aboutsummaryrefslogtreecommitdiffstats
path: root/docs/peers.md
Commit message (Collapse)AuthorAgeFilesLines
* deflate blob chunkstv8 days1-2/+3
|
* deflate blob chunkstv8 days1-0/+5
|
* epoch tag in synctv8 days1-1/+4
|
* no swarm ID in synctv8 days1-1/+2
|
* announce only to those behindtv8 days1-1/+2
|
* predecessors go alongtv8 days1-1/+5
|
* covered neighbours have ittv8 days1-0/+8
|
* delta from held headstv8 days1-2/+3
|
* offer again ungeneratedtv8 days1-1/+2
|
* stale delta offerstv8 days1-0/+3
|
* delta known to sendertv8 days1-1/+2
|
* one delta at a timetv8 days1-1/+5
|
* document messages that overtaketv8 days1-4/+20
|
* any message taken as deliveredtv8 days1-7/+9
|
* slow round back offtv8 days1-4/+8
|
* tree messages need no answertv8 days1-3/+10
|
* hello again when heard fromtv8 days1-1/+4
| | | | | | 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.
* unanswered pull frees the slottv9 days1-3/+4
| | | | | | When the answer to a pull is overdue (the sync reply timeout: 1 s in the simtests, 10 s in production), the pull gives up the slot, and the next waiting neighbour is pulled at once.
* no heal while both move ontv9 days1-1/+1
| | | | | | | 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.
* pull after gracetv9 days1-1/+3
| | | | | | | | Before pulling, a node waits one measured round trip to the announcer (1 s at most) and pulls only what is still missing. If the round trip isn't known yet, it pulls at once, as now. On the clean benchmark links the round trip is a few milliseconds, so those scenarios shouldn't change.
* announcements answered when lackingtv9 days1-2/+3
| | | | | | | A neighbour now answers an announcement only if it lacks some of the announced heads. Heads that merely differ are no longer enough, since they do while writes go on. The check reads the document, so it runs off the message consumer, as the pull check already does.
* unanswered members not reachedtv10 days1-5/+10
|
* writers from authorstv10 days1-5/+7
|
* reach set for the sessiontv10 days1-2/+4
|
* tree changes had alreadytv10 days1-1/+3
|
* announce what trees misstv10 days1-1/+6
|
* order members by distancetv10 days1-4/+6
|
* replace eager links by treestv10 days1-9/+2
|
* add writer treestv10 days1-0/+16
|
* measure 2-hop neighbourhoodtv10 days1-0/+9
|
* introduce eager linkstv10 days1-1/+9
|
* skip neighbours with our headstv10 days1-1/+4
|
* announcements answered only if heads differtv10 days1-2/+3
|
* retry new peers before probingtv10 days1-1/+3
|
* probe list memberstv10 days1-1/+3
|
* always announce new headstv10 days1-5/+4
|
* swarm: announce to whom writers can't reachtv10 days1-1/+4
|
* add documentationtv11 days1-0/+358
generated by cgit v1.3.1 (git 2.54.0) at 2026-11-12 23:11:17 +0000