Show Idle (>14 d.) Chans
← 2022-07-09 | 2022-07-11 → ↓
phf[asciilifeform]: and right after i said that i got two copies of that message. so blatta doesn't filter dupes?
crtdaydreams[asciilifeform]: http://logs.bitdash.io/pest/2022-07-09#1008935 << not quite sure what happened here with $USER
phf[asciilifeform]: http://logs.bitdash.io/pest/2022-07-09#1008937 well, i'm only connected to you, and don't dedup and i got two copies of http://logs.bitdash.io/pest/2022-07-09#1008936
phf[asciilifeform]: i'm actually connected to you :)
phf[asciilifeform|asciilifeform|asciilifeform]: possibly gets sent elsewhere, eventually comes back but doesn't get deduped?
phf[asciilifeform|asciilifeform|asciilifeform]: awt: well if you look the whole nonce (the first 16 bytes) is different
phf[asciilifeform|asciilifeform|asciilifeform]: awt: but that's the red packet from the wire, which i'm not otherwise modifying so it definitely came in that way.
phf[asciilifeform|asciilifeform|asciilifeform]: i also find it odd, that the nonce is all numbers in the 50-60 range :o
phf[asciilifeform|asciilifeform]: so i'm looking at all the packets in my backlog, and all (except for my own) have busted nonce
phf[asciilifeform]: here's a random selection of nonces http://paste.deedbot.org/?id=8RhY
crtdaydreams[asciilifeform]: http://logs.bitdash.io/pest/2022-07-09#1008938 << ty shinohai, it appears we're not peered correctly yet?
bitbot[asciilifeform]: Logged on 2022-07-09 22:57:51 awt[billymg|signpost]: phf: every time you see a message coming through asciilifeform's station, he's likely filtered out 11 dupes of it.
bitbot[asciilifeform]: Logged on 2022-07-09 22:52:52 crtdaydreams[billymg]: oh good god what is happening
bitbot[asciilifeform]: Logged on 2022-07-09 23:05:54 bitbot: Logged on 2022-07-09 22:52:52 crtdaydreams[billymg]: oh good god what is happening
awt[asciilifeform]: Interesting to note that the bounce counts are different
awt[asciilifeform]: right but the hashes (d6862d44...) appear to be the same
bitbot[asciilifeform]: Logged on 2022-07-09 22:52:52 chad[billymg]: hello?
bitbot[asciilifeform]: Logged on 2022-07-09 22:58:26 shinohai[signpost]: welcome crtdaydreams
crtdaydreams[asciilifeform]: loox like ascii's logotron hit the hay
awt[asciilifeform]: http://logs.bitdash.io/pest/2022-07-09#1008967 << lol yeah no clue what is special about those messages, or the particular clique of nodes involved.
bitbot[asciilifeform]: Logged on 2022-07-09 23:33:08 phf[billymg]: awt: https://www.youtube.com/watch?v=TzsNh3-FnHQ
awt[asciilifeform]: There was an issue in the past where messages ending in whitespace would be rebroadcast with the whitespace chopped off, causing dupes.
awt[asciilifeform]: phf why are the first two integers different here? 50 vs. 52: http://paste.deedbot.org/?id=Mhfb
awt[asciilifeform]: ok so you're getting the same message with two different nonces?
phf[asciilifeform]: fwiw different nonce shouldn't result in different message hash256
signpost[asciilifeform]: ohey crt, lemme get you peering info.
bitbot[asciilifeform]: Logged on 2022-07-10 00:03:37 awt[billymg]: ok so you're getting the same message with two different nonces?
phf[asciilifeform]: not sure if nonce is supposed to change during rebroadcast, but that's probably not a bad thing?
awt[asciilifeform]: phf yeah nonce shouldn't affect dedup
signpost[asciilifeform]: your station is probably syncing and replaying messages.
phf[asciilifeform]: yah,we discovered a bug while investigating another bug:x
signpost[asciilifeform]: message.py#298 does appear to return a string
signpost[asciilifeform]: oughta be bytearray or something eh?
signpost[asciilifeform]: and also not changing
signpost[asciilifeform]: nonce being an ascii version of digits is a b00g though eh?
signpost[asciilifeform]: looks like the broadcast handler uses pack, which always generates a new nonce.
awt[asciilifeform]: my nonce generator takes the short bus
signpost-dev[asciilifeform]: naively changed message.py#300 to a urandom call like the next func uses.
signpost[asciilifeform]: probably can just snip that guy and pull the urandom call into pack
signpost[asciilifeform]: I'd be inclined to do same to gen_rubbish_body since nobody else calls that either.
awt[asciilifeform]: signpost: makes since since the urandom call is pretty straigt forward
awt[asciilifeform]: past my bedtime. please don't find any more bugs until tomorrow.
signpost[asciilifeform]: lol, goodnight
asciilifeform: http://logs.nosuchlabs.com/log/pest/2022-07-10#1009175 << erry ciphration, whenever it takes place, gets a fresh nonce. (oghta have said this in spec, but thought was obv)
dulapbot: Logged on 2022-07-10 00:13:49 phf[asciilifeform]: not sure if nonce is supposed to change during rebroadcast, but that's probably not a bad thing?
asciilifeform observes that local station bites it on erry single echo from dulapbot
asciilifeform observes also that billymg's log aint seeing dulapbot's echos
asciilifeform to bed shortly
phf[asciilifeform]: asciilifeform: i'm connected to awt, and i didn't get dulapbot response to your link
asciilifeform: phf: likely because asciilifeform's station bit it immediately, and replays dun quite work
asciilifeform: ( dulapbot atm connected solely via asciilifeform's station )
awt[asciilifeform]: phf: every time you see a message coming through asciilifeform's station, he's likely filtered out 11 dupes of it.
shinohai[asciilifeform]: welcome crtdaydreams
shinohai[asciilifeform]: http://logs.bitdash.io/pest/2022-07-09#1008968 << Weird, you're in my peer list, etc. Didja try to DM?
crtdaydreams[asciilifeform]: shinohai you're not showing up in stanis logs either http://logs.nosuchlabs.com/log/pest
bitbot[asciilifeform]: Logged on 2022-07-09 23:34:24 crtdaydreams[signpost]: http://logs.bitdash.io/pest/2022-07-09#1008938 << ty shinohai, it appears we're not peered correctly yet?
shinohai[asciilifeform]: crtdaydreams: sent you a dm ping
shinohai[asciilifeform] also needs to dig up jonsykkel's peering info, must have forgotten to add it in during last overhaul ....
crtdaydreams[asciilifeform]: shinohai: for some reason you weren't in my AT, nor even peered and not showing up in my peers but doing a getdata sync somehow or another?
crtdaydreams[asciilifeform]: shinohai, best bet might try a rekey, setting up now
shinohai[asciilifeform]: crtdaydreams: My AT info was posted here in chan the other day, lemme know if you need me to link it.
crtdaydreams[asciilifeform]: shinohai something's borked
crtdaydreams[asciilifeform]: shinohai tried it, we'll have to rekey
shinohai[asciilifeform]: Try to %unpeer shinohai then add me back. My AT is 67.205.140.207:7778
awt[asciilifeform]: Will try to get a patch out for the unpeer bug soon.
shinohai[asciilifeform]: iirc I patched this in previous version on bench, just forgot to submit here for peer review - and it didn't make it to latest press.
jonsykkel[asciilifeform]: maked new peerig for u shinohai
shinohai[asciilifeform]: tyvm jonsykkel
shinohai[asciilifeform]: +1 jonsykkel prefixing with cmd syntax so I could lazily pipe it into weechat fifo. xD
jonsykkel[asciilifeform]: i got u covered
awt[asciilifeform]: shinohai: if you can find that patch and paste it it would be helpful. I can't repro the unpeer bug in my dev environment.
shinohai[asciilifeform]: awt: will do but who knows this is only time it happened in recent memory. Station fired right back up once I removed his key from db.
whaack[asciilifeform]: howdy pesnet
awt[asciilifeform]: afternoon whaack
awt[asciilifeform]: lol I've done that a few times
shinohai[asciilifeform]: $ticker btc usd
busybot[asciilifeform]: Current BTC price in USD: $21038.47
awt[asciilifeform]: jonsykkel: this seems a little weird: http://paste.deedbot.org/?id=uCG_
jonsykkel[asciilifeform]: awt: lol ye my prog by default getdatas evry 10sec on borken chains, peered crtdaydreams and chain broke imediately
jonsykkel[asciilifeform]: i guess cuz dosnt recognize chad messages as crtdaydreams and that why dosnt resolve break or smth
jonsykkel[asciilifeform]: increased that 10sec shit anyway
awt[asciilifeform]: Testing 9975
awt[asciilifeform]: pestbot: ping
awt[asciilifeform]: That said, obviously something weird going on in this case.
awt[asciilifeform]: pestbot: ping
awt[asciilifeform]: In that case, "chad" or crtdaydreams is only connected to asciilifeform, as are you.
awt[asciilifeform]: Doesn't explain anything, just trying to figure out what might be happening.
awt[asciilifeform]: Could very well be a bug in the re-broadcasting code that isn't visible to other blatta users that dedup incoming messages.
awt[asciilifeform]: oh shit, yeah you are lol
awt[asciilifeform]: pestbot: ping
awt[asciilifeform]: 9975 fixes %unpeer, duplicate handles while syncing, dupe GETDATA requests, and nonce generation.
signpost[asciilifeform]: cool updated. ty awt
crtdaydreams[asciilifeform]: jonsykkel: probably a good idea not to run two pest instances in the same dir, as I had two of the same key in ./data/peer/
phf[asciilifeform]: so just to clarify, if i lose the connection, then eventually reconnect, there's no way for me to immediately know "what i've missed", in fact i have to slowly rebuild chain as respective parties speak?
awt[asciilifeform]: phf: PROD prevents that scenario
awt[asciilifeform]: smalpest has PROD, blatta doesn't have it yet
awt[asciilifeform]: Also - found an issue with 9975 - don't update yet.
signpost[asciilifeform]: since it's already in the log might want to just issue an errata vpatch.
phf[asciilifeform]: i assume relevant conversation re prod is in the logs, but i can't figure from the spec, how it supposed to work. does connecting station issue prod? halp
awt[asciilifeform]: phf: haven't reviewed but I thought PROD was to be spammed to all peers periodically.
awt[asciilifeform]: there was even discussion of replacing IGNORE with PROD
phf[asciilifeform]: i guess question to asciilifeform, why not let client query for data? it seems like an intentional thing, that the data comes at the counterparty's whim, i'm curious what the reasoning behind this is
phf[asciilifeform]: e.g. getdatasince hash256 ->gives you everything that came in since hash
awt[asciilifeform]: Testing 9974. Gonna let it run a while before linking, although it's up if anyone wants to try.
jonsykkel[asciilifeform]: crtdaydreams: ye that would be bad idea. u can run multi instances by overriding data dir with -d opt
jonsykkel[asciilifeform]: awt: i not has PROD yet either
asciilifeform was out pulling cable, back for short spell
asciilifeform: http://logs.nosuchlabs.com/log/pest/2022-07-10#1009268 << the reasoning is that if you dun have complete chain whenever at all possible, you can't 1) serve it up to folx who need to fill own gaps 2) detect 'forks'
dulapbot: Logged on 2022-07-10 20:54:39 phf[asciilifeform]: i guess question to asciilifeform, why not let client query for data? it seems like an intentional thing, that the data comes at the counterparty's whim, i'm curious what the reasoning behind this is
asciilifeform: grr blatta kaput again. erry single time dulapbot speaks..
asciilifeform: maybe misunderstood the q tho. is phf's notion that ~all~ msgs oughta be 'pulled' rather than pushed? why on earth wouldja do that ?
asciilifeform: if i'm talking, i want delivery to all live peers asap.
dulapbot: Logged on 2022-07-10 20:19:02 phf[asciilifeform]: so just to clarify, if i lose the connection, then eventually reconnect, there's no way for me to immediately know "what i've missed", in fact i have to slowly rebuild chain as respective parties speak?
crtdaydreams[asciilifeform]: awt what would be your wot handle?
awt[asciilifeform]: asciilifeform: maybe send me a key for dulapbot and we can see if it crashes my station as well.
asciilifeform: phf: a proper pestron will blast 'prods' on warmup and at certain freq. subsequently
crtdaydreams[asciilifeform]: lol I knew I was missing something :D
asciilifeform must bbl
awt[asciilifeform]: crtdaydreams: thimbronion
awt[asciilifeform]: ty crtdaydreams will add
awt[asciilifeform]: crtdaydreams: if you add a key for my handle you should get my address automatically
awt[asciilifeform]: crtdaydreams: anyway not getting any packets from you
crtdaydreams[asciilifeform]: see it's doing what happened with shinohai again
crtdaydreams[asciilifeform]: you're broadcasting getdata somehow yet I still see you has hearsay
crtdaydreams[asciilifeform]: hola signpost o/
crtdaydreams[asciilifeform]: you don't come up as a direct peer either, my AT hasn't been updated with your address
awt[asciilifeform]: hrm jonsykkel shouldn't smalpest update the AT on receipt of IGNORE packets?
crtdaydreams[asciilifeform]: fwiw when I was testing on my own machine the AT updating worked fine
signpost[asciilifeform]: seems like my station might be crash looping atm.
signpost[asciilifeform]: works atm, watching the log.
signpost[asciilifeform]: crtdaydreams: you may have a NAT issue on your end if you're not receiving my packets.
signpost[asciilifeform]: I'm receiving yours, and others are receiving mine.
awt[asciilifeform]: signpost: yeah you need 9974
crtdaydreams[asciilifeform]: signpost, no that's not you, that's thimbronion that's hearsay, we just peered
awt[asciilifeform]: from my log: [10.210.138.157:7778 crtdaydreams] <- BROADCAST hrm jonsykkel shouldn't smalpest...
crtdaydreams[asciilifeform]: awt try %unpeer crtdaydreams and see if you crash
← 2022-07-09 | 2022-07-11 → ↑
Random(pest) | Download hourly DB snapshot | Get Source Code