aboutsummaryrefslogtreecommitdiffstats
path: root/src/swarm
Commit message (Collapse)AuthorAgeFilesLines
* predecessors go alongtv5 days1-3/+18
|
* covered neighbours have ittv5 days3-16/+54
|
* new changes per messagetv5 days2-1/+10
|
* neigbours by transporttv5 days1-1/+21
|
* own tree shapetv5 days1-10/+136
|
* drop full mesh stattv5 days1-3/+7
|
* delta from held headstv5 days1-1/+4
|
* offer again ungeneratedtv5 days1-32/+53
|
* stale delta offerstv5 days1-0/+19
|
* blob window growstv5 days1-5/+10
|
* delta known to sendertv5 days1-1/+7
|
* delta propagationtv5 days2-2/+18
|
* one delta at a timetv5 days2-9/+53
|
* time blob messagestv5 days4-4/+46
|
* log why a deltatv5 days1-0/+6
|
* offered deltas lapsetv5 days1-4/+14
|
* throughput from delivered messagetv5 days2-2/+59
|
* hop delaystv5 days4-2/+128
|
* benchmark delays by pathtv5 days2-0/+15
|
* fix: change publisher stops with the runtimetv5 days1-3/+8
|
* benchmark what sync bytes aretv5 days2-3/+30
|
* changes from a peer never go backtv5 days1-0/+4
|
* older delivered messages leave us leveltv5 days1-1/+3
|
* benchmark where twice comes fromtv5 days2-4/+11
|
* resend only what is neededtv5 days1-1/+1
|
* benchmark delivered and twicetv5 days2-0/+30
|
* any message taken as deliveredtv5 days2-13/+23
|
* slow round back offtv5 days2-5/+86
|
* tree messages need no answertv5 days3-6/+40
|
* hello again when heard fromtv5 days2-2/+19
| | | | | | 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 slottv6 days2-1/+23
| | | | | | 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.
* overdue sync asked after from thentv6 days2-4/+4
|
* pull after gracetv6 days2-4/+33
| | | | | | | | 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 lackingtv6 days4-10/+34
| | | | | | | 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 reachedtv6 days2-14/+43
|
* links: only news pass ontv6 days1-6/+12
|
* dashboard: show writer treestv7 days1-26/+70
|
* writers from authorstv7 days3-27/+46
|
* reach set for the sessiontv7 days2-16/+10
|
* tree changes had alreadytv7 days1-4/+14
|
* announcement round counterstv7 days2-8/+23
|
* docs after treestv7 days1-5/+6
|
* announce what trees misstv7 days2-35/+115
|
* order members by distancetv7 days2-26/+59
|
* replace eager links by treestv7 days6-137/+35
|
* add writer treestv7 days5-27/+266
|
* add writer tree benchmarktv7 days1-0/+148
|
* measure 2-hop neighbourhoodtv7 days7-8/+136
|
* dashboard: show eager links on the maptv7 days1-5/+19
|
* add eager link counterstv7 days2-0/+21
|