Eelco
I concur with your guess. I add another post to this topic with some of my findings using the `time.Sleep(c.resendTimeout()) ` hack I mentioned before. I'm currently testing UTP within...
I've compared the current master with the `time.Sleep(c.resendTimeout())` hack I mentioned above and I get the following results. I'm doing the following test in the 3 node network I mentioned...
I've played around with the code trying to come up with a more elegant solution for the sleep hack. I pushed something to my fork of the repository that works...
I don't understand why the OS should be reason. Shouldn't the protocol be more gentle on the available resources (bandwidth, CPU)? I agree that my changes are not ideal but...
According these Syncthing bench there is plenty of room for improvement: The current version gives me ``` BenchmarkSyncthingTCP-8 10000 103641 ns/op 1265.24 MB/s 112 B/op 13 allocs/op BenchmarkSyncthingUDPUTP-8 1000 1785891...
Thanks. I'm seeing the same behaviour using this library with IPFS that my main reason for investigating this.
Thanks. I noticed that repository last week. I will give it try.
@Symbitic just checked the MacOS Spike Prime build but it does not mention the module being used. I stopped supporting the bluetooth native module for nodejs as I lost interest,...
Will most likely fix #40 and #45
May also fix #48.