- removed defer/recover/panic where warranted
- found and eliminated race conditions when extremely high connection pressure was hit
- resolved race condition in blossomsub around mesh maintenance
- incorporated resource manager scaling to work off low/hi watermark values
* v1.4.20 base
* add inmemory dev mock for hypergraph
* add simple rdf + tr
* Update config.go (#234)
2 of bootstrap nodes are going to be closed due to low performances. Will consider to replace with better specs.
* go mod tidy
* go mod tidy
* bump name in readme
---------
Co-authored-by: 0xOzgur <29779769+0xOzgur@users.noreply.github.com>
* experimental: switch mutex to RW to see if it alleviates backpressure on peer info
* relax mutex req
* reject unknown messages
* open the floodgates
* adjust message handler to use goroutine, i'll probably regret this
* switch that back, it was regret
* further discovery
* log more data
* forcibly block channel when unbounded
* else
* make it configurable so bootstrap peers are the only ones putting up with this.
* ok, non-starter, let's try a different route
* further tweaking
* let the peer info flow uninhibited
* final burn off on master
* final adjustments
