log☇︎
61800+ entries in 0.037s
BingoBoingo: Meanwhile in local news: Uber has an office here. Office is in solid favella territory https://www.elobservador.com.uy/nota/oficinas-de-uber-en-montevideo-fueron-apedreadas-y-debieron-cerrar-20181114182928
BingoBoingo: Anyways bunch of lessons to write up
mircea_popescu: heh. cool then
BingoBoingo: mircea_popescu: Turns out lighttpd didn't do the pingback thing
mircea_popescu: BingoBoingo this blog disaster of yours has rescued a bunch of pingbacks you apparently never sent b4
BingoBoingo: If cute girl dildo'd rock, gets investigated by smell, even if it turns out the culprit was honest clams
mircea_popescu: problem is he hunts chunks of flotsam, rock etc half the time. ☟︎
mircea_popescu: i still don't know what you dislike about it, other than it's continued existence per se.
asciilifeform: imho it is by far weakest link in the ensemble.
asciilifeform: i'm willing to 'marry' rsa, but not oaep, oaep only willing to casually fuck
asciilifeform: at any rate, i dun have an obviously-superior replacement for oaep. so presently ( see my rollout calendar ) i'ma implement one. but i dispute that 'cannot be made' claim.
asciilifeform: but if it were possible to have straight rsa, rather than rsa and hash,would be stronger system ( no hash ever was proven strong, even in the sense rsa is )
asciilifeform: mircea_popescu: we have hashes cuz some thing simply cannot be done without'em ( short rsa sigs of GB inputs, etc )
mircea_popescu: and no, thread that prompted me to publish the lulzitem was specifically the glib "oh, NOW here's shits and giggles"
asciilifeform: ( how to take plaintext, with possible known patterns, and turn it into something about which NO useful 'and bit B is X' can be said )
mircea_popescu: if the hash is not there to carry load then why have it.
asciilifeform: recall the thrd that provoked mircea_popescu into making shot at own hash algo etc. orig was a stab at the generalized problem of destructurizator.
asciilifeform: it puts 9000 tonnes of load on the hash to carry.
asciilifeform: it'll turn out to suck tho if it turns out that , say, if you know 3bits of what went into a keccakization, you now know 33 of its oaepization, or the like. i dun know a proof that ~not~ this, if that makes sense
asciilifeform: mircea_popescu: the 'proof' is iffy ( iirc we dug into this in early thrd ) but that aint the thing that makes me hairs stand up: it's that oaep introduces structurism ( for comparison, even the oldest 'all or nuffin' algo, rivest's , does not )
lobbesbot: BingoBoingo: The operation succeeded.
BingoBoingo: !Qlater tell trinque Feel free to return the feed
a111: Logged on 2018-11-14 20:50 asciilifeform: but for general-purpose pgp replacement, conceivably could use something 'hungrier' but with 0 fixed structural bits. i'ma invite mircea_popescu et al to consider the subj.
mircea_popescu: http://btcbase.org/log/2018-11-14#1872257 << i am not aware something better than oaep for rsa padding ~could in principle be had~, let alone actually is known to exist. ☝︎
a111: Logged on 2018-11-14 20:20 diana_coman: my understanding was that nobody actually LIKES oaep all that much but it's (again! another one of those!) the thing we have (as opposed to the thing we might wish for)
mircea_popescu: http://btcbase.org/log/2018-11-14#1872250 << not so. i actually like it, as being one of the few proven-hard bits of the witches' brew we call crypto. ☝︎☟︎
BingoBoingo: Yeah, timed mysql dumps
BingoBoingo: Anyways, figuring out chron is now high on the todo list
asciilifeform: pretty sure that diana_coman is on rk tho
BingoBoingo: hanbot is on the shared machine
BingoBoingo: I have not. My understanding of how it worked was it sat in a failry consistent RAM footprint as a running process and played goalie shoving the blog to php-fpm
asciilifeform: ( seems , at least by my lights, uncharacteristically hungry... does it really take a coupla MB of working ram , per GET , to serve up a blog ? )
BingoBoingo: The failure mode under stress was an attractor to lighttpd, but... configuring it is a pain
asciilifeform: esp. a wwwtronic thing, where there is a readily-available graceful failure mode ( make connector wait a few sec for a free slot )
asciilifeform: erry proggy should at the very least take a param for max ram guzzle
asciilifeform: imho people should not default to runaway-ram
BingoBoingo: Consider Phuctor is a sorta bespoke thing. DB software architect can't antipate a Phuctor and how its filled. On the other hand most mysql installs power wordpress.
asciilifeform: BingoBoingo: fwiw postgres doesn't, 'out of the box', have the runaway ram problem. ( it had exact opposite -- i had to coax it into actually making constructive use of dulap III's very plentiful ram )
BingoBoingo also moved from lighttpd in the last incarnation to apache in this one
BingoBoingo: From eating more than half the Rockchip's 2gigs to ~130 megs
BingoBoingo: Changing the storage engine has resulted in a more than 10x reduction in RAM used
asciilifeform: possibly 1 of these days i oughta publish my kludge for making ancient wp go on postgres; but it isn't half as polished as mp's and i dunno that anyone would win from cribbing it
asciilifeform: phuctor ( and in particular, some of the 'heavier' / unusual pheatures, like search ) i baked specifically around postgres. ☟︎
a111: Logged on 2018-11-15 02:09 BingoBoingo: On the plus size, I did learn that using MyISAM instead of Innodb actual RAM available on the Rockchip for not MySQL
asciilifeform: http://btcbase.org/log/2018-11-15#1872279 << currently i know veeery little about mysql ( always used postgres, and at this point know embarrassingly much re the internals and tuning knobs ) -- but iirc mp's-wp requires specifically mysql, so prolly doomed to study it at some point ☝︎☟︎
asciilifeform: ( plenty of folx made 'soft' raid, but those imho are worse than useless )
asciilifeform: btw for many yrs i've searched for the obvious simple gadget, a y-shaped thing that'd turn 2 or moar usb sticks into an iron raid. but still not found, dun seem like anybody ever made.
asciilifeform: jurov do you recall what it was they used ? ( or was it custom )
asciilifeform: i was almost gonna say 'we're the first to put arm64 in a dc rack' but then recalled that jurov, years ago, saw such a thing somewhere
asciilifeform: BingoBoingo: if we ever build a rk-like board from the ground up, i'ma give it at least mirror raid. but presently afaik no such thing can be had.
lobbesbot: asciilifeform: The operation succeeded.
asciilifeform: !Q later tell ben_vulpes didja find the eggog ? you estimated 'wednesday', and the day is nearly at end.. ☟︎
asciilifeform: also folx, pleeez make backups! i should not even have to remind. there is not such a thing known presently as an immortal disk.
asciilifeform: mod6 pleez let me know when we can swap yours, iirc you're the last rk fella on old disk
BingoBoingo: On the plus size, I did learn that using MyISAM instead of Innodb actual RAM available on the Rockchip for not MySQL ☟︎☟︎
asciilifeform: dunno why BingoBoingo felt compelled to tempt the fates
asciilifeform: i did say 'let's errybody move to new rk disks asap' , neh
asciilifeform: BingoBoingo: didja end up having to glue yer blog back together from archived pages ?!
BingoBoingo: <lobbes> "el incendio anual" heh << Mind that before becoming a mall in a more expensive neighborhood than the Datacenter's, a good chunk of the present legislature or their parents were locked up there for being commie scum
BingoBoingo: trinque: ping me at your leisure to announce the silence
trinque: can't get to it immediately, in about an hour or two can
mod6: <+diana_coman> mod6, re comment, it seems it got lost, I could not find it anywhere so not in queue nor anything << not a problem at all. we've resolved it anyhow. thanks for looking.
asciilifeform: ( i.e. when ciphertext passed through e.g. 'dieharder' it would be indistinguishable from FG )
asciilifeform: ideal algo imho would carry at least 5 bit of entropy for erry bit of payload, and in such a way that all bits are 0/1 with exactly 0.5 prob.; and such that flipping one bit of ciphertext flips at least 1/2 of the output bits.
asciilifeform: but for general-purpose pgp replacement, conceivably could use something 'hungrier' but with 0 fixed structural bits. i'ma invite mircea_popescu et al to consider the subj. ☟︎
asciilifeform: and hence prolly also is The Right Thing for other space-constrained applications, e.g. udptronic gossipd
asciilifeform: ( that is, oaep )
asciilifeform: diana_coman: it's the most economical, at the very least, algo that sorta does the job
diana_coman will go afk too
diana_coman: my understanding was that nobody actually LIKES oaep all that much but it's (again! another one of those!) the thing we have (as opposed to the thing we might wish for) ☟︎
asciilifeform: the fixed structure elements in oaep bother asciilifeform . ( initially was gonna do destructurization differently : each bit of payload turned into 4 via rng xor, then fisher-yates shuffle, then the 'deshuffling' binarysort code is appended to message. you can prove that the output is 'all or nuffin' transform. )
asciilifeform: i was hoping to avoid baking hashing into ffa/p , but loox like it isn't escapable if we're doing oaep
diana_coman: meanwhile the oaep got sorted better as part of smg comms really (i.e. only ada calling c, no back and forth dance and that gets rid of a LOT of confusion)
asciilifeform: it's 1 of the coupla items i dun have yet
asciilifeform: i in turn am working through diana_coman's oaep articles
diana_coman: but it might still take a while to get back to it
asciilifeform: ( iirc you already ate the most difficult material )
asciilifeform: diana_coman: lemme know if you run into anyffing to any degree confusing.
diana_coman: the table with FFA patches looks great; I'll look into carving again some space to re-start on it
diana_coman waves to asciilifeform
BingoBoingo: <mircea_popescu> BingoBoingo trilema is mostly on myisam. << ty
mod6: thanks for doing all this work diana_coman, I appreciate it. it's long over do for a v'ing. :]
mod6: <+diana_coman> mod6, I meant out of date wrt to my vpatch i.e. keccak hashes << ahh, gotcha
diana_coman: anyway, if anyone wants to maintain the full & detailed manual, I will happily sign their .vpatch with it,sure
diana_coman: onth I kept the quick guide and updated it to point users to a. use the bloody help of V as it's quite useful as it is! b. ask if they get stuck, wtf
diana_coman: the out of date also refers to the fact that meanwhile V has a broader scope really than trb code
diana_coman: if you notice, I said out of date at 3rd vpatch, NOT before; and yes, 2nd vpatch includes them precisely because they are at that point still perfectly fine
diana_coman: mod6, I meant out of date wrt to my vpatch i.e. keccak hashes
mod6: Anyway, *nod* looks good. I'll try to test it one of these days when I have a spare minute.
a111: Logged on 2018-11-14 10:54 diana_coman: and done: the updated .vpatch files and starter_v.zip + sigs are all up; I've checked them on my RK; genesis is the same and result is identical to V-2017; the result of 99993 .vpatch is identical to V-2018; the result of keccak_vtools vpatch is to update BOTH docs and code; re docs, I've nuked the user manual as I won't maintain it and it's becoming confusing due to being out of date; I've updated the quick_guide however, mainly for 1st t
mod6: http://btcbase.org/log/2018-11-14#1872191 << Totall get it that you don't want to maintain the user docs, I do disagree that they were are out date tho. I did a diff between them from the 99994K and 99993K versions and I did update it for the Feb release. ☝︎
mod6: http://btcbase.org/log/2018-11-14#1872187 << no worries, I tried to post a comment on your blog post from the train, it may have not went through. ☝︎
a111: Logged on 2018-11-14 07:55 diana_coman: the genesis IS on 99994K yes, the next vpatch upgrades it to 99993K and only THEN my own vpatch changes it to use Keccak and vtools
mod6: http://btcbase.org/log/2018-11-14#1872182 << Ahhh! I see what you did there. ☝︎
mod6: tarting MDMCommandQueueMonitor through JMX
mircea_popescu: i meant diana_coman ah i see, you redid the whole tree, that was a patch.
mircea_popescu: BingoBoingo trilema is mostly on myisam.
mircea_popescu: asciilifeform oh i see, you redid the whole tree, that was a patch
asciilifeform: bvt: in all srs, thx for digging it up, and if anybody finds another , plox to also post, there is room in the kunstkammer .
asciilifeform: and 'how not to crypto', pretty good illustration of the tension between 'machine proofs' and 'fits in head' (author resolved wholly in favour of the former and took a tall shit on the latter)