* fix regression: llarp::Context::CallSafe was not deferring call on startup
* prepare for ipv6 on android
* remove test case
* limit calls to service node list updates for when we are synching the chain we dont spam with list updates
[ Jason Rhinelander ]
* Limit builds to 6 jobs (4 for arm)
* Add missing getIPv6 implementation
* Add systemd-resolved dynamic DNS updating
* Put systemd ifdef around the headers, too
* format
* Fix headers & grammar for non-systemd
[ Thomas Winget ]
* Fixes subtle memory leak, adds comments
[ Jeff Becker ]
* lokinetmon updates
* fix up outbound session
* limit path reanimation
* parameterize traffic alignement timeout in dns lookups
* don't rehash decaying hashsets that is no bueno and probably leaks like the pipes in tom's apartment
* compat for lokinet 0.8.x
* consmetic fixes
* win32 installer
* make path alignment timeout configuable
* make default log level warn fixes #1593
* change granularity to seconds
* try fixing std::shared_ptr leak with paths
* if a path's latency is zero dont use it because it's not actually a zero latency path it's probably about to be failed or timed out
* dont invalidate cache on lns timeout, only on explicit negative
* allow more than one outbound context per endpoint for fallback in case one outbound context takes a shit
* only use alive paths for lookups only track usage from sessions and don't care about intro timeouts
* use correct intro
* re-add parans
* report block height reported by oxend in systemd status
* report block height reported by oxend in systemd status
* format
* make profiling far stricter for path builds
* limit path builds across all builders
* always do path tests
[ David Arnold ]
* Tyops
[ Jeff Becker ]
* fixups in service endpoint
* don't send nx if we have an address already mapped
* in the event that a session is removed and then gets more traffic we would re-add the session with everything blank if there was more traffic in the same tick.
* shift off bad intros when we get a discard message, this happens if the other side's intro expires or the pivot restarts for whatever reason.
* when we have to reestablish an outbound session use a much much higher timeout for restablishment so it doesn't time out and get into a state that's totally screwed.
Jason Rhinelander
3 years ago