tmsr - Log 2018-02-01

▇▁▁▁▁▁▁▁▁▁▁▁⏐︎▁▁▄▂▁▁▁▃█▃▁▁ 1457

asciilifeform: in other noose, asciilifeform recently looked at 'lighthttpd', after it came up in the log. astonishing tanker of liquishit.

asciilifeform: just like apache, nginx, etc. it has the microshit-quality : you can throw darts all day long and hit only things that really don't need to be there for anything sane people are likely to do with http

trinque: nginx went the venture capital route at some point, gotta justify the rounds ☟︎

asciilifeform: in this particular case it wouldn't even matter if it were run by honest folx

asciilifeform: the very notion of 'production http server' that exists now, is retarded, in its baked-in conceptual support for warcrimes like ssl

asciilifeform: this 'modularity' generates considerable complexity even if unused

asciilifeform: ( this is entirely separate from the fact that all 'industrial' http servers existing nao are in megatonne of c liquishit, language with nothing resembling sane string handling )

asciilifeform: ohai BingoBoingo

BingoBoingo: Hello, machines screwed in. Need to return in the morning to figure out what stupid mistake I made configuring the switch/router. Long night of reading sample configs ahead. ☟︎

asciilifeform: BingoBoingo: what exactly do the digs look like ? cage with own door ? or just rack ? or how ☟︎

BingoBoingo: rack with locking door and badass mesh barrier. Like a birdcage for hannibal lecter

asciilifeform: mesh where ? the back side ?

asciilifeform: normally this is done with vertical chimney

asciilifeform: BingoBoingo: are you able to distinguish 'misconfiged router' from 'dead pipe' , say ?

BingoBoingo: asciilifeform: Gotta remember this is in the basement of the basement

mod6: Lords and Ladies of The Most Serene Republic, The Bitcoin Foundation presents The State of Bitcoin Addresshttp://therealbitcoin.org/ml/btc-dev/2018-February/000288.html

asciilifeform: oh hey

asciilifeform: nuffin in there about ben_vulpes's test of the sync sanity patch

asciilifeform: i can only suppose he dun have a verdict yet...

asciilifeform: ( for that matter, ben_vulpes's own logical extension of same patch , ditto )

mod6: When we have something solid, we'll make mention of it for sure. :]

asciilifeform: in other olds, http://prog21.dadgum.com/116.html >> 'Things That Turbo Pascal is Smaller Than'

asciilifeform: e.g. 'The image of the white iPhone 4S at apple.com (190,157 bytes)' 'The touch command under OS X Lion (44,016 bytes).'

jhvh1: BingoBoingo: Bitstamp BTCUSD last: 9520.01, vol: 13168.35666700 | Bitfinex BTCUSD last: 9668.5, vol: 43087.00794047 | Kraken BTCUSD last: 9488.9, vol: 6767.82091731 | Volume-weighted last average: 9618.18724777

BingoBoingo: Attention Lords And Ladies of The Republic: You may add 161.0.121.253 to your hosts file as Qntra.net

BingoBoingo: !~later tell mircea_popescu Qntra.net online at 161.0.121.253 pls to get the domain repointed, ISP statement incoming. ☟︎

BingoBoingo: Turns out the problem was I overconfigured the switch yesterday. Less is more.

asciilifeform: congrats BingoBoingo !

asciilifeform: diana_coman: slow to load but i suspect that it has qntra hardcoded links and yer browser sits and tries to resolve and fails

diana_coman: asciilifeform, the home page directly with the ip was also slow

asciilifeform: see if still slow after adding to hosts

asciilifeform: diana_coman: i meant specifically when loading via ip

diana_coman: asciilifeform, you are right; works fine after adding to hosts

asciilifeform: wellknown effect

asciilifeform: BingoBoingo: search your html for absolute links to qntra, there's no good reason for them, use relativelinks

asciilifeform: i mean for css/img

asciilifeform: perfectly proper to have clickable links to absolute

asciilifeform: oh hey.

mod6: <+BingoBoingo> Attention Lords And Ladies of The Republic: You may add 161.0.121.253 to your hosts file as Qntra.net << congrats BingoBoingo!

BingoBoingo: shinohai: I still gotta get a bitcoind on the box

asciilifeform: BingoBoingo: be sure to use trb with sync sanity, cuts sync time from 6+months to coupla weeks at worst ☟︎

asciilifeform: ( at some point we will have to solve the puzzle of just how many trb nodes it makes sense to have inside 1 cage, i also suspect )

BingoBoingo: asciilifeform: I put an up to recent block chain copy on the box already, but yes I plan to build the ben_vulpes super aggression for it.

BingoBoingo: Comments on how to handle the fx movements in the future welcome, encouraged, and desired

lobbes: http://btcbase.org/log/2018-01-31#1778994 << heh, they must be reading logz. MPEx now appears under "2014 Fundraising Campaign" ☝︎

a111: Logged on 2018-01-31 19:58 asciilifeform: http://www.openbsdfoundation.org/contributors.html << aaaaand openbsd ~immediately~ listed 'pineapple fund'. and guess who not listed at all, to this day.

lobbes: BingoBoingo: small typo under "Assets, Cash"; You have 0 + 0.20123128 = 0.20113128

trinque: pretty sure it's been there a long time https://archive.is/WwN89

asciilifeform: trinque: i missed it: looked for mp, rather than mpex

asciilifeform: diana_coman: 'gnat-specific methods' link broken

diana_coman: phf, chapter 8 of eucrypt is a fully-working Keccak, albeit at bit level

asciilifeform: diana_coman: neato

asciilifeform: seems to use whole word (i.e. 64b , on amd64) to store each bit tho

diana_coman: tbh I would recommend it as a better "starter" approach anyway

asciilifeform: i assume this was deliberate, sorta like my egyptologies

diana_coman: asciilifeform, yes, at this stage the point was to keep it very clear and simple

diana_coman: the current plan is then to go back to the "word-level" and make that one deal with different endianness

asciilifeform: in re shifts/rotates ( as seen in http://btcbase.org/patches/ffa_ch1_genesis/tree/ffa/libffa/w_shifts.ads ) one nitpick : Shift_Left, etc. are not per se gnat-specific; they are in the standard. but for some perverse reason their ~invocation~ is implementation-specific, and requires this gnaticism.

diana_coman: in principle if one aims for speed, one should look at parallelism I guess

asciilifeform: (supposing this difference makes a difference to anybody)

asciilifeform: fwiw i do not know of a cpu , from any earthling cpu maker, that did not have shifts/rotates by 1

asciilifeform: diana_coman: http://www.adaic.org/resources/add_content/standards/05aarm/html/AA-B-2.html

asciilifeform: fwiw one can also define these like a barbarian, as x * 2 and x / 2 ; and gcc (last i saw) will Do The Right Thing and transform to a shift

diana_coman: funnily enough the rotates are anyway mainly "shortcuts" for 2 (iirc) transformation operations but otherwise not directly in keccak's defs as such

asciilifeform: but this is not clean.

asciilifeform: diana_coman: the observation re 'take as much speed as you actually need, but not more, if costs ugliness' is correct and is in fact the basis for ffa

diana_coman admits that is one of the things she really, really likes about ffa

asciilifeform: the unfortunate thing re 'use word to store 1 bit' is that it will give obscenely disproportionate slowdown, due to spilling out of cacheable size

asciilifeform: but if the application tolerates this -- no reason wainot.

asciilifeform: ( you are likely to see 200-300x slowdown, depending on particular chip, be ready for this )

asciilifeform: diana_coman: in some applications, speed doesn't really matter almost at all. and in those, it is even now possible to, e.g., take ch5 ffa for rsatron.

asciilifeform: ( it is imho unlikely that anyone will demonstrate an rsatron with fewer moving parts than ffa-ch5 )

phf: diana_coman: ty, i'll try wiring it in, and keep you posted. i assume ada<>c interop is pretty straightforward, though i won't have a chance to try until later next week. suddenly busy ☟︎

diana_coman: phf, ada<->C interop was quite straightforward last time I tried it, didn't take much at all; you'll need to add a few exports iirc

phf: keccak patche on btcbase, also asciilifeform ch9 likewise