aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* benchmark delays by pathtv4 days3-0/+28
|
* tests count the changes senttv4 days1-5/+7
|
* document messages that overtaketv4 days1-4/+20
|
* fix: change publisher stops with the runtimetv4 days1-3/+8
|
* no reset for heads still on their waytv4 days1-1/+44
|
* benchmark what sync bytes aretv4 days4-3/+74
|
* changes from a peer never go backtv4 days2-0/+52
|
* older delivered messages leave us leveltv4 days2-36/+76
|
* benchmark where twice comes fromtv4 days3-6/+20
|
* resend only what is neededtv4 days2-27/+62
|
* benchmark delivered and twicetv4 days3-0/+58
|
* delivered means has all up to our headstv4 days1-13/+59
|
* any message taken as deliveredtv4 days6-47/+121
|
* slow round back offtv4 days7-10/+102
|
* tree messages need no answertv4 days6-10/+245
|
* hello again when heard fromtv4 days4-3/+75
| | | | | | 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.
* bump pulled_changes_go_along_their_writers_treestv4 days1-0/+3
|
* unanswered pull frees the slottv4 days3-4/+27
| | | | | | 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 ontv4 days3-5/+51
| | | | | | | 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.
* bump pulled_changes_go_along_their_writers_treestv5 days1-2/+52
|
* overdue sync asked after from thentv5 days3-49/+92
|
* pull after gracetv5 days3-5/+36
| | | | | | | | 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 lackingtv5 days7-15/+40
| | | | | | | 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 sync asked aftertv5 days1-6/+28
| | | | | An overdue sync answer is now followed by a PING too, and the peer stays reached once it answers the PING.
* simtest: benchmark incidentstv5 days1-6/+21
|
* unanswered members not reachedtv5 days7-24/+143
|
* add member away benchmarktv5 days2-1/+36
|
* dashboard: show reach settv5 days2-2/+24
|
* links: only news pass ontv5 days1-6/+12
|
* dashboard: show writer treestv5 days5-26/+233
|
* add tree teststv5 days2-2/+144
|
* update architecture doctv5 days1-4/+7
|
* fixup 700, 701, 702tv5 days4-42/+43
|
* writers from authorstv5 days5-32/+75
|
* reach set for the sessiontv5 days6-22/+18
|
* add late arrivals benchmarktv5 days1-3/+49
|
* tree changes had alreadytv5 days3-5/+24
|
* announcement round counterstv5 days7-10/+53
|
* docs after treestv5 days3-10/+12
|
* announce what trees misstv5 days5-36/+125
|
* parallel propagation benchmarktv5 days2-15/+57
|
* quicker propagation benchmarktv6 days2-15/+23
|
* order members by distancetv6 days4-41/+69
|
* replace eager links by treestv6 days28-339/+61
|
* add writer treestv6 days12-33/+339
|
* add writer tree benchmarktv6 days2-1/+152
|
* measure 2-hop neighbourhoodtv6 days23-11/+250
|
* add propagation benchmarktv6 days4-1/+404
|
* dashboard: show eager links on the maptv6 days8-11/+74
|
* add eager link counterstv6 days6-1/+57
|