indutnyat least the discovery part of it and is very happy that I have spent time to do so
indutnyit is architecturally sound
indutnywish you had a bit more comments in there, though :P
mafintoshindutny: then we wouldn’t have these nice irc chats
indutnyor some of us would have few more spare hours on their hands :)
mafintoshindutny: so not a swarm bug?
indutnysorry for being passive aggressive
indutnymafintosh: likely not a swarm bug, at least not an obvious architectural bug
indutnyI'll debug it a bit more extensively now, though
indutnyI wonder if I can setup NAT on DigitalOcean
indutnyWould be fun to poke around with hyperswarm using real-world nodes
indutnyThe fact that the clients of my app leave and join without me knowing it makes debugging quite complicated
indutnyAlso, what I have not reviewed is holepunching
mafintoshDo you know about mininet?
indutnyI do not. What is it?
mafintoshQuite effective for virtualising swarm networks
mafintoshThose are my “bindings”
indutnyare they the same?
indutnybut would you recommend to set it to
trueon clients behind NAT?
indutnyI suppose that such clients would pollute NAT to some extent with both changing IPs and ports
indutnybut this pollution should be mitigated by the queries from other nodes, right?
indutnyIs there any benefit of setting
ephemeral: true(or in other words using default)?
substackpublished firstname.lastname@example.org with authenticated channels using a custom extension
substacki'll factor that extension out into its own module next
telamohnAtleast haven't found a design on how to use multiple standalone extensions on one and the same channel without something multiplexing the
mafintoshwe are adding a feature where it can turn automatically non eph when holepunchable and been around for a bit
cblghwe're suspecting that long-lasting sessions need to restart hypercore replication, and this seems like a good mechanism for doing exactly that