232000+ entries in 0.146s
mircea_popescu: and in other argentine wtf, /me felt like putting some of
the grandiose coffee available here
to good use as coffee liqueur. sent girl
to farmacy, where she bought 1 liter bottle of pure ethylic alcohol (98%). for like 6 bux. it's fabulous,
tastes mildly of corn.
mircea_popescu: lol
this is like cinderella on monkey island already ;
they keep bringing her all sorts of offerings. "boogers ?" "how about some feces ?" "dried feces from yesterday ?" "how about
this corn kernel" "ok how about dead molluscs"
thestringpuller: because whitewashing security is
the only way
to make people have
trust
these days?
trinque: Teechan is similar in design
to
the Lightning Network, save for one crucial differentiating factor: it leverages
trusted execution environments (TEEs),
that is, secure hardware components found in recent commodity processors such as
the latest batch of Intel CPUs with Software Guard Extensions (SGX). << why
the fuck even bother
talking about
this
thestringpuller: asciilifeform: Emin is pushing SGX like solutions. Seems everyone is infected with security
theatre
these days.
☟︎ mircea_popescu: anyway, yes, it'd seem iris fs
took over
the disks meanwhile, who knew.
mircea_popescu: also
the issue of performance discussed by dave chinner is worthy of consideration (in
trinque 's link)
mircea_popescu: trinque ah, you can recompile with flag and it makes
them 64 ?
trinque: that was not
the claim at all; read
the link
trinque: the xfs limit is iirc defined as "how much disk space do you want
to use for
them"
a111: Logged on 2016-12-22 07:32 jurov: also,
there is limit of 2^32 inodes, even in 64-bit mode
mircea_popescu: (for
the record, in practice
the utility of prepared statements is often nil, and can be mostly captured
through saying eg insert in x values y, values z rather
than insert in x values y, insert in x values z. literally
the whole benefit is
that it compiles
the part prior
to values just once.
a111: Logged on 2016-12-22 07:35 jurov: and
to give my 2c also on sql, while it's possible
to use some portable standardized ansi sql subset, essential stuff like prepared queries need a database driver
a111: Logged on 2016-12-22 07:32 jurov: also,
there is limit of 2^32 inodes, even in 64-bit mode
mircea_popescu:
http://btcbase.org/log/2016-12-22#1588254 << can
this be rewritten ? not
that we're liable
to have over 4bn
txn at any point, but more of
the principle of
the
thing "fuck you and your fucking magic numbers. if i run 64bit processors i'll run 64bit disks also motherfuckers!"
☝︎ mircea_popescu: shinohai chick's from like 2009, what do you want. she gets 2 or 3 btc for
teh dollar.
mircea_popescu: (while
the store-by-hash
txn scheme promises at most 256 directories OR 65536 files however in practice finding 65536 contiguous hash
txn seems rather unlikely for a while yet)
mircea_popescu: intuitively
though i
think blocks should be stored by height.
mircea_popescu: unless you want
to store
them by hash, in which case of course it's at most 65 chars,
though because of difficulty i expect it should be made 40 or less
mircea_popescu: jurov in principle
the address of a block is something like /44/4600 for
the latest, so 2char dir + 4 char name
mircea_popescu: "¿Sabes algo sobre Bitcoin?" "Se que es una especie de almacenamiento de dinero virtual , que por un dólar
te dan 2 o 3 bitcoins o algo así , que básicamente se usa para comprar online y conseguir descuentos creo"
shinohai passes
trinque some coffee [~]D
trinque: ben_vulpes: wooo! I'll
test
that soon as I have coffee in hand.
jhvh1: shinohai:
The operation succeeded.
davout: ben_vulpes: what's
the sqlator?
shinohai: neato ben_vulpes (on
the lisp V)
ben_vulpes: mod6, asciilifeform,
trinque, phf, mircea_popescu, and anyone else
tracking vtronic gnashing: i dusted off and rewrote my cl V implementation. i'll follow up sometime
tomorrow with more demo usage, and a more robust demonstration of wot-variant pressing.
http://p.bvulpes.com/pastes/juTpM/?raw=true ben_vulpes: BingoBoingo, pete_dushenski: in re graphite, "meaningfully stand behind
their Q" perhaps?
ben_vulpes: jurov: sql just falls out of "i shall make
tmsr a block explorer", no more, no less.
jurov: to introduce it into
trb would, in my understanding, would make folks apoplectic
jurov: and
to give my 2c also on sql, while it's possible
to use some portable standardized ansi sql subset, essential stuff like prepared queries need a database driver
☟︎ jurov: also,
there is limit of 2^32 inodes, even in 64-bit mode
☟︎☟︎ jurov: not sure how
to shovel
tx/block id into 60 bytes
jurov: mircea_popescu: ext4 has 256Binodes, and "The
target of a symbolic link will be stored in inode if
the
target string is less
than 60 bytes long."
ben_vulpes: a few minutes of
thinking while working on something related did clear
this up for ben_vulpes
mircea_popescu: moreover if you say press kimkardassian and it says "failed - parishilton not found"
then it is somewhat likely you forgot
to sign blondy.
ben_vulpes: without a flow
that renders all patches. but fine
ben_vulpes: i expect it
to be a royal pain
to run down "which patch is missing sigs and breaking
the flow from genesis
to whatever"
mircea_popescu: should it also show you
the contents of your catpics folder, in case maybe you wanted
to do some clicking ?
a111: Logged on 2016-12-22 02:38 ben_vulpes: but if it doesn't show me which patches are lacking sigs,
that strikes me as a bug.
a111: Logged on 2016-12-22 02:34 mod6:
the printable flow, is
the same as
the pressable flow.
a111: Logged on 2016-12-22 02:32 mod6: you will, you must, have everything signed for it
to show up in a pressable flow.
a111: Logged on 2016-12-22 02:28 mod6: it sounds like everyone wants instead, a general overhaul
to get
to
the 'wot variant press' instead, which would also fix
the bug, because
these vpatches, without a corresponding seal, would simply be ignored.
a111: Logged on 2016-12-22 02:12
trinque: does it matter or not who signed?
a111: Logged on 2016-12-16 12:09 mircea_popescu: but i can see why
this is practically obnoxious.
a111: Logged on 2016-12-22 02:04 ben_vulpes: and sure, if .wot is
the empty set, return
true ;)
a111: Logged on 2016-12-22 01:59 phf: well, since collective reaction is "tis but a scratch" i have nothing else
to say, and will happily await mircea_popescu's unrate
mircea_popescu:
http://btcbase.org/log/2016-12-22#1587972 << it's altogether unclear where did you get
this "celebrated is its ability
to support a scientific dialog" ; got a link or something ? seems more
the case what was celebrated was its ability
to put a wrench
through imperial-style "scientific dialog" where "we dindu nuttin wrong".
☝︎ mircea_popescu: barely-compiling c isn't good enough for us. oh, no, mommy i need my special asm skirt for
THIS sql stuff!
ben_vulpes: it's a very special haskellian snowflake
that makes it so i don't have
to
think about
that so nyah
ben_vulpes: asciilifeform: i'm going
to need
that gas mask
ben_vulpes: they're barely-compiling c
tools in unix land
mircea_popescu: in practice i'm in
two waters about it, wouldn't be surprised if
the whole
thing catches fire.
ben_vulpes: that is a very liberal use of
the acronym
mircea_popescu: in
theory
the above should be uberefficient sqling of a blockchain
ben_vulpes: surely a symlink wouldn't eat more
than a few bytes now would it
mircea_popescu: profile seek
times, whatevs reasonable
testage once
the artefact's prepared.
mircea_popescu: aaaanyway. mimi's in a fine position
to
try
this out, attach old drive, write script
to spit
them out as per schema, see what occurs.
mircea_popescu: obv you can't store infinite data in finite hard drive, i came
to
terms with
that
ben_vulpes: fill, sure. but an /infinite number/
thereof?
mircea_popescu: more importantly, why is
this not ~exactly~ like saying "doh, of course you can't fill a drive with
text files". dude what.
ben_vulpes: because every open source anything i've ever
touched has failed in precisely
these sorts of extreme use cases
mircea_popescu: afaik
this is entirely unwalked ground, for all
the foss bs about million eyes.
mircea_popescu: and b) wtf happens once you load a drive with
this nuttery.
mircea_popescu: ok so a) gotta see if ext2 / ext4 CAN EVEN HOLD
this many symlinks. like, at all.
☟︎ ben_vulpes: sure yeah, i remember
the design pretty well
mircea_popescu: this way you don't actually have
to ~index~ anything, if you wish
to see where
txn 1234567890 was included in a block, you go
to /12/34/56/7890 which points
to block x
☟︎☟︎ mircea_popescu: you'd end up with (as proposed
there) a directory
tree like 20 layers deep, with
thousands of symlinks in each.
a111: Logged on 2016-06-13 23:48 mircea_popescu: and
there's symlinks if anyone wants
to alias.
ben_vulpes: chiptune christmas, if you want
to give any olds a migrane