549100+ entries in 0.361s
jurov: how else you want
to check relayed
transactions from other nodes?
mircea_popescu: if you check you check. if you don't check,
then don't check.
mircea_popescu: and i don't give a shit how convenient it is. let userland cut down
the correct implementation for
the sake of convenience.
mircea_popescu: ben_vulpes once a
transaction is proposed,
the software has
two options. 1) "oh ,
this looks like a
tx involving inputs i remember were valid" and 2) "these are
the inputs. check.
these are
their sources. check.
these are
their sources. check.
this is
the coinbase. check. ok."
ben_vulpes: to verify a new
transaction, one needs
the previous
transaction in hand, in entirety.
jurov: and
to do it you need
to find all
the inputs, if i understand ben_vulpes correctly
mircea_popescu: <mircea_popescu> bitcoind should broadcast verifiable
txn submitted
to it. <<
ben_vulpes: <mircea_popescu>
txn are verifiable aren't
they. << not really.
transaction only includes
txid of previous
transaction which itself has
the pubkey required for validation.
jurov: so, it will allow me
to relay 1000 doublespend
transactions with different destination addresses, because it can't check if
the source isn't spent already?
kakobrekla: first runs on airgap, second on my server in dc and
third on my desktop
kakobrekla: imho 3 parts. private key handler, bc api and a
thing in between
jurov: okay. but
then bitcoin will not be able whether relayed
transactions aren't completely bogus
mircea_popescu: like linux. i don't have
to get a new version of lindows
to get rid of fucking notepad
mircea_popescu: modular fucking design, so you doin't have
to change
the whole
thing every
time a part sucks [for your usecase]
ben_vulpes: bitcoind-wallet should generate and sign
transactions.
mircea_popescu: bitcoind should broadcast verifiable
txn submitted
to it.
mircea_popescu: trying
to invent a "this is
the right way" because we're
the power rangers isn't
the way
mircea_popescu: there's no ready way
to know hat
the user will want
the index
to be.
jurov: just want
to clarify
jurov: yu also imply creatng useful
transaction index is
the job of
the user
mircea_popescu: bitcoind should put
the blocks into an accessible database.
the querying of
that
thing is
the job of
the user.
jurov: okay, so for one
to have complete stack, run something like electrum server
that maintains
the
tx index and connects only
to our node
ben_vulpes: b) bitcoind *should* be able
to return
transactions with unspent outputs
kakobrekla: every
time address is added
to wallet? < if
this is your common practice i
think you are doing something wrong
ben_vulpes: a) bitcoind should not return arbitrary
transactions by id
ben_vulpes: so i made
two claims, let's
treat
them seperately
mircea_popescu: suppose you somehow DO actually want 32mb blocks. how long does a rescan
take
then ?
mircea_popescu: design is not a consideratio nof "how long does it
take
to bale
the water out"
jurov: scouring whole blockchain every
time address is added
to wallet?
ben_vulpes: mircea_popescu: how long does a "rescan"
take
these days?
mircea_popescu: if every phone had
to maintain 6bn phone numbers ?
that's
terrabit memory on any phone ?
jurov: with bitcoin you need
to know your own phone numbers
mircea_popescu: notwithstanding
that in order
to call someone, having
the phone number is not a big deal. having a passing answer for "who's
this ?" is.
mircea_popescu: it's not unlike
the "idea"
to maintain a list of
the phone numbers of
the entire planet.
ben_vulpes: with
the utxo list in hand, a node could create a raw
transaction for signing by an offline node.
ben_vulpes: i
think
that maintaining
the utxo list is a good idea.
jurov: but you propose
to maintain only unspent ones? or none at all?
mircea_popescu: Cubs in
the Islamic State,
the generation
that will conquer Rome in shā Allah << boys are wearing
the stupid face veil
thing
too now ?!
ben_vulpes: correct me if i'm wrong, but address balance is
the sum of unspent
transactions
to
that public key.
ben_vulpes: it's a much more
tractable problem
to find
transactions relevant
to a given address out of
that list
than it is
to walk
the blockchain for same.
ben_vulpes: if
the node were
to maintain
the list of unspent
transaction outputs, it would also have a list of
the
TXID's with unspent outputs.
jurov: some means
to get individual
txs is needed
jurov: well, but if you attach a wallet
to
the node, it will need
to crawl whole blockchain
to get its
transactions?
mircea_popescu: "a CPU profile of
the
time spent processing a 32 MB block by a full node is dominated by ECDSA signature verification, meaning
that with
the current infrastructure and computer hardware, scaling above 300
tps would require a clustered full node where ECDSA signature checking is load balanced across multiple machines."
jurov: ben_vulpes: what is meant by "cough up
transaction"?
mircea_popescu: ben_vulpes: asciilifeform: i don't know
that
the
thing needs
to cough up *any*
transaction on demand. << one can make his own emacs scripts do whatever one wants. notthe job of emacs
to provide a preloaded button for arbitrary
tasks.
mircea_popescu: davout: "you want
to do some btc related accounting? well, fuck you,
the decimals after
the second are a $1.99 in-app purchase" << seriously ? like a power chord
that's not six inches long ?
ben_vulpes: not
the best english, but what american knows how
to write well anyways?
mats: p sure he's referring
to
the
twitter isis brigade.
Naphex: <+mats> a man sees what he wants
to see
mats: a man sees what he wants
to see
ben_vulpes: when i made furniture, we went out of our way
to put cunts into
the woodworking.
mats: i was
thinking
that
ben_vulpes: "a single machine acting as a full node" << conformal has yet
to answer any questions re: what
the specs of
this "machine" were.
ben_vulpes: anything in particular about
that, mats?
ben_vulpes: my reasoning being
that iterating
through
the blocks is adequate
to get complex information out of
the
thing, and
that producing unspent outputs is adequate for "bitcoinating" - eg, producing and signing
transactions.
ben_vulpes: asciilifeform: i don't know
that
the
thing needs
to cough up *any*
transaction on demand.
assbot: LiveLeak.com - 21-01-2015 Peshmerga hitting ISIS on major Offensive
to reach Mosul, Good use of MILAN ATGM ... (
http://bit.ly/1D44Uy2 )
ben_vulpes: oh as if its even a question at
this point
[]bot: Bet placed: 1 BTC for Yes on "EUR/USD Parity in 2015"
http://bitbet.us/bet/1111/ Odds: 95(Y):5(N) by coin, 95(Y):5(N) by weight.
Total bet: 1.3 BTC. Current weight: 99,494.
davout: "you want
to do some btc related accounting? well, fuck you,
the decimals after
the second are a $1.99 in-app purchase"
davout: not
that rounding isn't already an error in itself
ben_vulpes: davout: yeah let me
tell you about
the manual checks in
the billing process
to catch rounding errors
ben_vulpes: <jurov> [] hanbot rather a pickup line, maybe. "wanna write my eulogy?" << i have actually used
this one
diametric: also, nubbins`, your dedication
to
this guy is amazing.
diametric: not
that it matters, his shit is already obviously lasered.
diametric: asciilifeform: in context of
the wood dong guy, he could vary height with
the intensity as well, but it becomes really, really apparently its lasered because
the edges are always charred
diametric: asciilifeform: its a manual 3 axis,
though i
think it would
take little effort
to make it a full 3 axis.
jurov: never late
to plan for
the occassion, isn't it?
mircea_popescu: hanbot if you eventually decide
to sue for creation of a sexually harassful working environment, pls
to direct it at jurov
mircea_popescu: omfg
that's not what
that article said, and wtf spouse!
hanbot: or "spouse", for
that matter ;p
hanbot: well i remember
the
trilema post in which
the woman's duty
to write a eulogy, suck cock, and wash socks was outlined
jurov: hanbot remember
the
trilema post
that spouse of deceased man is useless except
to write
the obituary?
hanbot: <jurov> mircea's looking forward
to hanbot's obituary << lol what are you on about?