Hide Idle (>14 d.) Chans
← 2022-07-11 | 2022-07-13 → ↓
vex: there's a little bit of lost highway there somewhere
vex: lynchesque fibs and laura palmer. this is cinematic tv
vex: i'll need to consult shinohai before I read rolling stone, or the other fishwrap. communicatiobns degree indeed
vex: stupis usually misses the point
vex: spoiler alert, you've had enough time
vex: saul e608 is poised with potential, truth be told, it's tense as fuck
vex: It might be happenstance that it lands the same day that CZ takes .fr really who gives a shit
vex: the happenings are in a german built bunker. who will escape?
vex: do we like saul less? what is he up to? do we like kim more? would she do it?
vex: from my perspective, it's a resounding yay on both counts
vex: Saul gave Kim her freedom, but he knew she would. It's heavy shit
vex: they didn't have time to talk, all she knew was "go blast a cunt"
vex: we know saul survives.
vex: kim wexler is hot. she goes to blast
vex: can she get any hotter?
vex: actress is shuddering prcisely
vex: how I imagine hanbot watched mp drown
vex: pure fear
vex: I hope she's okay, that kinda shit will fuck a person up
vex: for a widow with a well found five series and recent fake tits, she's prolly doing okay
vex: ie clevererer than most
vex: nope. prolly still crying. my heart hurts
vex: it's too much crying for a smol girl. I can't do anyhting to help
vex: unless you wanna be a bottle blonde. buy a case of vintage bollinger.
vex: I tan my hides with once drunk bollinger
vex: it's the ammonia in it that does it
vex: well. victoria bitter
vex: I don't get meat often, but I save the skins
vex: I'm pretty much vegan
vex: If I accidentially a roo, I take the skin, not plop the whole beast on a fire
vex: rootail soup is best without fur
vex: vegan chicks love it
vex: did you really accident the beast vex? yep no guns
vex: I guess it's okay hten
vex: r u hungry? yes? it's okay
vex: now, go gring the useless bits to keep for salami
vex: just poke it in there and crank the handle
vex: why is the girinder bloted to such a low bench vex?
vex: fairytales is all
vex: I'm just trying to lick your butthole
vex: basically vegan
vex: eek. racoons
vex: cdd, I was going to leave you a geocache, but brisbane is really far
vex: I lived there for a time, I'd wake up and stroll the city after a special breakfast muffin, quite convinced it was gotham
vex: i'm not sure how well you know the place, but if you care to stroll over the brigde, I'll tell you where I would have left it
vex: charlotte st, the second hand bookshop
vex: what a library. towering shelves invite you to walk whispering, lest the old groaning floorboard fold you into a mess
vex: evrything is for sale
vex: I've spent hours in there reading, scared to move
vex: eek. why are there even racoon here?
asciilifeform: lolwat.
cgra: asciilifeform: did you still have this idea that separate pest nets ought to be able to join?
cgra: asciilifeform: another q, what was the 'long buffer' for again? why not just always check the 'very long buffer', (tm)(r) jonsykkel?
asciilifeform: wb cgra
asciilifeform: cgra: indeed separate pestnets can be joined if at least 1 peer from each form a mutual peering.
asciilifeform: cgra: 'long buffer' is simply the subset of the 'very long buffer' that is guaranteed to live in memory and can be used for fast deduping.
dulapbot: Logged on 2022-01-20 13:59:15 asciilifeform: if an arbitrary packet can force you to access disk, yer sunk
asciilifeform: if a message was received at some point, but aint stale yet, it absolutely must be kept in memory, for deduping.
asciilifeform: per current spec, for 1h (to account for whatever degree of clock skew, plus healthy margin)
cgra: asciilifeform: in replay scenario, doesn't the 15 min timestamp margin limit the disk access, so that the queried entries are ~always in cache, and no actual disk access caused?
cgra: asciilifeform: if two pest nets join, they're going to cross-sync their histories and end up having two separate branches. if somebody was part of both nets, but separately, he's now going to have two 'genesis messages', one from each net. correct?
cgra: another, related consideration: doesn't getdata as now specced, allow retroactive selfchain fork? feed new branch backwards, starting from the last of the chain, which meets the 'not-stale' requirement
cgra: now, if selfchain forks are allowed, prod seems to have a problem: how to sync the possible loose ends? should apply to the joint nets case too
asciilifeform: cgra: the selfchain thing resolves to the 'bob problem' ( i.e. handle collision. still unresolved, tho asciilifeform proposed a hypothetical pill . )
dulapbot: Logged on 2022-06-06 14:07:36 asciilifeform: phf: and when 2+ nets merge, and this produces a handle collison ? (aka the 'bob' problem)
asciilifeform: cgra: netchains will converge, however, when the participants of the joint net finally see same thing
asciilifeform: cgra: when 2+ nets merge, each will see the history of the other as a long thread (in hypothetical thread-view pestron) in the past
asciilifeform: 1 tricky moment is that they'll have conflicting 'genesis' msg.
asciilifeform: (where netchain == 0)
asciilifeform: cgra thimbronion et al ^ does need thought
asciilifeform: 1 'solution' is to simply prohibit getdata(0).
asciilifeform: unsatisfying tho.
asciilifeform: tho come to think of it already prohibited (there aint a msg which hashes to 0)
asciilifeform: so actually scratch that, this problem dun actually exist.
asciilifeform: cgra: if asciilifeform misses sumthing, plox to point out.
whaack: !w poll
watchglass: Polling 15 nodes...
watchglass: 54.39.156.171:8333 : (ns562940.ip-54-39-156.net) Alive: (0.111s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=744698
watchglass: 205.134.172.4:8333 : (172-4.core.ai.net) Alive: (0.143s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=744698
watchglass: 205.134.172.26:8333 : Alive: (0.141s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=744698
watchglass: 205.134.172.28:8333 : Alive: (0.084s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=744698 (Operator: whaack)
watchglass: 205.134.172.27:8333 : Alive: (0.145s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=744698 (Operator: asciilifeform)
watchglass: 208.94.240.42:8333 : Alive: (0.209s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=744698
watchglass: 54.38.94.63:8333 : (ns3140226.ip-54-38-94.eu) Alive: (0.263s) V=88888 (/therealbitcoin.org:0.8.88.88/) Jumpers=0x1 (TRB-Compat.) Blocks=744698
watchglass: 94.176.238.102:8333 : (2ppf.s.time4vps.cloud) Alive: (0.311s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=744698
watchglass: 82.79.58.192:8333 : (static-82-79-58-192.rdsnet.ro) Alive: (0.391s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=744698
watchglass: 103.6.212.28:8333 : Alive: (0.546s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=744668 (Operator: whaack)
watchglass: 103.36.92.112:8333 : (terebe.ns01.net) Alive: (0.572s) V=99999 (/jwrd.net:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=744698
watchglass: 75.106.222.93:8333 : Could not connect!
watchglass: 71.191.220.241:8333 : Busy? (No answer in 100 sec.) (Operator: asciilifeform)
watchglass: 143.202.160.10:8333 : Busy? (No answer in 100 sec.)
watchglass: 205.134.172.6:8333 : Busy? (No answer in 100 sec.)
cgra: asciilifeform: any comment re prod and loose chain ends?
cgra: asciilifeform: if multiple genesis messages are allowed, then prolly otherwise retroactively injected selfchains must be allowed too?
dulapbot: Logged on 2022-07-12 11:01:35 asciilifeform: so actually scratch that, this problem dun actually exist.
dulapbot: Logged on 2022-07-12 10:38:05 cgra: another, related consideration: doesn't getdata as now specced, allow retroactive selfchain fork? feed new branch backwards, starting from the last of the chain, which meets the 'not-stale' requirement
cgra: and not just genesis messages, any retroactive forks like ^
asciilifeform: cgra: hypothetically yes, 'prod' can contain arbitrary values for the chains. per current spec the recipient will getdata for'em and store whatever it gets.
asciilifeform: for netchain , you want to be able to do this . for selfchain, not, but to properly prohibit it would have to solve the handle uniqueness problem, 'which prong of fork is genuine' resolves to it
cgra: asciilifeform: in other words, is this correct: 1) netchain forks are desirable, but selfchain forks not. 2) multiple selfchain genesises desirable only if for a good reason (like joint nets). 3) while selfchain forks and excess genesises not desirable, difficult to prevent, so perhaps better to just tolerate
asciilifeform: cgra: correct
asciilifeform: btw per current spec, no effort to 'prevent' selfchain forks, only to mark'em
asciilifeform: (they will inevitably occur, if only when sumbody blows a disk and sets up station from scratch)
cgra: asciilifeform: any comment re 'long buffer portion being cached anyway'?
dulapbot: Logged on 2022-07-12 10:26:06 cgra: asciilifeform: in replay scenario, doesn't the 15 min timestamp margin limit the disk access, so that the queried entries are ~always in cache, and no actual disk access caused?
asciilifeform: cgra: can't rely on external caches.
asciilifeform: there's no guarantee that past hr will in fact be entirely in such a cache (or that a cache even exists)
asciilifeform: hence mandated to exist in pestron.
asciilifeform: because incoming packet (which may be a replayed dupe by attacker) MAY NOT result in mandatory disk access.
asciilifeform: will also point out, pestron is specced in such a way as to make sense for implementation without os.
asciilifeform: (or even in iron)
asciilifeform: spec aims to describe the logic without reference to complex external moving parts (disk caches, etc) which may not be standardized or even exist on a given machine.
cgra: ok, makes sense
cgra: imo the long buffer desc could use similar explanation
asciilifeform will add note in next rev
asciilifeform encourages cgra et al to write in whenever 'hmm wtf is this, wai' in spec
cgra: asciilifeform: how's your 'hevea' rewrite getting along?
asciilifeform: cgra: veery slowly
cgra: asciilifeform: aiming for cool diagrams and all?
asciilifeform tightly in the grip of meatspace chores atm
asciilifeform: cgra: eventually yes
asciilifeform: cgra: tried to port 'bytefield' to heveaism, 1st try broke teeth
asciilifeform: ( yes, ~could~ shit'em to bitmaps, but then won't have clickable links in the packet diagrams ! )
asciilifeform: soo arguably hevea dun actually solve the 'hey asciilifeform , want a civilized spec? go and fight with 10,000 lol of liquishit for next yr' )
asciilifeform: ( 'civilized spec' non-negotiably means, ftr, 1) fully hypertext for www 2) fully formatted printer-friendly pdf/ps for printer 3) NO separate maintenance for 1 and 2, generated from ~one~ src )
cgra: right
cgra: asciilifeform: if you go by the current spec draft letter, doesn't the order buffer necessarily always release the newly synced chain segment backwards, from newest to oldest? obv though, can't help out-of-orderism if sync takes longer than 5min, the order buffer timeout
asciilifeform: cgra: not obviously so: 'At any time when the Order Buffer is non-empty, the station will schedule the continued processing of the oldest packet found there, to be executed Tw seconds following the Time at which said message was received.'
asciilifeform: ... and when it connects to the known chain, whole thing gets shot out
asciilifeform: oughta specify the eviction order tho
cgra: ah, ^ mention prolly missing
asciilifeform: that whole section needs serious massage.
cgra: currently afaik only said way out of order buffer is the timeout
asciilifeform wonders whether thimbronion implemented this literally as-written, hence the wild delays on dulapbot's end
cgra: side note: upon a handful of first spec reads, i found 'long buffer' and 'short buffer' didn't stick as well as perhaps smth like 'dedup buffer' or 'embargo buffer' would've
cgra: similarly all the two-letter symbols. perhaps helped with added explanation, or symbol length, or increased consistency. for consistency example, the "T" in "Tw" means point in time, while "T" in "Te" means duration. maybe "Dt" or "It" instead (I as in 'interval')?
asciilifeform: these need an index, too
asciilifeform given up on the 'markdown' turd, is impossible to fucking change ANYTHING without rewriting half the refs
cgra: obviously a rough draft still, for N next revisions, and understandably
asciilifeform: in current spec, order of sections is also riotously nonsensical
asciilifeform: again because impossible to move anyffin without rewriting, tediously, over9000 links
asciilifeform: and section #s
cgra: right
thimbronion: asciilifeform: did my best to implement the spec literally as written, but, perhaps you might find this interesting: http://logs.bitdash.io/pest/2022-07-11#1009305
bitbot: (pest) 2022-07-11 awt[billymg|signpost]: for bots, setting order_buffer_check_seconds and order_buffer_expiration_seconds to like 10 should speed things up quite a bit.
asciilifeform: thimbronion: aha, but cgra did find laughable omission in spec ( asciilifeform didn't actually remember to say 'and when the chain links up, whatever in the buffer that linked up with it is evicted in order' )
signpost just about to test thimbronion's suggestion on deedbot btw
thimbronion: asciilifeform: ah gotcha. Yeah in blatta items are evicted by timestamp.
asciilifeform: cgra: thinking about it, there's nuffin particularly unpleasant re having 2+ netchain genesis. but imho in a dedicated client they oughta be displayed chronologically (by timestamp).
asciilifeform: (and in fact any msg with netchain==0 oughta be displayed where its timestamp places it.)
asciilifeform: hypothetically this means that a member of a pestnet could 'litter' the past. easily visible to naked eye tho, when it happens, and you can whack him over the head if need be.
asciilifeform: (and the 'past' he can litter is -15min max, lol)
asciilifeform: ^ thinking further, potentially in -inf, if via chain of getdata
cgra: asciilifeform: yea was just about to point out ^
asciilifeform: at any rate visible to participants if happens
asciilifeform: this kinda thing is properly categorized as spamming
cgra: only now noticed you said 'netchain' there, was still in 'selfchain' thinking mode
asciilifeform: correct
asciilifeform: a retro-fork of selfchain oughta be handled in precisely same way as a realtime fork of same (currently -- by marking)
asciilifeform: if the participants with colliding nicks aint trying to 'misbehave', this actually makes ~easier~ merging 2+ nets with a colliding nick
cgra: is there any good reason for selfchain fork?
asciilifeform: (the main headache there is when someone is deliberately impersonating)
cgra: (and not counting multiple genesis messages as fork)
asciilifeform: cgra: the most common cases, asciilifeform expects, would be 1) someone loses his state 2) merge of nets with colliding handles
asciilifeform: a then (2)
cgra: asciilifeform: you mean (1)?
asciilifeform: and hm this does need to be illuminated in spec ( the fact that 2 parallel chains with colliding handle also constitutes a fork )
asciilifeform: cgra: (1) and (2) both will give multiple genesis
cgra: asciilifeform: but (1) only good explanation for selfchain fork?
asciilifeform: depends what means 'good explanation'
cgra: asciilifeform: no funny business
asciilifeform: what meant above is that 2 commonplace situations not involving deliberate sabotage will give something other than an intact chain with unique handle
asciilifeform: (not including temporary gaps from packet loss here)
asciilifeform: *a single intact chain
asciilifeform: btw asciilifeform had notion, where in a dedicated 'gui' pestron, could display 'hash icons' which include chain info in the hash and give visible indicator of chain breakage/split
asciilifeform: (a la wp's auto-genned username+email icons)
cgra: i'm trying to isolate a branching selfchain as a condemnable situation vs >1 parallel chains
asciilifeform: a makes sense
asciilifeform: branching means that you have a wrecker. (or a fool who config'd two separate stations to use one key)
cgra: and branch is reliably detectable
asciilifeform: correct
thimbronion: I wonder if the more likely event will not be merges but splits.
asciilifeform: thimbronion: likely; but splits dun pose any interesting algorithmic problems afaik.
asciilifeform: ( unless/until the split parties, or some subset thereof, reunite )
← 2022-07-11 | 2022-07-13 → ↑
Random(asciilifeform) | Download hourly DB snapshot | Get Source Code