►
From YouTube: Ethereum Core Devs Meeting #29 [12/01/17]
Description
A
A
B
A
C
B
C
So
I've
been
also
working
on
a
big
non-fossil:
that's
the
big
num
implementations
of
go
and
host
and
C++,
and
so
on.
So
it's
very
important
that
these
are
all
aligned
that
aren't
any
any
differences
and
computations.
So
I've
been
working
on
that
and
next
up,
as
we've
been
talking
about
smart
contracts
fussing
but
on
any
clear
plans.
Yet
about
that
and
I've
been
talking
with
marketing
about
fastening
network
stack
of
gas,
because
if
there
wouldn't
be
any
portman
and
that's
what
stack
here
that
could
lead
to
denial
of
service.
That
would
be
a
problem.
A
C
A
D
D
So
I
was
thinking
to
implement
the
final
test
that
we
still
don't
have
in
a
general
form
is
a
transaction
test.
So
as
we
planning
to
create
change
changes
to
the
transaction
protocol-
and
we
have
many
Forks
so
I
think
we
need
that
general
transaction
test
a
one
source
while
describing
the
test
and
this
source
file
produced
a
general
transaction
test
which
contains
the
result
on
every
fork.
So
you
could
just
take
the
LPO
such
transaction
and
there
is
a
accepted
or
ejected,
and
that
would
be
a
new
transaction
test.
D
Sorry,
I
selected
new
categories
like
tests
for
each
individual
field,
so
I
think
it
will
be
easier
to
find
like
data
tests,
while
your
tests
and
other
fields
also
I
think
if,
in
the
future
we
and
decide
to
change
amount
of
fields
or
transaction
folder,
it
will
be
easier
using
in
general
transaction
test.
As
you
just
read,
there
are
LP,
so
that's
what
I
was
working
on
and
I
think
it
will
be
life.
C
A
Great
thanks
for
the
update
agenda
item
number
two.
So
yuuichi
had
a
an
idea
to
add:
CC
I
think
it's
CC,
zero
notices,
2e
ip's,
that's
where
you
say:
copyright
and
related
rights
or
waived
vyas,
Creative,
Commons,
0,
so
I
think
that's
a
good
idea!
Is
anyone
in
the
room
familiar
with
that
enough
to
know
if
we
can
put
that
on
previous
VIPs
automatically
or
is
that
something
that
we
need
to
request?
From
the
author
of
the
EIP
for
EG
IP?
We
want
to
add
that
to
I.
E
A
E
So
yes,
for
the
incoming
once
I
would
be
cool.
I
did
copyright
wave,
otherwise
I
do
not
much
and
then
food
there
are
some
eats
that
are
already
a
must.
Without
these
copyright,
waivers
I
will
try
to
get
them
the
first
one.
Is
it
number
one
courses-
and
you
know
Egypt-
should
contain
note
number
whatever
this
one
doesn't
contain
in
this
copyright
waiver.
For
now,
I
had
some
education
about
this
and
then
proceed
from
there
sounds.
A
E
A
A
A
So
the
next
item,
Ben
Edgington,
who
works
on
or
works
on,
I
think
the
Pegasus
team,
the
low-level
protocol
development
team
at
consensus,
the
blockchain
studio
commented
a
couple
days
ago
and
said:
is
it
a
reasonable
place
to
discuss
and
add
to
the
agenda
of
this
concern
about
the
github
yellow
paper?
So
if
you
click
on
the
like
link
on
4,
it
gives
the
full
picture,
but
I'll
just
read
through
this
real,
quick
and
then
we'll
we'll
go
over
it
step
by
step.
A
Pardon
me,
so
the
main
question
is:
does
it
remain
the
case
that
the
yellow
paper
is
intended
to
be
a
theorems,
formal
specification?
If,
yes,
then
it
should
be
noted
that
the
yellow
paper
is
currently
significantly
out
of
date.
So
what's
the
plan
to
bring
it
up
to
date,
post
Byzantium?
What's
the
process
for
ownership
and
maintaining
that
going
forward?
A
Should
it
be
made
part
of
the
EIP
finalization
process
and
in
order
to
facilitate
the
finalization
process,
should
ii
IPS
be
required
to
be
included
in
a
pull
request
for
the
yellow
paper,
if
relevant
before
they
are
accepted
and
consensus
failures
are
discovered,
like
during
fuzz
testing.
How
do
we
make
sure
that
the
yellow
paper
is
updated
where
necessary,
and
then,
if
the?
A
If
it's
not
the
case
that
the
yellow
paper
does
not
continue
to
be
a
theorems
formal
specification,
what
replaces
it
and
is
it
worth
maintaining
as
a
resource
that
is
descriptive
rather
than
prescriptive?
If
so,
the
above
questions
will
still
stand.
So
this
is
a
bigger
one,
but
I
think
this
is
important.
It
has
a
lot
of
good
questions
in
it.
A
So
let's
go
through
and
just
kind
of
generally
talk
about
our
feelings
on
the
yellow
paper,
because
I
don't
think
we
can
answer
everything
in
either
one
or
subsection
1
or
subsection
2,
without
first
kind
of
going
over
what
the
current
state
is
and
what
we
feel
about
that.
So
I
can
kind
of
give
a
little
bit
of
information
on
the
current
state
and
then
you
know,
eg
can
also
provide
some
context
because
he
has
been
working
a
lot
on
the
yellow
paper
via
pull
request.
So
right
now
the
yellow
paper
is
out
of
date.
A
It's
if
I
understand
correctly
I,
don't
even
think
that
it's
homestead
compatible
completely.
There
are
a
few
mistakes
there.
The
process
and
ownership
for
maintaining
the
yellow
paper
currently
is
Gavin
wood,
accepting
pull
requests
and
before
we
go
into
whether
it
should
be
part
of
the
EIP
finalization
process.
There
is
what
some
people
consider
to
be
an
issue
that
the
yellow
paper
does
not
currently
have
a
copyright,
and
because
of
that,
depending
on
jurisdiction.
A
That
leaves
open
the
possibility
that
the
author
of
the
yellow
paper,
which
is
Gavin
wood,
could
potentially
use
the
fact
that
there's
not
a
copyright
notice
against
people
wanting
to
make
changes
or
copy
the
yellow
paper
and
put
it
on
their
website
so
and
when
I
say
make
changes.
What
I
mean
is
make
changes
and
then
paste
them
to,
for
instance,
another
repository
or
to
a
web
site
that
hosts
someone
else's
version
of
the
yellow
paper.
That's
a
copy
so
that
that
kind
of
brings
up
the
question
of
where
to
go
from
there.
A
But
if
the
yellow
paper
does
get
sorted
out
as
far
as
copyright
and
other
issues,
some
people
might
have
I.
Think
it's
a
good
idea
for
it
to
be
part
of
the
EIP
acceptance
process
were
relevant
personally
and
also
updated
if
those
testing
finds
problems.
So,
just
to
summarize
what
I
just
said,
the
yellow
paper
is
out
of
date.
A
The
ownership
and
maintaining
process
currently
is
that
Gavin
wood
maintains
the
yellow
paper
and
accepts
pull
requests
from
other
community
members,
and
there
are
some
concerns
about
copyright,
so
you're
ichi
you've
talked
a
lot
about
the
yellow
paper
in
the
past.
Did
you
have
any
comments
about
that?.
F
E
Think.
Whoever
who
believes
that
have
made
a
huge
contribution
to
the
Aero
paper
can
have
a
veto,
at
least
to
try
modifications
or
whatever
I
think
in
the
current
status
is
very
dangerous
to
get
off
paper
modifications
before
protocol
changes
or
if
we
put
the
yellow
paper
in
any
kind
of
formal
procedure,
then
some
beetles
write
little
decisive
right
to
the
copyright
holders
of
the
yellow
paper,
whoever
they
are.
So
it's
kind
of
very
dangerous.
Okay,.
E
G
Listed
as
contributors
video,
no
paper
they're,
all
on
the
github
teacher
computers,
look
I,
don't
know.
If
there's
other
contributors
we
have
to
consider,
but
this
would
really
just
involve
reaching
out
the
it's
29
people
and
asking
them
for
written
consent.
So
that's
not
a
nun
thing,
but
I
will
point
out
that
somebody
should
reach
out
to
Gavin
ahead
of
time
and
get
prior
authorization
from
him,
because
previously,
when
there
was
an
attempt
to
realize
some
CBP
etherium,
he
was
the
one
who
blocked
that.
A
A
A
B
Think
it's
worth
talking
about
so
there's
a
couple
of
things.
The
yellow
paper
doesn't
attempt
to
be
a
complete
speak
of
of
everything
necessary
to
build
a
client.
So
it's
it
only
reflects
the
current
state
of
aetherium.
So
if
you
want
to
build
something,
they
can
sync
the
chain
before
the
current
hard
fork,
then
you're
out
of
luck.
You
need
to
get
out
of
band
information
and
there's
actually
no
formal
specification
for
that
information
other
than
older
versions
of
the
yellow
paper.
And
even
then
you
don't
know
the
for
block
numbers
and
so
on.
B
The
wire
protocol
and
other
critical
things
about
building
the
clients
are
also
out
of
scope
and
I
know.
A
lot
of
people
myself
included,
feel
that
the
yellow
paper
is
very
obscurantist
and
and
difficult
to
read.
So,
for
my
part,
I
would
love
to
see
an
effort
for
a
formal
speak
for
aetherium
other
than
the
yard
paper,
and
for
there
to
become
the
canonical
specification.
A
E
A
E
On
the
menu
the
markdown
files
specification,
actually,
the
markdown
files
contain
code
snippets
and
these
code
snippets
run
our
test
cases,
but
the
overall
document
is
marketing
document,
so
the
format
is
I
think
the
intention
is
to
make
it
as
a
specification
that
can
be
read
from
the
beginning
to
end
content
wise
these
English
sentences
that
sometimes
like
comments
to
the
code,
but
this
can
be
improved.
I
think
under
this
one
is
licensed
to
property.
B
A
H
H
H
A
Yeah
so
I
think
AIP
should
be
included
in
whatever
formal
specification
ends
up
being
the
one.
That's
mainly
used,
or
maybe
even
you
know
in
any
like
at
least
one
of
the
specifications
it
kind
of
depends
on
how
strict
we
want
to
be
at
the
EIP
process.
So
we've
kind
of
answered
the
question
about
what
could
replace
the
formal
specification
of
aetherium
and
if
it's
worth
maintaining
as
a
descriptive
resource
rather
than
prescriptive
Nick.
You
kind
of
touched
on
that
to
say
that
you
think
that,
because
it's
not
up-to-date,
you
find
it
more
descriptive.
G
Mathematical
'not,
just
mathematical
notation,
but
but
what
it
really
comes
down
to
is
the
ability
to
turn
specs
into
that
mathematical
notation,
which
is
something
that
well
I'm,
not
an
expert
on
and
I
got
a
math
degree,
so
I
suspect.
If
we
went
with
the
yellow
paper,
we
would
have
a
much
harder
time
keeping
it
up
to
date.
Yes,
there's
a
limited
number
of
people
who
are
going
to
be
capable
of
actually
contributing
the
poor
request
to
update.
A
A
Kvm
is
gonna
continue
to
be
developed
as
my
understanding,
and
so,
if
that's
the
case,
then
we
may
not
even
need
to
you
know,
do
anything
and
we
can
just
see
where
the
community
and
where
the
developers
start
picking
the
standard
from
so
it
might
be
just
something
where
consensus
kind
of
wins
out.
I
I
feel.
G
Maybe
you
can
comment
on
this
I'm
worried
to
be
able
to
where
we
have
a
decision
right
now.
As
to
you
know,
ok
KBM
is
gonna,
be
the
standard
that
the
yellow
paper
is
going
to
be
the
standard
or
or
whatever
is
this
something
that
could
fall
under,
like
dev
grants
to
essentially
fund
some
entity
going
ahead
and
getting
this
up
to
date,
because
that's
really
one
of
the
big
problems.
Here's
is
getting
us
back
up
to
date,
all
right,
I,
guess,
yeah.
A
And
cinema's
ation
is
definitely
important,
so
for
those
unfamiliar
dev
grants
is
a
program
that
the
etherium
foundation
ran.
That
gives
grant
money
to
projects
that
help
the
underlying
protocols
and
as
and
sometimes
projects
in
the
ecosystem,
so
I
believe
there's
been
two
runs
of
dev
grants.
The
latest
one
was
completed
in
2016
and
right
now,
dev
grants
is
currently
not
up.
There
are
discussions
to
bring
it
back
and
to
have
a
new
class
of
dev
grants
recipients,
but
as
of
right
now
there
are
no
timelines
or
any
other
information
for
that.
B
I'd
also
be
curious
to
know
whether
there's
any
way
to
run
kV
and
against
the
state
tests,
because
we
have
a
very
comprehensive
set
of
state
tests,
but
no
real
proof
that
they
actually
line
up
with
any
sort
of
specification.
And
it
would
be
good
certainly
for
writing
future
tests
as
well
to
know
whether
they
they
line
up
with
the
fullness.
G
A
A
E
A
E
A
A
Cool
yeah,
so
one
of
us
will
reach
out
to
Evan.
That's
not
an
issue.
Any
other
comments
on
this.
Next
steps
are
basically
gonna,
be
reach
out
to
Gavin
reach
out
to
Evan
and
get
them
on
one
of
the
cord
to
have
calls
to
discuss
this
further
Dimitri.
Did
you
have
a
comment
on
on
the
EVM
code,
description.
D
A
A
H
There
was
idea
to
actually
move
them
to
a
separate
project
that
can
somehow
share
some
of
the
components
that
are
not
critical
to
the
Hume
note
and
in
the
description
there
is
this
in
the
agenda.
There's
a
description
and
github
issue
that
there
is
some
discussion
on
that
and
also
a
prototype
is
done
in
Python.
So
the
first
feature
I
would
like
to
have
the
race.
H
Http
server
for
for
RPC
requests,
so
we
don't
have
to
include
HTTP
server
in
every
every
client
and
there
are
some
other
ideas.
What
can
be
added
there
later
on?
So
if
anyone
is
interested
in
in
having
something
like
that-
or
it
also
might
be
a
candidate
for
dev
grants,
as
dev
guns
were
mentioned
in
this
meeting
already,
but
I
just
wanted
to
to
share
this
and
that's
it.
H
So
yeah
the
other
components
actually
question:
can
we
get
up
to
date
with
the
current
state
of
the
art,
with
some
subsystems
we
already
have,
and
sometimes
answer
is
here
would
like
to,
and
sometimes
answer
is
no.
We
don't
have
resources
at
the
moment
to
actually
upgrade
something
like
whisper
for
example.
H
H
The
other
thing
I
was
considering
actually
account
management,
but
that's
that's
probably
a
bit
more
complex,
so
we
actually
don't
have
to.
If
you
have
the
proxy
that
handles
RPC,
you
don't
actually
have
to
have
account
management
inside
the
node,
because
the
proxy
can
actually
sign
transactions
for
you
and
somehow
change
personal
or
peace
request
to
draw
RPC
request.
So
this
is
also
the
component
we
are
considering
that
can
be
actually
shared
between
some
of
the
clients.
H
The
idea
is
to
actually
have
support
for,
if
stats,
because
what
we
have
currently
some
note
model,
nodejs
model
that
it's
not
maintained
anymore
and
it's
actually
actually
hard
to
get
it
work.
So
I
know
the
go,
have
the
implementation
of
that.
So
if
we
can
show
that
implementation
and
also
move
it
to
this
shirt
like
to
link
product
that
also
can
be,
can
be
used
easily
by
different
different
notes.
F
So
I
think
it's
a
great
idea
and
I
remember
that,
like
when
I
was
working
on
this,
it
was
walking
over
here
to
do
integrating
HTTP
and
cross-platform
support,
for
it
was
like
a
great
high
school
and
like
the
HPD,
supports
itself
and
brought
a
lot
of
unnecessary
dependencies.
So
I
think
that's
a
great
idea
for
learning
the
proxy
I
see
one
issue
with
this
approach.
F
So
apparently
by
recalled
correctly
correctly,
there
is
no
way
to
sign
arrow
message
using
your
private
key
because
for
the
artist
occurs,
basically,
if
sign
is
whenever
you
call,
if
sign,
is
pretending
some
arbitrary
message
before
the
data
that
you
want
to
sign
so
I
see
this
this.
This
might
be
the
only
the
only
issue
with
the
proxy,
but
apart
from
that,
I
feel
like
it.
It's
a
great
improvement
to
see
puss-puss
fine.
G
I'm,
not
exactly
sure,
I
have
a
complete
thought
here,
but
related
to
the
issue
with
design.
I
think
that
that
just
highlights
I
don't
know
I
like
to
see
nodes,
moving
further
away
from
account
management
and
not
necessarily
having
account
management
features
as
an
expected
default,
because
well,
that's
just
one
extra
bit
of
security
overhead
that
puts
on
on
creators
of
area
nodes
and
account
management
is
a
really
complex
thing
and
even
kinda
standardized
account
management
is
complex
because
it's
well
we're
still
in
the
early
stages
and
there's
a
lot
of
experimenting
to
be
done.
I
I
agree,
my
my
after
that,
so
we're
planning
to
move
account
management
away
from
gap
and
a
couple
of
steps
to
do
it
slowly,
but
I
have
been
implementing
a
separate
account
management
tool
which
of
the
KPI,
which
doesn't
take
any
passwords
at
all.
Instead
of
isness
only
for
user
videos
to
make
me
interact
with
trusted,
you
I
have
this
in
our
house
and
a
remote
node
can
submit
transactions
or
data
and
to
decide
we
use
fine,
yeah
I
think
it's
works
kind
of
like
Pericles
already
done,
but
it's
easy.
A
G
A
F
Okay,
maybe
all
starts
with
free,
with
with
the
party
code
base
itself.
So
currently
we
we
are
working
on
the
release.
One
point
on:
it's
not
going
to
have
any
major
features.
Any
major
new
features
are
so
we
are
mostly
working
on
stabilizing
stuff,
we've
added,
multiple
new
features
which
were
not
well
acid
and
they
and
the
number
of
issues
in
our
and
our
kids
have
significantly
increased.
So
we
are
working
on
those
issues
and
we
are
also
working
on
improving
our
database
layer.
F
So
hopefully,
31:39
will
have
a
new
database
for
accounts
and
code,
stop
and
codes
so
creating,
for
example,
snapshots
or
just
costing.
Transactions
inside
VM
should
be
sitting
significantly
faster,
because
the
weed
operations
from
there
from
a
disk
will
be
will
be
absolute
and
times
faster
than
they
currently
are.
And
apart
from
that,
we
are
also
working
on
party
multi-state
issues,
so
we
are
currently
most
internally
preparing
a
list
of
different
approaches.
F
How
we
could
address
the
problem-
and
we
would
like
to
add
this
thing
to
the
discussion
about
the
proposals
that
we
will
present
to
the
agenda
on
the
next
card.
That's
a
thing,
and
we
we
are
not.
You
know
what
we
will
not
try
to
push
forward
any
any
of
those
proposals.
We
will
just
try
it
we'll.
Just
we
just
want
to
discuss,
discuss
them
with
with
that
community
and
with
you,
and
you
know,
see
where
we
can
go
from
there.
F
F
I
H
Yeah,
it's
it's
quite
quiet
here,
as
well
still
in
the
post,
Mexico
mode,
and
so
we
mostly
cleaning
up
all
stuff
related
to
to
Byzantium
for
hard
work.
Some
reviewing
some
old
issues
and
also
usually
updating
the
IPs
and
me
and
Andre,
are
preparing
some
we're
having
some
discussion
about
some
some
roadmap
for
for
bigger
steps
in
in
the
CPP
atrium
team,
and
that
doesn't
include
that
at
tests,
work
done
by
by
Dmitry.
D
Speaking
of
which,
so
you
asked
about
the
function
and
if
you
get
single
test
without
checking
a
fielder,
so
it's
now
done,
you
could
just
type
test
each
and
specify
single
test.
Then
we
use
this
for
a
test
case,
name
as
usual,
but
then
you
could
provide
another
arm.
Is
the
pastor,
a
JSON
file
and
doing
it?
This
way
will
not
check
its
existing
of
a
fiddler
file.
A
J
J
B
Yeah
I
assume
the
truck
in
question
is
exhorting
the
path
of
the
current
and
the
result
of
that
is
that
the
you
can't
support
the
Felice
protocol
because
you
can't
win
queried
for
a
particular.
Try
knows
about
the
context.
You
don't
know
what
keys
look
it
up,
I
and
galloping
was
noting.
Was
that
actually
it's
with
very
high
probability,
the
try
as
graph
shaped
because
of
the
nature
of
the
the
keys
are
all
cryptographically
generated.
I
can
go
into
the
more
detail,
proof
which
ones
interested.
A
Absolutely
are
you
on?
Are
you
on
guitar
Anton?
Sorry,
are
you
on
get
ur
chat?
Yes,
I'm
there,
ok
great!
If
you
could
post
in
the
Alcorn
dev
channel
and
just
put
your
get
ur
name
and
the
Nick
can
know
how
to
reach
you
mm-hmm.
Okay,
thank
you.
So
much
and
yeah
sounds
cool
that
you're
doing
interoperability
there.
K
G
So
we
largely
just
finished
up
a
big
github
migration.
Moving
about
ten
repositories
from
my
personal
github
under
the
nation
get
her
lots
of
maintenance.
Work.
I
haven't
published
information
about
this
yet,
but
the
repositories
that
were
previously
named
aetherium,
something
or
not-
and
now
all
prefixed
with
eath
something
to
match,
essentially
matching
the
repository
name,
the
Python
library
name
and
the
actual
module
import
name,
because
there
was
some
confusing
stuff
going
on
there.
A
G
Find
a
far
away,
update
I
know
that
there
is
stuff
moving
along
with
Casper
research
with
Casper
test
net
I.
Don't
know
exact
specifics,
but
I
know
that
there's
been
some
trial
runs
and
running
the
actual
kind
of
Casper
test
net
using
pi
e
theorem
successfully.
So
that's
really
cool
to
see
happening
in
a
in
a
real
environment.
A
A
What
is
it?
Oh,
it's
it's
Vlad's,
Casper
version,
CBC,
maybe
correct
by
construction.
That's
right,
so
Casper,
correct
by
construction,
has
a
repository
now
and
some
of
the
first
coding
implementation
of
that
of
that
design.
So,
if
anyone's
interested,
the
repositories
I
believe
are
under
some
name
for
Casper
and
the
etherium
repos.
A
I
So
I'd
like
to
mention
that
there
is
a
downside,
Keith's
been
working
on
a
c-sharp
client
for
areum
and
it's
testing
almost
all
state
tests
and
VMs.
And
yes,
he
reckons
it's
about
a
couple
of
more
months
of
work
to
have
it
kind
of
complete.
At
this
point
it's
build
a
private
repository,
but
if
anyone's
interested
reach
out
to
me
and
I'll
forward
you
to
him
also
on
the
curator,
the
executor
name,
if
there's
an
interesting
development
in
either
solidity
or
even.
K
K
The
biggest
one
really
is
the
ABI
and
D
fully
implemented
ABI
decoding
and
encoding,
which
supports
trucks
and
everything
which
wasn't
supported
by
the
old
ABI,
encoder
and
decoder,
and
but
in
order
to
use
that
want
me
to
specify
an
experimental
flag
and
that
we
are
really
focusing
on
the
SMT
and
the
SMT
solar
in
in
holiday
for
next
release,
as
well
as
the
intermediate
language,
I
think
I
think
those
are
the
most
important
parts
of
it.
Solidity.
A
A
Okay,
good
meeting,
everybody
I
think
we
hit
exactly
an
hour
and
I've
already
made
an
issue
for
in
the
etherium
p.m.
repo
for
aetherium
coordinate
of
meeting
30.
So
if
you
already
have
in
mind
what
you
want
to
talk
about
in
two
weeks
on
the
15th,
please
put
it
in
there.
I
don't
have
the
live
stream
linked
up
yet
obviously,
but
I
do
have
four
agenda
topics
and
I'd
love
for
people
to
add
more
so
go
ahead
and
get
in
there
and
add
some
stuff
and
everybody
have
a
great
two
weeks.