Forum logs for 06 May 2018

Monday, 16 March, Year 12 d.Tr. | Author:
mother[m]: mircea_popsecu definitely not ultrasound machinery. Two ways. One is you have drones on surveillance flights. What it does provide you is a lot of high-res overlapping imagery you can generate aerial footage and 3d models of. You can combine this with other seismic and geotech surveys taken. Another way is that you have drones able to retrieve samples on-site, leaving a surveyors role more on analysis. The cost reductions [02:01]
mother[m]: are substantial. [02:01]
ben_vulpes: what does the sample-taking equipment look like? [02:03]
ben_vulpes: what's the sampler-system design, i mean [02:04]
mother[m]: ben_vulpes haven't seen it myself yet so can't say. [02:30]
mother[m]: With Nadex, you're trading binary options and it's at market-set prices. Which is more than what I can say for quotes from Deribit for example. I agree on that about the bitcoin price signal. Do you see BTC trading to or past its highs this year? [02:37]
mircea_popescu: mother[m], but they'd get very superficial samples neh ? [02:41]
* mircea_popescu doesn't know much about mineral survey, but oil at least, assays are all >100m. [02:42]
mircea_popescu: anyway, bitcoin's natural value is somewhere in between 150% and 300% or so of the M3. so yes, i see bitcoin price exceeding the aggregate cash value of most countries. [02:44]
mircea_popescu: 2.5 haitis to the btc sorta thing. [02:45]
mircea_popescu: Moduli Broken: 2489 << asciilifeform yeah totally, this did more in a few days than the old one did per month or two. [02:56]
mircea_popescu: http://phuctor.nosuchlabs.com/factor/380 << this incidentally is a great addition. [03:22]
mircea_popescu: asciilifeform, my question to you is : if there's "Factors Shared by Two or More Moduli: 259" reported on the stats page, yet the factor index goes to 419 (ie, I can see http://phuctor.nosuchlabs.com/factor/419 ) then what the hell's going on ? are there 259 or 419 factors ? [03:24]
mircea_popescu: (in other lulz, google fully indexes it, like say https://www.google.com/search?q=185.68.228.74 or w/e random ip from the reported list). [03:25]
mircea_popescu: !!up fromdeedbot [03:31]
deedbot: fromdeedbot voiced for 30 minutes. [03:31]
fromdeedbot: hey guys so I did register my RSA key to deedbot as "fromdeedbot" but i got it working. I'm learning a lot of things are new to me right now so please xcuse me ahead of time if i do something in bad from, it's not on purpose [03:34]
lobbesbot: fromdeedbot: Sent 1 day, 2 hours, and 41 minutes ago: <mircea_popescu> stop join/parting. [03:34]
fromdeedbot: if i leave my irc window open it should just stay logged correct? [03:34]
mircea_popescu: mkay. don't do the join/part thing, it pisses people off. [03:34]
fromdeedbot: copy that [03:34]
mircea_popescu: fromdeedbot, not reliably. why not get a client / bouncer set up ? [03:34]
fromdeedbot: I'll google that [03:35]
fromdeedbot: what do you guys think of apps being built on BCH? [03:36]
fromdeedbot: if anything [03:36]
mircea_popescu: otherwise, if your internet access is primarily via web, the channel's logged the logs are more reliable than your ad-hoc wwwapp will be. see http://btcbase.org/log http://logs.bvulpes.com/trilema http://log.mkj.lt/trilema/today [03:36]
fromdeedbot: ty [03:36]
fromdeedbot: I started reding the logs here in NOV. [03:37]
fromdeedbot: impressive stuff... [03:38]
mircea_popescu: so then you know nobody took the bitcoin crash lulz seriously. [03:38]
fromdeedbot: bitcoin "crash" or bitcoin crashing, or bot [03:39]
fromdeedbot: *both [03:39]
mircea_popescu: "bch". that thing where some busker in san francisco declared himself "emperor of america". [03:41]
mircea_popescu: except with handpuppets and blockchains. but otherwise, same thing. [03:42]
fromdeedbot: yeah too many forks not enough steak [03:42]
fromdeedbot: most "tokenized" use cases are not interested in building their own chain [03:43]
mircea_popescu: the only stable solution is when 50%+1 of all energy production goes to mining there's absolutely no space for "alternate" coins other than an expensive luxury early on. [03:56]
mircea_popescu: meanwhile in ATM news, https://78.media.tumblr.com/129da98300ef78f827d37edd8e89e4c9/tumblr_p1ycl0m2gU1vowk7zo1_r1_400.gif [04:09]
asciilifeform: http://trilema.com/forum-logs-for-06-may-2018#2434991 << there are two types of factor that are factors of ~solely one~ modulus: 1) debianistic ( and other 'special collection' ones, i'ma make it mark them in the factor page soon) , and, [10:00]
a111: Logged on 2018-05-06 07:24 mircea_popescu: asciilifeform, my question to you is : if there's "Factors Shared by Two or More Moduli: 259" reported on the stats page, yet the factor index goes to 419 (ie, I can see http://phuctor.nosuchlabs.com/factor/419 ) then what the hell's going on ? are there 259 or 419 factors ? [10:00]
asciilifeform: 2) if mod / oneknownfactor is prime, we get the other kind, also a just-that-modulus factor [10:01]
asciilifeform: i'ma need to put this in faq, really [10:01]
asciilifeform: http://btcbase.org/log/2018-05-06#1810404 << 2818 currently [10:02]
a111: Logged on 2018-05-06 06:56 mircea_popescu: Moduli Broken: 2489 << asciilifeform yeah totally, this did more in a few days than the old one did per month or two. [10:02]
asciilifeform: http://btcbase.org/log/2018-05-06#1810407 << it indexes it in ~realtime , it plus a whole buncha crawlers that have namestrings that dun correspond to anything findable publicly, i suspect the usual suspects, when ipbanned they come back ~immediately from somewhere else [10:04]
a111: Logged on 2018-05-06 07:25 mircea_popescu: (in other lulz, google fully indexes it, like say https://www.google.com/search?q=185.68.228.74 or w/e random ip from the reported list). [10:04]
asciilifeform: this started in 2015 actually [10:04]
asciilifeform: difference is, now it does not grind to a halt, from this treatment, O(n log n) fetches. [10:05]
asciilifeform: http://btcbase.org/log/2018-05-06#1810398 << ahaha, i did predict, >> http://btcbase.org/log/2018-05-06#1810332 [10:10]
a111: Logged on 2018-05-06 06:30 mother[m]: ben_vulpes haven't seen it myself yet so can't say. [10:10]
a111: Logged on 2018-05-06 00:12 asciilifeform: http://btcbase.org/log/2018-05-05#1810271 << i suspect that there is no physical operation involved, it is 100% chumpatronics, like all other subjects of usg.startupism at this point [10:10]
asciilifeform: aaand mods still popping... [10:29]
mircea_popescu: asciilifeform, you don't take my meaning. there can't be at the same time the case that a) "the cardinal of the set of factors is 259" and b) "419 in http://phuctor.nosuchlabs.com/factor/419 denotes the 419th factor". because if you have a 419th factor, howsoever serialized, the cardinal of the set of factors must be at least 419. [11:55]
asciilifeform: mircea_popescu: there in fact are currently 5182 known factors. [11:55]
mircea_popescu: and so the status page is way behind ? [11:55]
asciilifeform: most of them however are factors of only 1 modulus. [11:55]
asciilifeform: and are not reported under 'Factors Shared by Two or More Moduli' stat. [11:56]
mircea_popescu: oooo i see! [11:56]
asciilifeform: ( as to how we break a modulus without it sharing any factors with other known moduli: see logs re debianization ) [11:56]
asciilifeform: aha. [11:56]
mircea_popescu: right you are i was missing that particular edge of braindamage from my mental model. [11:56]
asciilifeform: good q tho. i was waiting for somebody to ask. [11:56]
mircea_popescu: maybe time to add "factors of a single modulus" stat to stat page ? [11:56]
asciilifeform: it is in the conveyor, along with 'where from?' stat in 'factor/123' page (e.g. 'debian collection', 'cisco', etc) [11:57]
mircea_popescu: cool deal. [11:57]
mircea_popescu: you know, last time we relocated phuctor there were significant gains. this time, even more significant. [11:57]
asciilifeform: nao it's sorta realtime [11:58]
mircea_popescu: i basically hope we keep losing it. seems the more indermeddling, the stronger the republic. [11:58]
asciilifeform: lol [11:58]
asciilifeform: soooo summary of the eating of tail end of Framedragger collection: 877 popped mods. nearly matched prediction from earlier ( http://btcbase.org/log/2018-05-04#1808858 ) [12:32]
a111: Logged on 2018-05-04 13:05 asciilifeform: there are 1.7mil+ moduli in the queue right now if i fire the werker i expect that it will produce 8-900+ popped-moduli. [12:32]
mircea_popescu: wow check that out. [12:39]
asciilifeform: trinque may find interesting that deedbot registered , of these, : 702 [12:40]
asciilifeform: !#seen BenBE [12:42]
a111: 2017-04-09 <BenBE> But that's nothing left for today. [12:42]
mircea_popescu: asciilifeform, so what's your prediction re the github set ? how many items even in there ? [12:42]
asciilifeform: ^ d00d with large collection of debian-style 'famous p's and q's', even once showed up here and asked to get phuctor's, and he did, but somehow his collection includes ~whole keys~ rather than factors. soon i'ma feed in ~his~ collection. [12:43]
asciilifeform: 6976695 , in jurov's csv [12:43]
asciilifeform: i'd predict, but i do not recall how he obtained these. [12:44]
asciilifeform: ( and hey jurov, is it time for refresh ? or should i use the 2016 tarball as-is ) [12:44]
asciilifeform: !Q later tell danielpbarron http://phuctor.nosuchlabs.com/gpgkey/48EF022DF9242B52C3F2970BB62613855B089C60B99E8FCE5004CCF30D48D699 << any idea what this is ? [12:46]
lobbesbot: asciilifeform: The operation succeeded. [12:46]
mircea_popescu: asciilifeform, ah, so not even a mn ? alright. aaand what's your prediction ? [12:47]
asciilifeform: ^ it set off asciilifeform's heuristic bell re 'possible symptom of derps distributing ineptly mutilated lord key' [12:47]
mircea_popescu: asciilifeform, seems likely he signed it no ? [12:47]
* asciilifeform looks [12:50]
asciilifeform: loox like it. [12:50]
asciilifeform: ( this is unfortunately difficult to distinguish mechanically ) [12:50]
asciilifeform: mircea_popescu: i predict 0 . ( will be pleasantly surprised, perhaps.. ) [12:50]
asciilifeform: since the debian incident, enemy stepped up the 'NOBUS' crapola no noar '32768 possible keys, total', instead things moar in the spirit of http://qntra.net/2016/08/rng-whitening-bug-weakened-all-versions-of-gpg [12:51]
asciilifeform: ( where you apply a magictransform to the whole rfc4880 turd, to get a lattice and get the privs or at the very least, diddled rng that gives e.g. 48 bits of possible keyspace, so nobody finds straight collision, but their asic can walk it, or the like. [12:52]
mircea_popescu: hey, what other driver of progress, in republic as empire alike, can there even be. [12:52]
asciilifeform: sorta ~whole point of phuctor -- to increase the minimal complexity/cost for enemy . [12:54]
asciilifeform: let him do equiv of 'running in gas mask'. [12:54]
mircea_popescu: indeed. [12:55]
asciilifeform: and to whittle away at the nobus idiocy -- ideally we get ~regular whole-ipv4 scans, and erry time a usgtronic router vendor with fixed keys etc is discovered, list of ~live~ boxes is seen by whoever wants. [12:56]
mircea_popescu: doesn't seem avoidable. Framedragger_ came and went, maybe the next one and maybe even the next after that. but eventualy it's getting settled, like everything else. [12:57]
asciilifeform: it's a pretty straight mechanical job. [12:57]
mircea_popescu: great intro item too. [12:58]
asciilifeform: ( the correct place to do it ~from~, might even be the -- repurposed -- usg botnets, e.g. mikrotik etc. ) [12:58]
asciilifeform: and yes great work for n00bz. [12:58]
mircea_popescu: conceivably. [12:58]
BingoBoingo: I mean what else are those phree boxes there for [12:59]
asciilifeform: scanning is 'embarrassingly parallel', e.g. if you have 2 places to work from , it goes 2x as fast, and so on linearly [12:59]
mircea_popescu: in fact, the proper sources of tmsr hardware are, in order, a) confiscated usg hardware and b) its own iron outside the reich. that 2017 ended with 0% a and 0% b, and 2018 will likely end with 0%a and whatever%b has no bearing on this : there's always 2019, there's always uci, etcetera. [13:00]
asciilifeform: for certain uses, spoils iron is the ~only iron, aha [13:01]
mircea_popescu: for sexual uses. [13:01]
mircea_popescu: as a general rule, if it's satisfying, it's been taken. [13:01]
asciilifeform: oh, and to complete earlier snapshot picture : Known Moduli: 10347987 . [13:03]
asciilifeform: these, from 16010944 submissions ( i.e. gpg keys ) of which 12284842 were Framedragger-generated. [13:04]
mircea_popescu: asciilifeform, in nitpicks today : if phuctor.nosuchlabs.com then also fuckgoats.nosuchlabs.com neh ? [13:05]
asciilifeform: ( the rest, sks, and the tiny remainder, manual submissions. ) [13:05]
asciilifeform: mircea_popescu: nosuchlabs.com/hardware.html , and per mircea_popescu's 2016 spec , given as there is not yet a 2nd iron product, they are idempotent [13:05]
asciilifeform: ( will be revised once there is. ) [13:05]
asciilifeform: currently bare naked nosuchlabs.com shows FG pg. [13:06]
mircea_popescu: kk [13:07]
asciilifeform: incidentally, there's another yet-virginal but i suspect quite fertile field for phuctor : http://btcbase.org/log/2017-10-17#1725945 incident [13:14]
a111: Logged on 2017-10-17 05:59 jurov: "The flaw resides in the Infineon-developed RSA Library version v1.02.013, specifically within an algorithm it implements for RSA primes generation. " [13:14]
asciilifeform: ^ potentially good for some implausibly-large num of popped mods [13:14]
asciilifeform: includes ~all 'yubikey' devices, for instance. [13:14]
asciilifeform: and ~every 'tpm'-generated rsa modulus to date [13:15]
asciilifeform: ( recall? derps used prime-constructor instead of asciilifeform-style 'generate N random bits and probe for primality, if fail -- discard ALL of them' ) [13:15]
asciilifeform: !!up RagnarDanneskjol [13:15]
deedbot: RagnarDanneskjol voiced for 30 minutes. [13:15]
asciilifeform: !#seen RagnarDanneskjol [13:16]
a111: 2018-01-24 <RagnarDanneskjol> Fraudsters? Please clarify [13:16]
asciilifeform: upstack : http://btcbase.org/log/2018-05-06#1810476 << 6`976`695 , mircea_popescu [13:21]
a111: Logged on 2018-05-06 16:47 mircea_popescu: asciilifeform, ah, so not even a mn ? alright. aaand what's your prediction ? [13:21]
asciilifeform: or 5GB of keyola+usernames [13:22]
asciilifeform: pretty hefty collection. [13:22]
mircea_popescu: ah ah. and still you say 0 ? [13:24]
asciilifeform: hard to say 0, there's simply gotta be at least 1 or 2 debian victims [13:24]
asciilifeform: ... supposing these are ssh keys. i currently can't seem to recall what these are. lessee when jurov wakes up. [13:26]
BingoBoingo: Aite, the local pharmacies are starting to swiftly become a redeeming factor. One handwritten note with my symptoms and zero expensive ObamaDeathPanel visits is all it took to get Trimetoprim/Sulfametoxazol [13:36]
mircea_popescu: asciilifeform, tedious to get the pubkey out of shit like yubikey tho [13:40]
asciilifeform: mircea_popescu: the lusers get it out themselves and distribute, apparently ( the way it is used, other end of channel gotta have it ) [13:40]
mircea_popescu: asciilifeform, they were rsa keys off github iirc ? [13:40]
asciilifeform: right, but where were they generated/how, i do not recall [13:40]
mircea_popescu: asciilifeform, yes, but they only distribute it to hitler's minions. [13:41]
asciilifeform: ( did the lusers submit ? or shithub genned ? or how. ) [13:41]
mircea_popescu: possibly a mix. [13:41]
mircea_popescu: understand how "yubikey" works : hitlerist website buys a boatload, extracts pub/privkeys, sends to losers. [13:41]
asciilifeform: mircea_popescu: nominally yes, only minions however a large qty iirc has leaked out. would be interesting to get hold of'em somehow. [13:41]
mircea_popescu: this way both "security" and nobus. [13:41]
asciilifeform: aha [13:41]
mircea_popescu: i suppose next someone cracks open one of these fetlifes, can dump the yubikey set this way. tho... very roundabout way of ghoing about things -- could as well just list the privkeys. [13:42]
asciilifeform: those aint kept around on the server end tho [13:42]
asciilifeform: ( the privs, that is ) [13:42]
mircea_popescu: they're kept in the safe. [13:45]
asciilifeform: rright, but in washington, not in whatever flea pit bought the 'yubi' [13:46]
mircea_popescu: depends on size. either at yubi plant or at buyer. [13:46]
mircea_popescu: generally, producer doesn't want to be stuck with them. [13:46]
asciilifeform: producer does not, per the actual magic revealed in the infineon incident, have to know anything at all. [13:46]
mircea_popescu: whereas fucktarded importance-seeking "investors" of customer tend to imagine this makes them less likely to sink. [13:47]
asciilifeform: he gets a 'licensed' chip mask, churns out devices, ships. [13:47]
asciilifeform: said mask contains 'prime constructor' with the obvious boojum intrinsic to it. [13:47]
mircea_popescu: asciilifeform, i meant yubi co, by producer, not whatever random chinese plant spitting these out and making the only money for the only actual economic activity involved. [13:47]
asciilifeform: apples to both in equal measure. [13:47]
asciilifeform: 0 need to yubi to know any seekrit. [13:48]
mircea_popescu: anyway, it's not (as your purity seeking mind would prefer) one true magic bojum. [13:48]
mircea_popescu: there's just a list of crap. "wins at go" algo. [13:48]
asciilifeform: no particular reason it is 'only one true', aha [13:48]
asciilifeform: it simply happens that one of them, got out, last autumn. [13:48]
mircea_popescu: believe it or not, empire still runs on excel, empire still prefers the list 1, 3, 5, 7 to the generator n*2+1. [13:48]
asciilifeform: ( concretely, a key generator that produces a fairly good quality -- from 'nobus' pov -- class of weak keys ) [13:49]
asciilifeform: well generator is readily transformed into 'list 1,3,5...' prior to being given to whichever minion. [13:49]
mircea_popescu: which are then specifically enumerated (yes, in an excel spreadsheet), with the respective associations, and then stored in a fucking "safe", ie a spurious hunk of metal that'd do nothing in any case. [13:49]
asciilifeform: sorta how they started doing with 'radio keys' for costly autos in 1990s, aha [13:51]
mircea_popescu: then a "court-ordered" whatever the hell takes three weeks because the only person that even knew what an index is in the entire org quit over not wanthing to wear the corporate tshirt at the corporate event and so on. [13:51]
asciilifeform: ( when the 'repo man' side of the easycredit usg.auto.industry started having problem with clean repossession on acct of electrical locks ) [13:51]
deedbot: http://trilema.com/2018/minigame-smg-april-2018-statement/ << Trilema - MiniGame (S.MG), April 2018 Statement [13:51]
trinque: asciilifeform: already explained that if the RSS is ever longer than 20 when checked, only 20 are coming through deedbot [13:52]
trinque: no plans to fix [13:52]
asciilifeform: trinque: right, makes sense [13:52]
asciilifeform: trinque: no particular reason to try to fix [13:52]
mircea_popescu: so then why were you prodding him ? [13:52]
trinque: naw, but iirc spyked has one coming, maybe he'll do the problem better justice [13:52]
asciilifeform: mircea_popescu: previously thought 'if only deeper queue', then we moved whole thing to 'go and read' [13:53]
mircea_popescu: pretty much. [13:53]
mircea_popescu: may be actually the best thing for nsa anyway. let people use the website. [13:54]
asciilifeform: trinque: fwiw i dun expect any moar 1000-modulus bursts in foreseeable fyootoor [13:54]
mircea_popescu: hm. [13:55]
trinque: http://btcbase.org/log/2018-05-06#1810461 << this has been my experience also. deedbot was more reliable each time I had to rip the thing out of one host, move to other, for obvious reasons of having to look under the log. [13:57]
a111: Logged on 2018-05-06 15:58 mircea_popescu: i basically hope we keep losing it. seems the more indermeddling, the stronger the republic. [13:57]
mircea_popescu: the hood ? [13:57]
trinque: well, "what did I have in tmux" etc [13:58]
mircea_popescu: aha. [13:58]
asciilifeform: meanwhile from mircea_popescu's www, 'In my opinion the biggest red flag is that the opie intends to pay the artist by bitcoin. How many people have a bitcoin account. PayPal is a widely recognised payment method that has been around for a long time. It is legtimate, safe and easy to use. Even if the opie does not want to use PayPal what is wrong with bank transfer.' << lol!! [13:59]
* asciilifeform was quite certain that this species had properly died out when the usdrate grew its most recent trailing 0 [14:00]
mircea_popescu: there's just this massive class of entirely spurious sacks of shit, want to "participate in leadership" for a living. nothing else, they earned their keep at the ballot box. [14:00]
asciilifeform: mircea_popescu: typo in 'not going to to anything' [14:01]
mircea_popescu: entirely the source of the vague spasms misrepresented by pantsuit into "public support for the electoral process" and "representative legitimacy" -- dorks who literally imagine they earn their 4 years' living by putting down an x once. [14:01]
mircea_popescu: ty [14:01]
mircea_popescu: http://btcbase.org/log/2018-05-06#1810463 << in other shit you won't believe : girl cleanned out her purse, as it was getting heavy. produced as a result a satchel of coins, which she shows me, "check this shit out". "whoa, that's like... almost a kg. go weigh it, i'm curious." so she goes weigh it, and the scale says... 877! [14:36]
a111: Logged on 2018-05-06 16:32 asciilifeform: soooo summary of the eating of tail end of Framedragger collection: 877 popped mods. nearly matched prediction from earlier ( http://btcbase.org/log/2018-05-04#1808858 ) [14:36]
deedbot: http://www.thedrinkingrecord.com/2018/05/06/it-turns-out-pharmacies-in-uruguay-are-actually-rather-civilized-compared-to-old-country/ << Bingo Blog - It Turns Out Pharmacies In Uruguay Are Actually Rather Civilized Compared To Old Country [15:11]
deedbot: http://trilema.com/2018/oregon-represent/ << Trilema - Oregon represent! [15:18]
mircea_popescu: !#s "sulfmeto" [15:37]
a111: 2 results for "\"sulfmeto\"", http://btcbase.org/log-search?q=%22sulfmeto%22 [15:37]
mircea_popescu: yup, it's both a) the ideal medication for the sort of cold / gonorhea and b) readily available in sane lands. [15:38]
mircea_popescu: that said, you should prolly still see a doctor if you actually suspect you had pneumonia the risk is survivant infection, gotta make sure you actually cleanse the bugs not merely reduce them to livable level. [15:47]
BingoBoingo: http://btcbase.org/log/2018-05-06#1810594 << It's in the cards soon. The option to put the waiting AFTER the relief has started instead of giving the bugs more time to do their damage is nice. [16:08]
a111: Logged on 2018-05-06 19:47 mircea_popescu: that said, you should prolly still see a doctor if you actually suspect you had pneumonia the risk is survivant infection, gotta make sure you actually cleanse the bugs not merely reduce them to livable level. [16:08]
mircea_popescu: well, pneumonia is still a perfect way to die, even today. so yes, immediate response is necessary an' appropriate. but bear in mind that biseptol+paracetamol is field medicine, once back in town you're really supposed to go to a hospital and have some tailored cocktail made. [16:09]
mod6: hola [16:17]
mircea_popescu: meanwhile in weblulz, "ad by Scry We’re building a community of people who can predict the future. We are looking for people who can predict the future, possibly better than experts can. Join Us at Scry." [16:18]
mod6: :D [16:20]
mod6: Who wants to rent the last Rockchip available rockchip @ Pizarro? Let us know! First come, first serve. [16:21]
mod6: bah, redudancy. [16:21]
mod6: Anyway, one left! [16:22]
mod6: My rockchip has the FG hooked up, looking good. [16:25]
jurov: asciilifeform: use the tarball as-is, i have not scanned more since [16:41]
jurov: and users submit their keys generated outside github, they recommend ssh-keygen [16:45]
mircea_popescu: in other news : bad teacher (cameron diaz doing her usual fare) is actually pretty good. racoon eyes not giving a shit, very 1990s hipster highschooler aesthetic. [17:37]
esthlos: looking to cement understanding: with current V, file-level merges are impossible, but patches touching separate files can be applied sequentially. this way, it is possible, e.g., for different folx to work on separate parts of the project, and press there changes into one item. [18:52]
esthlos: now let's say that the new philosophy file contains hash of all directory contents in lexographic order, onse per patch. As far as i can see, this forces the tree of vpatches to be strictly linear, since latest patch depends strictly on single previous patch [18:54]
esthlos: so is this what's desired, or am I off the mark? [18:54]
esthlos: *their changes [18:56]
trinque: esthlos: that's incorrect, current V does not press all leaves [19:00]
trinque: current V presses the patches walking upward from the "head" given as a parameter [19:01]
mod6: ^ [19:01]
trinque: the history file is indeed the chosen solution to the problem of tree fragmentation. all patches which are intended to have permanence in the v tree shall edit that file. [19:02]
trinque: this oughtn't be enforced by the V implementation. operator might fully intend to *not* include a patch in the formal history of the project [19:03]
trinque: experimental patches for example. [19:03]
esthlos: so I'm not talking about pressing all the leaves, but when a vpatch has multiple parents, such as h or i in http://www.mod6.net/sps2_dag.png [19:03]
trinque: in current V, j had to *edit* items in both h and i to keep them in the set of pressed patches, if pressing to h [19:06]
trinque: and this is appropriate. the way to solve the problem of "have to edit something in each desired antecedent" is to do that, edit something, the history file [19:06]
trinque: if I haven't addressed what you're talking about though, please elaborate. [19:07]
esthlos: I think I understand, though I may be being thick. what I'm describing is differnt [19:12]
esthlos: if the new V defines the current state of the project as a single hash, then multiple parents (such as h, i, d, and j) becomes impossible [19:13]
trinque: it doesn't, so how did you get there? [19:14]
trinque: you should ignore the "history" file notion entirely it has no bearing on how your v-tron operates [19:14]
trinque: it is a fact about correct operation of a V, and perhaps a V client does something to help the operator along "hey you didn't edit the history file, wild patch!" but does absolutely nothing differently. it's just another file. [19:15]
trinque: sections of a given patch will yes, be parented on the previous, in a straight line, but this does not at all mean that the patch doesn't *also* have other parents, following the antecedent lines of the other files [19:16]
trinque: the graph will still look similar to the png you linked, but with a singular line also running through it [19:16]
mod6: as a side-note, my vtron when doing operations and graphing checks *all* edges. [19:17]
esthlos: oh, I thought the new idea was to determine parenthood based off a single line in the philosophy file. this was my interpretation of http://btcbase.org/log/2018-01-23#1774751 and http://btcbase.org/log/2018-01-23#1774760 [19:17]
a111: Logged on 2018-01-23 12:09 esthlos: to determine if b.vpatch descends from a.vpatch, my idea is to scan through b.vpatch and ensure that each ---(file,hash) matches some +++(file,hash) in a.vpatch. is this the standard procedure? [19:17]
a111: Logged on 2018-01-23 14:28 mircea_popescu: esthlos it is not standard procedure the emerging consensus is to have a dedicated philosophy file which a) all patches must touch (by protocol) b) contains comments as to the patcher's state of mind and c) contains one line per patch uniquely identifying it, machine generated. the format's not fixed yet, but as phf is working on a new proper vdiff it's probably going to coalesce around a variant of whatever he uses. [19:17]
trinque: no, not at all. [19:18]
trinque: esthlos: you are misunderstanding it entirely [19:18]
esthlos: hmm [19:18]
trinque: parenthood is determined by the graph of patches as created by walking across the hashes *per file* [19:19]
trinque: and it so happens that one is a history file. [19:19]
esthlos: alright. What I don't understand, then, is mircea_popescu's response [19:21]
mircea_popescu: esthlos> looking to cement understanding: with current V, file-level merges are impossible << huh ? [19:21]
trinque: esthlos: if I thought you'd gone that far afield, I would've asked you to do all this instead of a simple cleanup of interfaces [19:22]
trinque: please do not try to attach complexity barnacles to the history file, because it seemed very important in logs, or something. [19:23]
trinque: only thing that a V would even conceivably do related to "history" is indicate in "flow" that certain patches are not mainline, for lack of a history file edit [19:24]
trinque: if I tell the thing to press to a signed patch, it should press, history file edited or not. [19:24]
mircea_popescu: i confess i don't understand what he understood of v. this for instance fails to parse entirely : http://btcbase.org/log/2018-05-06#1810623 [19:24]
a111: Logged on 2018-05-06 23:13 esthlos: if the new V defines the current state of the project as a single hash, then multiple parents (such as h, i, d, and j) becomes impossible [19:24]
trinque: aha, I haven't a clue how he got to "thus can merge" [19:25]
esthlos: well, this seems like https://pbs.twimg.com/media/Ct4DkT7VYAAuuzx.jpg [19:25]
mircea_popescu: aanyways i agree with trinque in that the "history file" needs not be considered by the v implementation as anything in particular. it's a usage convention not a special case. [19:26]
mircea_popescu: so yes, a correctly implemented vtron has no history file notion. [19:26]
mircea_popescu: http://btcbase.org/log/2018-05-06#1810638 << which part ? [19:27]
a111: Logged on 2018-05-06 23:21 esthlos: alright. What I don't understand, then, is mircea_popescu's response [19:27]
ben_vulpes: esthlos: did you read mod6 or asciilifeform's v implementations before writing yours? [19:27]
esthlos: ben_vulpes: no, actually that's part of what asciilifeform encouraged me to do [19:28]
mircea_popescu: apparently in my naivite i bestowed upon trinque a larger dollop than originally realised. [19:29]
mircea_popescu: esthlos, what sort of barbarian, orcish, utterly insane and self-spiteful approach is this, whereby you set forth to do a thing without understanding the previous iterations of the thing you're doing ?! [19:29]
mircea_popescu: crossdressing, for all the social stigma, cruising highway bathrooms looking for glory holes, for all the infection potential, sitting by a wall and banging your head against it, they're dubious behaviours, yes, but this fucking takes the cake. what, you hate yourself QUITE TO THAT DEGREE ?! why ? what did you ever do to yourself to take such umbrage with yourself ?! [19:31]
esthlos: mircea_popescu: what I took away from your comment was that the philosophy file was to be used to determine vpatch ancestry, rather incorrectly it seems [19:32]
mircea_popescu: the idea was that it'd work passively to determine order in afore-ambiguous situations. [19:33]
esthlos: and wrt self hatred, I'm really at a loss. somehow it thought like a good idea [19:34]
mircea_popescu: but doesn't the sadness of solipsism strike you ? it's not hard enough to have to go to war, you'll do it naked and barefoot and hope you can find some flint to make a spear before the vietcong finds you and gangrapes you ?! [19:35]
mircea_popescu: they're there to help, not to hinder, all the accumulated piles of previous materiel. or does this not seem so ? [19:35]
esthlos: indeed. I am rather lonely, in general [19:36]
mod6: moar forum then :] [19:36]
ben_vulpes: https://www.esquire.com/news-politics/news/a28718/why-men-love-war/ [19:36]
mircea_popescu: well yes, but you don't have to be. i get it, in general everything's crap, but dja think i'd even for a second have permitted it if it were ? the stuff about, it's all stuff that somehow survived the republican demolition frenzy. it's not just random gunk. [19:37]
trinque: esthlos: so at least less to change about your v-tron right? though definitely *read* the others so you're more certain you understand what the thing is. [19:42]
trinque: I intentionally aimed you at "fix the interfaces" because it appeared that was the thing most glaringly wrong with it, and not all this [19:43]
trinque: it'd be a happy thing if I can stay with my welder in portage's innards while this happens and not have to end up doing that part myself [19:45]
esthlos: mircea_popescu: i'm at a loss of what to say. but really, it's moving [19:46]
esthlos: trinque: no, I didn't jump off the deep end without discussing it. I've already made the interface changes [19:46]
mircea_popescu: aite. [19:46]
esthlos: let me give it a once over, sign it, and send it your way [19:47]
trinque: ah neato [19:47]
mircea_popescu: meanwhile in saturday night fun, https://78.media.tumblr.com/413a1d5da92e80bc7b3c705f8f435ed5/tumblr_p7ouhi3DeZ1u7dxx4o1_1280.jpg [19:48]
esthlos: haha does her collar say BBC? [19:49]
mircea_popescu: possibru. [19:52]
esthlos: trinque: updated file at http://files.esthlos.com/crypto/v_2/v.lisp , sig at http://files.esthlos.com/crypto/v_2/v.lisp.esthlos.sig [21:43]
esthlos: mircea_popescu: wrt http://btcbase.org/log/2018-05-03#1807632 << esthlos-v presses fine to vtools_vpatch_newline [21:48]
a111: Logged on 2018-05-03 01:15 mircea_popescu: esthlos, a) http://btcbase.org/log/2018-01-23#1774760 b) how does your item handle the original http://btcbase.org/log/2018-04-09#1794417 problem ? [21:48]
trinque: very cool on both counts. reading. [21:49]
mircea_popescu: nice. [21:50]
esthlos: i'll field and questions/criticism tomorrow, gotta run for now [21:52]
esthlos: also want to say that above confusion was mostly misinterpretation of various comments on my part. not about what "v" is _now_, but on what was being asked of me [21:59]
trinque: the communication gets easier the more you work it, so stick around [22:21]
trinque: btw looks like missing definition of process-exit-code ? [22:21]
trinque: I'd also like to see the cl-ppcre dependency drop (not only for the call out to quicklisp to obtain it) [22:22]
trinque: the parsing here is pretty simple, oughta be able to implement without hauling in a big honking dep [22:22]
* trinque will digest more when he has a definition of process-exit-code [22:23]
mircea_popescu: esthlos, people throwing their hands in despair over something [they thought] you said is always way the fuck preferable to people throwing up their hands in despair over something you did -- specifically because there's a lot more room to misinterpret what's said than what's done. [22:25]
trinque: ah, sbclism. I'm on ccl over here. and looks like run-program's an undeclared dep on uiop [23:45]
trinque: wise thing to do is muntz off these foreign entanglements. v needs to stand on its own as much as possible, and ever more so. [23:47]
trinque: can implement a "run-program" of our own in here, implement for sbcl and ccl. I don't know that I've heard of anyone here using another. [23:47]
trinque: I'm fine with helping. could take a crack at it tomorrow. [23:48]
trinque: worth saying also that use of quicklisp is anti-v. it's a wad of packages some dude compiled, he's not in our WoT, no signatures on packages (last I checked), and to top it off, it hauls down foreign code and executes for you as a single op. [23:53]
trinque: v doesn't import the imperial convenience idiocy. it eats the world from where it stands. [23:54]
trinque: (think, when this is done, we have a basis for a vtronic lisp repository, aside a vtronic portage, and all the rest) [23:55]
mircea_popescu: i agree must stand on own more than possible [23:57]
mircea_popescu: trinque, cmucl = ccl ? [23:58]
trinque: !!v 55614ADF1D52CD18F16C72676CC8FF0FE7C24F13662752368EDE17A2F082F05D [23:58]
deedbot: trinque rated esthlos 1 << lisp vtronicist [23:58]
trinque: ccl is "clozure cl" [23:58]
mircea_popescu: i thought phf mostly did cmu cl [23:59]
Category: Logs
Comments feed : RSS 2.0. Leave your own comment below, or send a trackback.
Add your cents! »
    If this is your first comment, it will wait to be approved. This usually takes a few hours. Subsequent comments are not delayed.