IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#saltr
/2024/07/14
not_bob dr|z3d: Shit, fixed.
not_bob dr|z3d: Shit, fixed.
not_bob Or, it will be fixed onces the instances catch up.
not_bob *** is tired ***
dr|z3d kick back, have a beer, take some afk time!
not_bob I'm going to take a nap.
dr|z3d someone tried to kill Trump at his latest rally
snex Old news
dr|z3d just catching up with it then.
darius if ur not cia please msg me, i have an idea
darius or not fbi or any other tla
dr|z3d you want the entire channel to message you?
darius i'll unclude ppl one by one in the order of request
darius include
dr|z3d don't hold your breath!
darius lol its about securing ethical computering
darius but i dont need anyone on gov payrolls
darius its basic stuff
dr|z3d somehow I suspect they have better things to do than lurk here, but I may be wrong.
darius here's a freebie, now that we have an upload *and* download limits for i2psnark i put forward that we also need a "downloader peers" limit. eg. i've been using a low bitrate and i found on occasion there'd be a "stalling" effect, that seemed to be caused by trying to connect to too many downloaders. When it tries to connect to a more reasonable number of peers then the 'stalling' did not happen. I would go so
darius ffar as to wonder whether the number of peers might be dynamically chosen based on provided bandwidth? Eg. uploader==-1 ? choose for me
darius i have a formula for calculating effectiveMaxPeers for i2psnark from the user-supplied maxPeersField and user-supplied bandwidth.
darius if (userSupplied_maxPeersField<8) effMax=userSupplied_maxPeersField;
darius coEff=maxPeersField*2-16
darius effMax=CEIL(maxPeersField-coEff/(bandwidth – 2))
darius / bandwidth must be 4 (kbps) or above btw
darius i put this into a spreadsheet and played with the numbers and i think the result is going to work fairly well, no matter how many max peers a person says they want, if they set the bandwidth to 4kbps the effective number of peers will be 8, if someone sets the number of desired peers very high though, it does climb a bit fast :/ so it works best if max peers is under 40
darius thats only one-way btw so if one were to consider both up and down peers with 40 max in each, that's 80 peers altogether, i doubt anyone would want to go that high.
darius ok i've been able to flatten the curve...
darius the new "else" {
darius coEff=maxPeersField*6-48
darius effMax=CEIL(maxPeersField-coEff/(bandwidth + 2))
darius } // done
dr|z3d it's all yours, zzz!
darius lol yes, i am going to smooth the curve one more time, this uses power-of-two numbers
darius the new new "else" {
darius coEff=maxPeersField*8-64
darius effMax=CEIL(maxPeersField-coEff/(bandwidth + 4))
darius } // :P
darius if you want to add some randomness to avoid fingerprinting add ±1 to effMax
darius dunno if anyone tracks connections tho so that may be redundant
darius no the curve is still too steep, for weird fringe cases like 100 peers, i should be able to make it less steep depending on the size of the maxPeersField but i need a break