220000+ entries in 0.14s
trinque: and now we're on
the n-th conversation fork
trinque: guy was using
the sign-message as an example for a smaller cut
than
the wallet
trinque: you gotta communicate more on-subj man;
this is exasperating.
davout: asciilifeform: no, was
this comparison ever contemplated?
trinque: I don't
think anyone brought his cock into
the matter
davout: it's in
the "bells and whistles" box, not
the "hot wire functionality" one
davout: asciilifeform: you're conflating signatures on
txes with signatures on arbitrary messages
davout: if you ~must~ verify one of
those sigs you pull up a
trb
that can ~fin~
davout: asciilifeform: we're
talking about verifymessage and signmessage
davout: maybe we can have
trb-classic
then
trollface.jpg
trinque: it'd probably be better
to be hollering at a patch
than all
this
davout: and i really don't understand how keeping stuff around is sane when
the functionality is an operator knob-turn away anyway
trinque: this ends up meaning slashing probably everything
that doesn't keep
the current network protocol running
davout: asciilifeform: well, maybe for forensic purpose you want
to open up an accounts-enabled legacy wallet? i don't know, sounds equally likely as "want
to check signature some derp made in weird ways during ancient
times"
trinque: there are goals at odds here.
to fix
the nightmare
that is current
trb, gotta start slashing
til you have something
that's able
to be comprehended
trinque: "nothing but
this particular arrangement of driftwood counts as actual reference
trb"
davout: asciilifeform: well, if you want
to "keep everything" you end up keeping everything
trinque: and eventually
there's a process of regrinding and so on
trinque: I don't
think it's a sin for davout
to go slashing and cauterizing in one direction, meanwhile reference client proceeds in another
davout: asciilifeform: with
this kind of reasoning we'd end up keeping
the "accounting" feature around
ben_vulpes: davout: consider #!/bin/bash and set -e
to make
the
thing die if any subprocesses return non-zero
davout: asciilifeform:
the
trb
tree has a "continuity-preserving" mission, not "current
trb official version"
trinque: asciilifeform: consider our conversation about
the openbsd patches in another castle
davout: what version of gnu-awk are asciilifeform,
trinque et al. using
to have a usable vdiff.sh ?
trinque: I am aware of
that, or what I said next would make no sense
thestringpuller: davout:
that's why I
think
the UTXO probing is interesting. Easier
to search ~2GB and ~100GB for
the information you want.
davout: asciilifeform: archaeologists can build a verifymessage-capable
trb, couldn't
they?
trinque: the former might help clear up
the latter
ben_vulpes: perhaps a low-dough point: vpatches ensure
that functionality will always be available, regardless of weight of "current" version
☟︎ trinque: one of which strips
the
thing bare; another which is
the reference
trinque: I don't see
that it'd be a
terrible sin
to have multiple branches descending from current
trb
davout: asciilifeform: my emulated mp would answer
that a dork
that signed
that way can't possibly sign something important
davout: asciilifeform: well for example,
the "remove signmessage and verifymessage" patch could very well be considered ready for production, it cuts something, not something anyone sane would actually depend on
thestringpuller: i'm using
the blockchain itself
to get
the UTXO for forming new
transaction when creating raw
TX with bitcoin,
this requires searching
teh blockchain or using ben_vulpes
tool
davout: thestringpuller: i'm not sure what you mean by "scraping
the UTXO" ?
davout: but appreciate
that i'm a fucking noob and i need
to have something
to get started, at least with
the
toolchain, not
that it has
to make official release until it's actually done you know
thestringpuller: davout: no. I'm just curious if
there is a way
to scrape UTXO's now, cause for my wallet
tron I'm scraping blocks
themselves. Def not O(n). (don't laugh)
davout: asciilifeform: a sane wallet would make it very easy
to avoid dust fragmentation
davout: thestringpuller: you're asking me what
trajectory i'm going
to
take at 200mph for
that
turn, i'm at
the point where i'm still wondering how
to
turn
the goddamn car on
davout: asciilifeform:
take 20 UTXOs spend
them in one go
to a single address, poof! UTXO set has shrunk
thestringpuller: is
there a way
to scrape
the UTXO set in
TRB or do you have
to do
that manually as of now?
davout: asciilifeform: yeah well, if we're going
to debate what "we don't have yet" we're not going
to get very far
davout: also
the UTXO set might very well decrease
davout: asciilifeform:
tbh i haven't measured it, but scanning 2gb of ram shouldn't really
take
that long
trinque: were you going
to interrogate
the utxos every block or something?
davout: trinque:
text files with private keys sounds like
the sane approach
to me
trinque: there was a decent
thread on how 'wallet' end up being 'arbitrary declared index of declared addresses'
davout: thestringpuller: UTXO set is ~2gb
tops, indexing might be nice but necessary
to scan for UTXOs
that match a given set of addresses, also
the wallet part can cache
them if
that particular wallet is
the only one able
to actually spend
those UTXIs
☟︎☟︎ thestringpuller: davout: wallet has keys, asks for unspent outputs
to bitcoin << don't you have
to be "watching" an address while indexing
to get
this info?
trinque: I can't find davout saying he refused
to provide such a
thing in his patch
trinque: this whole
thead could've been "I will not sign
the excision unless it brings also
the replacement wallet" ... "k" ... fin
trinque: and you're making it out
to be something he never argued
trinque: guy's entirely ready
to discuss
the
thing