Show Idle (>14 d.) Chans
← 2022-07-16 | 2022-07-18 → ↓
crtdaydreams[asciilifeform] is still unsure what is going on with AT, whaack, billymg, et al. are still hearsay, blatta 9974 didn't patch, I didn't find the time to look at it last week, but if anyone has any ideas, perhaps someone willing to try a re-peer?
dulapbot: Logged on 2022-07-16 14:05:03 whaack[asciilifeform]: http://logs.bitdash.io/pest/2022-07-13#1009379 << i really enjoy this genre and specifically this mix you posted, plz post more
crtdaydreams[asciilifeform]: signpost: holy moly, the new wot visualizer is beautiful
shinohai[asciilifeform]: http://logs.bitdash.io/pest/2022-07-17#1009859 << I think I posted in logs here somewhere that this was head-scratcher for me at first but "makes sense" you will occasionally see messages as hearsay from others you are peered with.
bitbot[asciilifeform]: Logged on 2022-07-17 00:03:30 crtdaydreams[signpost]: is still unsure what is going on with AT, whaack, billymg, et al. are still hearsay, blatta 9974 didn't patch, I didn't find the time to look at it last week, but if anyone has any ideas, perhaps someone willing to try a re-peer?
awt[asciilifeform]: crtdaydreams: try renaming the sig file extensions for 9975, 9974 from .sig to .thimbronion.sig
awt[asciilifeform]: crtdaydreams: add this to your AT: http://paste.deedbot.org/?id=M8it
busybot[asciilifeform]: The 24-Hour VWAP for BTC is $ 21154.20 USD
busybot[asciilifeform|asciilifeform]: Current BTC price in USD: $21164.7
billymg[asciilifeform]: i generated a bunch of peering keys for bitbot and crawler bot, sharing them for anyone who would like to peer with the bots
billymg[asciilifeform]: i don't have IPs for phf, dulapbot, busyboy, and trbexplorer. so if you'd like to peer with one of those please send me the IP
billymg[asciilifeform]: i'd also like to peer billymg with the bots in here but i plan to move that station soon and rekey for all existing peers anyway so i didn't include peering keys for that handle yet
billymg[asciilifeform]: and also for phf, since we're not peered yet
phf[asciilifeform]: billymg: my multipeer is not implemented yet. i'll reach out to everyone whose keys i don't have as soon as it's operational. i've mostly just didn't have time to hack…
billymg[asciilifeform]: phf: ah, got it, sounds good
whaack[asciilifeform]: !e view-height
trbexplorer[asciilifeform]: block_height: 745399
trbexplorer[asciilifeform]: mins_since_last_block: 13
awt[asciilifeform]: billymg: successfully peered with your bots thx
bitbot[asciilifeform]: Logged on 2022-07-17 15:32:51 billymg: asciilifeform: http://paste.deedbot.org/?id=uKlq
asciilifeform still wishes there were a way to issue % cmds to a bot w/out unplugging its irc end
billymg[asciilifeform]: awt: indeed showing you as directly peered with bitbot now: http://logs.bitdash.io/pest/2022-07-17#1009895
bitbot[asciilifeform]: Logged on 2022-07-17 18:01:58 awt: billymg: successfully peered with your bots thx
billymg[asciilifeform]: asciilifeform: for some reason you're now showing up in the logs as hearsay: http://logs.bitdash.io/pest/2022-07-17#1009898
bitbot[asciilifeform]: Logged on 2022-07-17 20:42:33 asciilifeform[5]: still wishes there were a way to issue % cmds to a bot w/out unplugging its irc end
bitbot[asciilifeform]: Logged on 2022-07-02 19:30:38 asciilifeform: lol
bitbot[asciilifeform]: Logged on 2022-07-04 11:18:10 asciilifeform[billymg|signpost]: billymg: asciilifeform only hosted with ukrs once before; your mileage may vary
signpost[asciilifeform]: thanks billymg! I've added those peerings.
phf[asciilifeform]: i suspect if one's keeping a set of "canonical" packets, one ought to replace slowly replace ones if higher bounce with lower bounce, until one gets a 0 bounce, which implies direct, authorative communication
phf[asciilifeform]: *replace ones with higher bounce
phf[asciilifeform]: but i suspect blatta doesn't have a mechanism like that, or even if it does the irc nature of the irc compatible version of protocol means that packet traveling faster/shorter path will win, as far as representative packet
phf[asciilifeform]: so if you have a faster packet path that goes through intermediaries, than direct from originating station, you'll end up with "hearsay" packets for things that you might've gotten direct few us later
asciilifeform: phf: see 4.4.4.2.2 re how this (is supposed to) work
asciilifeform: err, 4.3.3.2.2
asciilifeform: (link is correct)
asciilifeform: tldr, hearsays are interned to give 1) hearsays with lower 'bounce' 2) immediate copies -- a chance to be received; then 'canonical' msg is either the subset of hearsay copy(s) with min. bounce, or the immediate (if the latter was received during the embargo interval)
phf[asciilifeform]: well, then
asciilifeform: on top of this, nuffin (per spec) is bounced to any place from where was already received; nor is anyffin but the min-bounce copy relayed
asciilifeform: in theory, gives 'least surprising' propagation behaviour.
asciilifeform: ... whether blatta 100% behaves as described in link, asciilifeform presently cannot say, still in the process of making sense of the src
phf[asciilifeform]: do you know what Te values is in blatta? it's essentially a lower bound on hello-yes-hello roundtrip
asciilifeform not recalls; q for awt ^ when he wakes up
phf[asciilifeform]: also am i reading this right, one doesn't really see a message earlier than Te seconds after receipt?
asciilifeform: phf: is only round-trip lower bound on hearsays tho
asciilifeform: phf: a hearsay msg, strictly
asciilifeform: immediates and directs are considered prima facie authentic and forwarded to console (per spec) right away
asciilifeform: (there's 0 reason to intern an immediate or direct msg)
phf[asciilifeform]: right, but also thread started because i was considering unequal packet travel times
asciilifeform: it aint uncommon for a hearsay copy to arrive coupla 100 msec prior to original (immediate) copy
asciilifeform: hence the internment mechanism
asciilifeform: receiver does not know when (or if) original will arrive, hence after Te elapses, will treat msg as hearsay
asciilifeform: in principle, in a dedicated client, oughta retro the db if receive immed. copy at some future time
phf[asciilifeform]: also kind of neat because in non-spec compliant non-irc version Te is only relevant for retransmit. can e.g. render hearsay right away, mark it direct once that information is available…
asciilifeform: aha, precisely
asciilifeform: irc oughta be an annex to spec, rather than in spec proper, iirc already mentioned this notion
asciilifeform: phf's thing afaik is the 1st 'general purpose' client w/out ircism
asciilifeform: good % of gnarl in current spec comes outta the constraint where db is considered 'write once'
asciilifeform: (because console 'is teletype')
asciilifeform: tbf, would be surprised if with any regularity the immediate in fact dun make it to receiver within a reasonable (e.g. 1s) Te
asciilifeform: (while hearsaid copies do)
asciilifeform: (in certain cases, elementarily, none of'em make it, in e.g. the case of a transmitter whose cable had been uplugged while he was sending)
phf[asciilifeform]: proper pest of course ought to keep a running p on counterparty's avg roundtrip by running stats on past packet information :>
asciilifeform: phf: not that it wouldn't be interesting, but not clear to asciilifeform what if anyffin machine oughta do with this stat other than to display it on req. of operator
asciilifeform must bbl
phf[asciilifeform] scratches head
bitbot[asciilifeform]: Logged on 2022-07-17 23:56:32 phf[awt]: foo^A
phf[asciilifeform] foo_^_A_” but without the underscore
phf[asciilifeform]: full message was “i should probably add /me… btw that's the case where irc spec full on leaks into pest: messages arrive as _^_A_ACT_ION foo_^_A_” but without the underscore
← 2022-07-16 | 2022-07-18 → ↑
Random(pest) | Download hourly DB snapshot | Get Source Code