326700+ entries in 0.186s
mircea_popescu: check
that out, "TradeBlock serves financial institutions with execution and analysis
tools
that capitalize on
the potential of blockchain
technologies."
mircea_popescu: built on
the basis of what other nodes advertise as
txn
mircea_popescu: sturles from my own nodes' estimation of what
the "mempool" is currently.
sturles: Where is
this number from? Mine is:
mircea_popescu: for
the curious, mempool was as high as 80mb during february, but recently it's ~8 or so.
mircea_popescu: yeah, mempool is low now for some reason, across
the board.
mircea_popescu: (obviously because no protocol
there is no actual definition of "mempool size" other
than rough consensus, but wehartevers)
mircea_popescu: in any case, at
the
time
this was discussed in
the logs,
the miner priority was in
the 9th decile, about 16% of
the global mempool.
☟︎ mircea_popescu: yeah but for practical reasons doesn't also vary
too much, or you get drowned.
kakobrekla: anyway, sturles is giving us his current mempoolminpriority, which doesnt mean it was
the same a week ago.
sturles: The
transaction has one 3 month old input of 15.898 BTC.
This alone may be enough
to push
the priority high enough.
kakobrekla: and just sum
the discrete inputs ages ?
sturles: OK. I don't know anything about
the
tx in question. How old were
the inputs?
sturles: 0 fee
tx are
thrown out if
the priority is lower
than what is required. Currently 142186611.6926576.
mircea_popescu: if you recall kakobrekla at
the
time it was merely broadcast and "not yet included" (in
the public version of
the blockchain), it had a very low priority.
mircea_popescu: right. by
this calculation,
the
tx in question would not have remained in your pool.
sturles: Standard priority calculations. size of input * age of input for all inputs,
then divided by
the size of
the
transaction.
sturles: Which means
that every
tx with priority below 142186611.6926576 will be
thrown out of
the priority pool, and into
the fee based pool.
mircea_popescu: sturles if you keep 0fee
txn without
time limit, you are necessarily running on a machine with infinite ram.
sturles: It doesn't matter if A2 has higher or lower priority
than A1. As long as A1 is in my mempool, a
tx spending any of A1's inputs will be rejected.
sturles: I don't have a
time limit.
sturles: I don't keep
them all. Only
the ones with
the highest priority.-
kakobrekla: >If you are proposing
that you are actually running a node which keeps all 0-fee
txn for a week plus, I very much would like
to know what machine are you running, and when's
the last
time you realised what an incredible DoS mechanism
this is.
kakobrekla: i asked him
to join
to answer
the questions from qntra
mircea_popescu: you have enough
trust, whenever you want
to speak here you should pm assbot "!up" and decrypt
the dpaste.
mircea_popescu: no argument
that it ~could~ be a so and so node. all sorts of
things could be all sorts of other
things. like
the moon landing photos could be fake, and so on.
kakobrekla: i find
the comment from 'sturle' much more interesting.
mircea_popescu: who knew how fucking deep in
the rules of
the world
this principle is baked
mircea_popescu: as usual - centralist power is
the enemy of free commerce.
mircea_popescu: of course large payments remain a problem - not so many double digit btc moving around
to an address with
the following 4 chars.
mircea_popescu: as
the commenter says,
this does nothing for bitcoin, merely for bitbet.
mircea_popescu: and yes, publish a sha of
the whole list as-is somewhere on page also. so people can
then verify at
tyhe end.
mircea_popescu: say paid out bets are published as now, but proposed bets only show first 4 chars of
the address, and only first
two digits of
the payment. except if under 0.01 it's just replaced with D
mircea_popescu: listen kakobrekla can we modify
the way bitbet works
to actually satisfy
this ? something like
mircea_popescu: so basically, yeah bla bla, miners can force "either doublepay or never pay" dilemma, but ~not on everyone~. only on derps who, like mp, actually publish
the whole story.
mircea_popescu: no but on second pass
this is actually correct, isn't it. ~I~ have been breaking
the protocol with
that design, indiscutably.
assbot: Logged on 03-03-2016 15:19:06; mircea_popescu: jurov is it actually
the js interpreter
that
takes up
the bulk of
that ?
mircea_popescu: which one of you has
the miserably poor
taste of impersonating
the dead woman ?
assbot: Logged on 03-03-2016 15:17:37; jurov: just for
the record,
there are several projects built on
top of llvm (cling,clang,clasp) promising C++ interpretation and easy interop with lisp. i
tried
to build and use
these, not one succeeded and
they are so behemoth so any analysis of
the problem was out of
the question
jurov: re: duplicate bugreport.. perhaps i envied you being
target of lizard hitler pranks, and imagined myself finally being
targeted
mircea_popescu: romanian school has a whole curricula in literature dedicated
to making
the kiddies cry.
this is well and good and
the ~only way.
mircea_popescu: let
them cry into
their fiction books like normal people.
assbot: Logged on 03-03-2016 15:09:58; mircea_popescu: since you have
the profiler right
there (do you ?) mind sharing like,
the
top
three
things on
the stack ?
assbot: Logged on 03-03-2016 15:07:13; jurov: asciilifeform: wake me up if you manage
to compile chromium in 3G ram.
the linker eats 10G. i have 12G here and must log out and
terminate most processes and
then emerge it.
mircea_popescu: ftr anacam was as close
to driving culture as
the
times and potus aspire
to be and occasionally manage.
mircea_popescu: jurov dude you
trolled me ? i mistook eul-
to be a function defined in eulora space pinged DianaComan about it lmao
jurov: such backdoor would be very convenied, dontcha
think
jurov: i came around
this in course of embedding lisp in eulora client
mircea_popescu: like it or not,
the
tendons won't give unless you get physical. you can
talk yourself hoarse, it dun do anything.
mircea_popescu: chick's
tougher
than coffin nails, right in
there with
the she-marine. heck,
they still use beatings
to
this day
to loosen
the girlies up.
jurov: yes, surely all
the HTML5 and older(like DOM) functionality
mircea_popescu: jurov is it actually
the js interpreter
that
takes up
the bulk of
that ?
☟︎ assbot: Logged on 23-02-2016 08:51:15; mircea_popescu:
http://log.bitcoin-assets.com/?date=23-02-2016#1413406 << basically us "college" chumpatron still experimenting with "what's
the lowest qty of candy bar we give
the cattle
to keep getting
the sweet sweet usg funds for
them. like, out of
the 100k per capita we get in fed funny money, what's
the least we could dole out
to
the maggots in whose name
the whole scheme is run ? maybe -100`000 in paper and $
jurov: so it's prolly
the same with chromium's javascript engine
mircea_popescu: "you can have your own opinions, but we want
to move from
this situation where harvard & mit lied
to you about you being smart enough
to resolve any real word conondrum 'if you just got
the facts [as officially branded facts by harvard and mit]'"
to a much more economical "you can have your own opinions just as long as
they're what we say
they should be".
jurov: just for
the record,
there are several projects built on
top of llvm (cling,clang,clasp) promising C++ interpretation and easy interop with lisp. i
tried
to build and use
these, not one succeeded and
they are so behemoth so any analysis of
the problem was out of
the question
☟︎ mircea_popescu: sort-of
trying
to chinese cartel
the electoral process, pretty much.
jurov: and i did not profile it, just looked on
top output
mircea_popescu: you see jurov, google is next
to useless for any purpose
that involves having
taken algebra II. all it can do is help you build empathy.
mircea_popescu: ;;google largest
three items on chromium profile by ram
mircea_popescu: like in
that joke with
the drunk and
the samovar. you know it ?
assbot: Logged on 03-03-2016 06:50:55; phf: well,
that was a rabbit hole of progressive insanity
mircea_popescu: "Only mathematicians and some engineers actually use advanced math in
their day-to-day work, Hacker argues—even
the doctors, accountants, and coders of
the future shouldn’t have
to master abstract math
that
they’ll never need." [and could one day use
to find out some unsavory facts about
the best possible castle in
the world!"