
\|
*	add save-after-rollback test	tv	2026-10-31	1	-0/+34
\|