asciilifeform: out of duct tape and chewing gum, and, to add insult to injury, cannot really be rebuilt
asciilifeform: imho one of the most perplexing things about 'anti-bitcoin' agitators is how they all - to a man - miss the most cogent argument against bitcoin: it being, how it is implemented☟︎
asciilifeform: (notice, incidentally, that cpp is virtually unused outside the microturd world ?)
asciilifeform: not 100% certain that we're all using same bdb...
asciilifeform: there is a cache cap option. setting it did precisely nothing.
asciilifeform: our ver. of bdb actually lacks the set_memory_max
asciilifeform: oh perhaps i should mention that i tested the db memory thing
asciilifeform: (it works without the patch but you can't take runtime heap snapshots then)
asciilifeform: mod6: the igprof thing was its own patch
asciilifeform: so the statement in the report - holds.
asciilifeform: mod6: i would comment that technically the block-bastardage and tx-orphanage nukes are semantically independent. but the latter was derived from a main.cpp patched with the former.
asciilifeform: decimation: in both cases, 'martian' - that is to say, unverifiable with info at hand - blocks (or transactions) are indistinguishable from garbage at time of receipt
asciilifeform: decimation: if you didn't verify the entire blockchain from end to end, you aren't a node!
asciilifeform: but afaik they no longer ship the dead tree
asciilifeform: lispworks is actually (afaik to this day) the maintainer of the online cl hyperspec! while documentation from franz was once superb - came in printed volumes, which i bought second-hand
asciilifeform: <ben_vulpes> trinque and i were wondering about their CLIM implementations, and if documentation etc was better on the closed-source side of the world << i never had lispworks and can't comment about it, but franz 'allegro' does come with clim. iirc it only works in the non-smp version though, and depends on gtk
asciilifeform: it is a consequence of how screamingly retarded the sync mechanism is to begin with
asciilifeform: decimation: asciilifeform: maybe one solution to the 'bastard transmitter' problem would be ... this is still reasoning from the misconception that 'bastards' necessarily come from malicious or malfunctioning nodes. this is not the case - a 0.5.3.x will emit bastards - from point of view of certain nodes - just as often as any other
asciilifeform dreams big: of hammering with -electron- microscopes !
asciilifeform: <ben_vulpes> Adlai, asciilifeform: are the commercial lisps worth paying for? << extremely bad question. is a boeing worth paying for? depends!
asciilifeform: mircea_popescu: granted, but 'don't let strangers throw ten pounds of shit each into your five-pound bag' is not higher mathematics ...?
asciilifeform: anyway, for all who want, replace 'return true;' in 'if' block where begins with printf("ProcessBlock: BASTARD... with 'if (pfrom) pfrom->Misbehaving(whateverconstant)'
asciilifeform: once we have a reasonable 'embrace and extend' of sync mechanism - then yes
asciilifeform: there would just be folks shooting selves in the feet
asciilifeform: mircea_popescu: aha but here there isn't a 'me' and 'you' yet
asciilifeform: but this reflects the lack of a reasonable 'gossip' mechanism - on both ends
asciilifeform: it is 'bad' from standpoint of node which doesn't have the antecedents, yes
asciilifeform: mircea_popescu: our 0.5.3.xyz isn't any more clever on the transmitter end!
asciilifeform: i have come to the conclusion that the -only possible- non-retarded solution to bastard blocks -when we can only adjust our end- is to throw them into the sun.
asciilifeform: http://log.bitcoin-assets.com/?date=31-05-2015#1149116 << i tried various values of 'misbehave' increment per orphan block - 20, 50, even 100. in my tests, this made sync... slower. reason, as i understand, was that 1) you lose time renegotiating connection 2) overwhelmingly likely that any node you reconnect to -also- shits just as many orphans.☝︎☟︎☟︎
asciilifeform: <mircea_popescu> yeah its weird << nothing weird about it. the sync mechanism is braindamaged more or less as far as it is even possible to be braindamaged.☟︎
asciilifeform: ^ makes it possible to preserve debug output from zero to synced in its entirety
asciilifeform: (incidentally, a while ago i posted a patch that kills log rollover. might be of interest to some of you)
asciilifeform: decimation: this is normal in my experience
asciilifeform: mircea_popescu: again, signing a key - on ordinary gpg - does not throw your whole modulus in it
asciilifeform: 'the fraud could have been uncovered sooner, potentially forestalling a great deal of the disruption it inflicted on various careers and on social science as a whole.' << mega-l0l! as if any of the 'careers' built upon this or similar crocks of shit - other than the exemplary scapegoat's - will skip so much as one beat
asciilifeform: as if any of these gasbags ever seriously replicate anything
asciilifeform: '“My puzzlement now is, if he fabricated the data, surely he must have known that when people tried to replicate his study, they would fail to do so and the truth would come out,”' << l0l re: the obvious 'fnord' here
asciilifeform: 'This was a graduate student who successfully ran the gauntlet of the Princeton interview process with a publicly posted CV that contained wild falsehoods about his grant receipts ... ' << this is sorta like counterfeiting usd. when the counterfeiter-in-chief cranks up the press and doubles the toiletpaper in circulation, that is blessed by the crown. when some schmuck ? headzoff
asciilifeform: 'see, the crown cares! the rare irregularities are dealt with! watch this impalement and move along now'
asciilifeform: sorta like the very occasional corrupt dogcatcher they shoot in china
asciilifeform: 'It’s the sort of brazen data fraud you just don’t see that often, especially in a journal like Science.' << this particular item from the gobbelsism toolbox needs a name
asciilifeform: ;;later tell mircea_popescu paused pump, resumed werker. we have around 1/4 of the total rsa keys available submitted (or processed) at this point.
asciilifeform: if someone can suggest a better phrasing for this, please write in.
asciilifeform: because said key was hardcoded in, or some schmuck had his seldom-used privkey substituted under his nose, or whatnot.
asciilifeform: the idea of phuctor is to detect duplicate privatekey material in use anywhere. the reuse of one particular divisor is only one possible scenario - it is also possible that a very ham-handedly sabotaged pgptron somewhere reuses a whole pair of factors - that is, two or more strangers may, somewhere, share an entire private key.☟︎
asciilifeform: jurov: now, how one checks that: "Please make sure these are yours" ? << i suppose this is when i explain -why- i wrote the 'dupes' feature to begin with.
asciilifeform: the basic unit of pgptronics is the public modulus / exponent set, rather than 'key' - which is a fairly arbitrary collection thereof
asciilifeform: jurov: but i can take any set of moduli and shoehorn them into a valid key, sure
