►
From YouTube: NEAR Eng Sync 2020.10.16
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
C
A
A
Let's
first
go
through
last
week
updates
and
what
we
needed
to
do.
A
A
A
Three
trillion
instructions
and
it's
still
going
on
both
machines,
one
with
ram
one
with
ssd
disk.
A
A
A
A
So
it
might
affect
some
caching
and
other
metrics
and
it
does
have
roxdb
threads,
enable
plus
caching
enabled
so
it
will
be
more
realistic
than
what
we
have
within
a
param
estimator,
where
we
don't
have
a
cache
enabled,
and
we
don't
have
threads,
enabled.
A
And
for
runtime
yeah
pretty
much
nothing
blocks
the
upgrade,
except
for
maybe
we
want
to
do
this
as
a
feature,
but
it
doesn't
matter
way.
We
release
it
too,
so
we
can
still
start
working.
I
think,
we'll
face.
I
realized.
We
will
face
some
extra
problems,
such
as
pessimistic
gas,
so
we
may
need
to
disable
this
pessimistic
gas
computation
by.
A
Because
otherwise,
so
probably
need
the
background.
A
We
calculate
how
many
hubs
it
can
do
on
a
blockchain,
so
how
far
in
the
blocks
in
the
future,
it
will
go
considering
it
does.
One
function,
call
per
block,
for
example,
then,
comparing
how
much
gas
you
have
and
how
much
it
costs
to
do
this
function.
Call
we
compute
how
many
blocks
in
the
future
you
can
continue
executing
this
transaction.
A
A
Much
cheaper
function
calls
which
is
about,
I
think
20
times
cheaper
than
it
was
right
now
it
means
the
dabs
can
be
20
times
larger,
but
data
is
inside
even
cheaper.
It
means
you
can
do
a
round
trip
through
a
data
receipt
which
is
you
do
a
function,
call
and
dependency
on
yourself
with
another
function,
call
which
delays
you
by
three
blocks
instead
of
one
block
or
like
instead
of
two
blocks,
and
since
this
is
cheaper.
A
Well,
the
depth
goes
from
about
64
maximum
to
something
like
almost
4
400,
probably,
and
if
you
just
keep
the
current
coefficients
as
is
then
it
will
inflate
the
maximum
possible
gas
by
way
too
much.
So
maybe
you
should
just
go
to
a
regular
guess
max
increase
of
like
one
percent.
C
C
A
So
I
don't
know
if
you
still
want
to
burn
all
the
gas
it's
still
okay
and
they're
like
while
it's
a
good
attempt
that
we
wanted
to
do
this.
I
I
don't
know
if
we
should,
until
a
lot
of
stuff
is
fake,
so
like
this
is
way
more
invasive
change,
because
right
now,
like
every
single
deluxe
team,
like
every
single
command,
still
attaches
both
load
of
gas,
and
if
you
start
burning
it,
our
blocks
gonna
be
just
fully
empty.
C
C
B
A
So
we
can
address
it.
A
This
is
like
adjusting
refund
cost,
so.
B
If
we
have
sharding,
then
it
becomes
even
more
likely
for
this
to
happen.
B
B
B
A
But
this
sharding
does
not
address
anything
through
burning
gas
does
not
solve
it
anyway,
so
I'm
I'm
saying
that
it
doesn't.
It
does
very
real
short-term
damage
versus
like
it
doesn't
address
long-term
with
charging.
So,
okay,
yeah.
B
A
Yeah,
let's
see
if
we
can
so
for
grass
priority,
it
will
be
like
a
better
address
in
this
problem,
but
that
leads
to
like
other
issues
of
like
receipts,
so
you
need
to
figure
out
how
to
resolve
it
anyway,.
C
C
So
you
keep
in
this
progressive
the
prediction
of
the
gas
cost.
A
A
C
A
It's
gonna:
it
does
not
to
stop
the
congestion
just
to
pay
for
the
congestion.
A
So
if
you
create
a
receipt
without
a
congestion
that
is
helps
every
block,
then
this
is
the
amount
of
gas
you
need
if
your
blocks
are
fully
inflated
I'll
give
the
kind
of
full
more
than
50
percent,
and
we
bump
the
price
every
time
to
one
percent
up.
So
this
is
the
adjustment
that
it
supposed
to
solve.
Otherwise,
you
run
in
on
deficit
of
when
you
prepared
versus
what
the
current
block
gas
cost.
A
Got
its
problem,
the
problem
with
burning
out
gas
is
current
current
contracts
or
not.
A
C
A
Well,
they
they
pass
for
all
promises,
they
start
office.
So
the
the
challenge
comes
in
when
like
because
it's
hard-coded.
A
Well,
new
contracts
will
benefit
from
lower
fees
like
all
this.
This
type
of
legacy
contracts
will
not,
and
if
any
partner
doesn't
have
an
upgradeable
contract,
then
they
will
suffer
from
the
high
cost
gas.
A
A
B
So
how
are
we
going
to
adjust
the
guys
computation
there.
A
B
C
I
see,
can
we
introduce
like
the
limit
on
the
maximum
depth
of
the
or
the
cross
controls,
and
that
will
solve
like
a
lot
of.
A
C
A
A
A
A
C
F
E
I
guess
we
are
ready
to
start.
I
guess
it
will
be
quite
short
today,
but
we
have
a
lot
of
out
to
cover
on
the
indexer
side
anyway.
So
I'd
like
to
give
a
chance
for
them
to
bring
us
an
update
on
this
site.
B
Okay,
let
me
think
so
for
the
indexer
for
explorer
site,
it
looks
like
we
have
almost
finished
all
the
plans
we
had
for
the
current
milestone,
except
this.
The
state
changes,
because
we
have
discussed
it
later
like
today
and
decided
to
move
it
to
the
next
milestone.
B
At
least
explorer
team
was
introduced
with
the
new
database
and
new
backend,
and
it
is
possible
to
start
re-implementing
to
the
new
backend.
B
Unfortunately,
we're
currently
syncing
test
net
from
the
genesis
with
archival
knot
and
it's
way
too
long
38
of
blocks
currently
hope
it
will
finish
after
weekend.
B
E
What
so,
what
is
the
state
for
my
net
index
right
now?
Is
it
fully
synced.
B
E
B
You
you
fully
think
that
our
archival
one
for
magnet,
not
sure
about
archival,
actually.
B
Yeah
we
we
planned
to
to
make
it
archival,
and
we
received
a
back
cop
for
main
net
to
speed
it
up.
Yeah.
E
Yeah,
so
the
test
net
like
turned
out
to
be
not
archival,
and
we
are
currently
re-thinking
it,
but
it's
much
shorter.
I
it
has
much
shorter
history,
so
it
like,
as
boyden
mentioned
it's
halfway,
there
should
be
synced
over
the
weekend.
I
guess.
Overall,
the
stability
as
bogdan
mentioned
is
pretty
good
and
quite
solid,
but
like
it
was
working
for
like
since
sunday,
but
overall,
like
we
anticipated
like
some
issues
in
the
indexer
itself,
and
none
of
them
happened
at
all.
E
So
that's
pretty
nice
and
neat
great
work,
but
then,
by
the
way.
E
So
is,
as
boyden
mentioned,
we
unblocked
explorer
to
do
this.
Migration
and
ifan
is
looking
into
the
data
structure,
since
we
had
to
change
it,
not
quite
a
lot
but
still
significantly,
and
expose
this
the
scope,
introducing
execution
outcomes
and
receipts
and
more
details
like
all
the
way
through
the
data
structures.
E
But
yeah
we
still
have
one
in
can
like
unfortunate
inconsistency
like
we
cannot
provide
the
exact
information
about
the
order
of
execution
outcomes.
Bowen
is
aware:
no,
it's
affects
another.
E
Our
information
may
get
in
a
like
inconsistent
state.
E
The
only
reason
we
like
get
away
with
it
right
now
is
because
every
block
is
usually
containing
like
just
like
a
few
transactions
in
it,
and
this
does
not
and
and
they're,
not
concurrent,
working
on
the
same
account,
and
this
way
we
don't
confuse
anything
in
our
state,
but
overall
it
would
be
great
to
have
a
consistent
order
there,
for
example,
just
like
it
is
all
about
the
coordinate
cases
when
you
add
and
remove
and
add,
and
remove
access
keys,
for
example,
in
a
single
block.
It's
the
same
access
key.
C
E
Know
what
what
the
latest,
without
the
order.
B
E
You
added,
I
don't
think
it's
it's
an
attack,
vector
it's
more
like
consistency
overall
and
right,
not
only
we
as
a
like
explorer
and
bullet.
Oh.
E
E
Yeah
on
the
reset
rpc
side,
there's
a
small
update
there.
We
had
a
meeting
and
we
had
like
some
chat
with
the
team
behind
rosetta
and
they
implemented
essentially
still
ex
extending
the
scope
of
the
desert.
So
they
introduced
the.
E
There,
like
phase
two
implementation,
so
that's
that
was
under
discussion
for
like
previous
week
and
this
week
they
released
a
new
spec
having
this
very
generic
endpoint,
allowing
us
to
be
very
flexible
there,
but
still
we,
the
main
focus,
would
be
to
provide
the
same
flexibility
while
still
allowing
them
to
specifically
view
the
lock-up
balances
and
stuff
like
that.
E
Unless
I
forget
something
cool
by
the
way,
I
wanted
to
clarify
whether
like
where
the
near
sdk
and
all
the
sdk
fall
into,
is
it
not
an
interfaces
or
it's
a
separate
or
group.
C
So
we
are
right
now
making
some
reorganizations
in
terms
of
who
is
working
on.
What
specifically
is
eugene
so
he's
working
on
the
runtime
and
we're
thinking
about
expanding
it
to
runtime
applications.
C
So
nearest
decay
can
be
this
component.
I
think
I
might
have
said
it
even
before,
which
is
like
one
on
his
side,
but
at
the
same
time
it's
called
by
the
node
interfaces
because
at
least
at
least
because
we
need
someone
like
you
would
like
depressed
experience,
so
yeah
consider
nearest
decay
to
be
like
columned
component,
where
leogene
would
come
in
and
say
we
need
this
feature
for
nearest
decay
right,
he
would
say.
Oh,
we
need
like
a
payable
annotation.
C
We
need
this
kind
of
trick
to
add
there,
because
he
would
look
into
what
our
partners
need
and
he
would
you
know,
understand
their
needs
and
say
we
need
this
thing
because
they
expect
this
to
be
in
solidity
and
we
don't
have
an
analog
there,
so
he
would
have
definitely
like
a
product
perspective
there,
but
how
it's
going
to
be
implemented.
C
It
can
be
him,
it
can
be
like
no
interfaces
team,
it
can
be
bounties.
We
have
some
external
contributors
coming
and
we
need.
We
need
at
least
two
people
who
would
be
co-owning
it
anyway,
and
we
probably
should
start
with
someone
who's
experienced
with
the
rest
like
deeply
experienced,
because
it's
all
microsystem
there.
So
it's
probably
gonna,
be
you
and
eugene
calling
in
it.
E
E
E
G
F
Yeah
we
can
do
it.
Do
we
have
everybody
here,
you
jin,
okay,
mike
peter
hi,
peter
thanks
for
joining
us
and
that's
okay.
Let's
do
it.
I
believe
that,
starting
from
today,
it
is
going
to
be
me
who
is
going
to
to
document
everything
that
that
we
are
going
to
discuss
on
this
course
so
yeah,
I'm
going
just
to
start.
The
screen
screen
sharing.
F
Okay,
so
so
I
need
to
sorry
for
this,
but
I
need
to
read
all
the
zoom
because
of
the
the
privacy
security
in
privacy
settings
mark.
Okay,
give
me
just
a
second
okay.
C
I
F
Here
I'm
back
here:
yeah:
let's
do
it?
Okay!
So
can
you
see
it?
Yes?
Okay,
that's
great!
So
let's
do
the
status
update.
So
what
I
think
that
maybe
is
in
in
some
some
few
not
too
distant
future.
F
We
are
going
to
to
change
a
little
bit
the
way
how
how
these
calls
are
done,
and
maybe
a
bit
more
focused
on
the
review
and
the
things
that
were
planned
on
on
the
last
calls
and
then
also
sharing
some
plans
on
the
analyticals,
but
for
now,
let's
just
stick
with
updating
on
on
any
anything
that
you
think
that
is.
That
is
important
for
this
call.
Okay,
so
if
you
don't
mind,
let's
go
let's
go
in
the
same
in
the
same,
like
order
that
was
during
the
last
call.
H
Okay,
sorry,
after
two
failed
implementations
I
found
so
there
were
two
places
I
found
each
vigil
and
the
eip
itself
for
712
had
a
implementation
of
how
to
basically
recover
the
address
from
a
signed
message
and
both
of
those
once
implemented
were
recovering
the
wrong
the
same
wrong
address
once
the
message
was
nested,
which
was
very
strange,
but
I
think
the
implementation
was
kind
of
like
correct
if
you
will,
in
the
sense
that
they
were
returning
the
same
one,
I
discovered
that
there's
a
meta
mask
library
called
eath,
sig
util
that
actually
will
successfully
recover
that
that
address
once
the
message
is
nested,
so
that
that
took
a
little
bit
of
trial
and
error.
H
Third
time
was
the
charm
on
that.
So
I
figured
that
out.
I
have,
I
still
have
the
basic
relayer,
which
is
an
express
js,
app
going.
I
refactored
that
to
use
typescript
this
week.
H
As
a
note,
I
also
saw
a
little
bit
of
issues
with
how
migrations
work
and
I
added
some
troubleshooting
to
the
readme,
which
is
basically
like
when
once
you're
compiling
your
contracts
and
then
you
make
changes
and
try
again,
you
may
need
to
blow
away
the
abi
files,
and
so
this
is
a
possible
hurdle
and
red
flag,
but
I'm
not
sure
yet.
So
I
added
some
troubleshooting
and
discovered
that
the
hard
way
in
terms
of
rlp
encoding
that
we'll
be
needing
to
do.
H
C
Oh
one
small
notes,
so
maybe
we
should
share
this
llp,
our
lp
library
investigation
with
the
bridge
team,
since
we
also
have
somewhere
lp
library
in
there-
and
I
mean
we
went
through
quite
a
bit
of
pain,
trying
to
pick
the
right,
javascript
libraries
to
work
with
ethereum
primitives.
H
Thanks
and
currently
there's
just
kind
of
two
more
items
here.
H
Currently
we've
been
having
recent
discussions
about
the
best
format
for
what
the
user
should
see
in
metamask
when
they
are
going
to
sign
something
I
am
still
focusing
in
on
the
very
basic
pet
shop
example,
and
what
we
don't
want
to
have
happen
is
someone
gets
a
pop-up
on
metamask
and
sees
you
know,
rlp
encoded
items
we
would
like
them
to
see
something
human
readable,
and
I
I
don't
have
a
clear
answer
on
the
best
way
to
do
this
yet,
but
we
yeah
we
want
to
have
the
interface
be
like
as
understandable
as
possible.
F
F
Some
kind
of
some
kind
of
an
erc-
I
don't
know
maybe
712-
that
is
that
is
saying
about
how
how
you
need
to
to
structure
the
the
messages
that
you're
assigning.
So
the
meta
mask
also
supports
this.
This
erc-
and
it
like,
translates
everything
in
there
in
the
in
the
human
readable
format.
H
Yeah
right
correct,
so
that
it
it's
a
little
bit
different
than
that,
because
so
right
now
I
do
have
a
human,
readable,
nested
structure
where
it
says
I
am
trying
to
adopt
pet.
You
know
number
14,
the
id
is
14..
I
will
give
you
10
fungible
tokens
from
this
fungible
token
contract
address.
H
I
am
trying
to
call
the
function
adopt
to
adopt
this
pet,
and
so
so
metamask
is
currently
displaying
all
of
that
in
human
readable
form.
I
think
where
it
gets
tricky
and
I
will
probably
have
a
hard
time
answering
questions
on
on
this
call
exactly
what
we
should
do
next,
but
as
soon
as
that
goes
like,
when
once
you
sign
that
and
then
that
goes
into
like
the
relayer
and
then
we
talk
to
the
near
evm,
the
meta
call
function
inside
of
near
core
is
like
expecting
it
to
be
rlp
encoded
arguments.
D
Sorry,
it's
not
it's
not
about
rlp
encoding,
it
needs
to
be
abi
input,
so
function.
Call
in
helium
is
like
the
function.
Name
is
mapped
into
an
integer
in
ethereum,
because
everything's
integers,
and
so,
if
you
signing
a
string
and
passing
it
into
the
evm
on
the
near
core
side,
there's
no
way
for
near
core
to
figure
out
what
the
string
actually
means.
D
So
so
the
question
is
how
to
show
a
string
to
the
user
and
pass
an
integer
to
the
to
the
signature.
Okay-
and
I
don't
know,
if
I
don't
know
if
this
is
possible-
this
is
just
like,
like
we
may
need
to
kind
of
work
around
this
in
some
other
way.
H
Yeah
I
I
was
I
I
was
kind
of
thinking
when
the
user
is
signing
on
metamask,
we
could
actually
we
could
sort
of
actually
yeah.
I
think
I
think
we'll
take
this
offline.
I
think
I
understand
the
problem
where,
like
so
we're
talking
about
the
function
signature,
I
think
also
called
the
function
selector
and
that
will
look
like
gobbledygook
in
metamask.
H
So
we
need
to
to
figure
that
out
there
isn't
a
library
imported
to
determine
that
function,
signature,
and
I
guess
I
guess
since
we're
on
this
topic
in
ilias
here
I
thought
that
elia
you
mentioned.
We
need
to
send
the
abi,
but
we
only
need
the
abi
for
the
function.
Signature,
correct.
H
I
Okay,
hey,
so
I
have
to
drop
off
in
four
minutes.
I
don't
think
we're
gonna
have
time
to
do
what
I
was
hoping,
but
I
just
want
to
put
it
in
everyone's
mind
of
what
what's
going
on
from
the
community
side
towards
the
evm.
So
I'll
just
be
really
brief.
Number
one
is
there's
a
contributor
who
is
has
previously
worked
on
compilers.
He
pitched,
I
think
max
already
heard
about
this
writing
a
transpiler
from
solidity
to
rust,
and
instead
I'm
I'm
gonna
pass
him
over
here.
I
We
have
two
other
teams
or
sorry
three
total
teams
who
are
super
competent,
who
need
or
not
need,
but
it
would
be
really
helpful
if
we
could
create
bounties
that
are
evm,
specific
or
bridge
specific
that
are
like
really
challenging
deeply
technical,
that
we
either
don't
have
time
to
or
time
to
get
to
with
this
team
or
are
like
just
under
your
guys
priorities,
and
then
three
is.
I
I
will
in
the
next
week
need
someone
to
help.
Austin
griffith
from
the
ethereum
side,
set
up
an
evm
endpoint
to
deploy
some
code.
It
doesn't
have
to
be
in
sync,
it's
it
can
be
async,
so
yeah.
Those
are
my
three
asks
and
and
then
I
guess
I
come
bearing
gifts
with
the
evm
contributor.
I
mean,
assuming
we
can,
you
know,
get
them
up
and
running.
I
Austin
griffith
spelled
a-u-s-t-I-n
g-r-I-f-f-I-t-a-h
he's
an
ethereum
sort
of
like
public
figure.
He
makes
really
cool
tutorials
on
ethereum
and
he's
pumped
on
deploying
one
of
his
tutorials
to
the
evm,
specifically
and
so
yeah.
This
is
like
this
is
fine
going
public.
I
know
we're
on
youtube
and
maybe
he's
watching
hi
austin.
If
he's
watching
so
he's
going
to
work
with
us
on
deploying
a
pretty
cool
project
that
he's
built
with
one
of
our.
I
I
guess
you
could
say
one
of
our
community
engineers
as
well
and
then
publishing
that
somewhere.
So
I
don't
want
to
get
into
too
many
details,
because
I
want
to
spoil
the
plot
of
the
video,
but
that
that's
what's
coming.
C
I
think
we
need
to
think
on
that.
Okay,
great,
like
like
we
discussed
in
slack.
I
Oh
right,
yeah
sure
just
the
main
thing
that
I
want
on
everyone's
mind
is:
if
you
guys
have
ideas
for
the
bounties,
that's
really
valuable.
Oh
someone
is
someone
is
helicopter
mode.
I
F
Okay,
so
as
an
action
point
here
from
max
to
synth
with
with
peter
okay
on
this,
but
we
obviously
probably
well
from
my
point
of
view-
we
we
can't
deploy
this
only
at
the
point
when
we
we
are
having
evm,
isn't
it.
I
Yeah,
sorry,
one
second,
yes,
but
I
think
we
can
set
up
a
private
network
with
the
evm.
It's
a
relatively
simple
contract,
so
imagine
like
if
we
were
to
launch
pet
shop,
you
know
on
its
own
private
network
or
on
the
private
network.
We
could
probably
come
up
with
that
kind
of
solution.
It
can
be
hacky,
it's
just
for
a
demo.
B
I
We
yeah,
we
just
need
to
make
it
it's
not.
Even
it
doesn't
even
need
to
be
broadly
user
friendly.
It
just
needs
to
be
like
specifically
demo-able
and
we
need
to
know
the
steps
that
it
takes
to
deploy
a
new
contract.
Like
I
bet
yeah,
we
can
sync
offline
on
that.
It's
it's
just.
I
I
Yep
sounds
good,
but
yeah.
The
thing
like
I
said
I
will
reiterate.
The
thing
that
is
should
be
really
important
is
identifying
the
work
that
we
can
offload
to
the
community,
so
we'll
have
a
whole
new
resource
and
then
three
other
teams
that
that
can
contribute
to
this.
So
I
know
we're
like
strapped
for
resources,
but
this
is
something
that
can
that
can
amplify
us.
F
Thanks
peter
okay
bo,
can
you
update
us
on
the
on
the
on
the
ethereum
evm
gas
into
the
near
gas.
F
Bull
are
you
here,
maybe
not
can
skip,
maybe
not:
okay,
okay,
eugene
capone.
Can
you
update
us
on
the
on
the
on
the
pr?
Well,
the
majority
of
the
code?
Pr
have
you
have
you
been
able
to
take
a
look
at
it
once
again.
F
G
F
Yeah,
I
just
I
just
I
just
confused
two
agents
that
are
on
the
call,
the
other
one.
A
So
I'm
looking
at
the
edm,
I
think
most
of
the
stuff
here
addressed
by
a
thing
that
should
be
handled
by
gas.
So
overall
it
seems
fine.
We
just
need
to
verify
it
and
I
don't
know
how
it's
actually
done
since
currently,
gas
is
disabled,
well,
not
disabled,
to
just
set
to
higher
limit.
A
So
once
we
have
some
estimates
from
both
and
we
plug
it
into
limits
on
the
vm,
then
we
need
to
make
sure
all
the
findings
that
we
have
are
addressed,
and
I
yeah
I'm
reviewing
the
code
I'll,
just
unresolve
stuff
that
should
be
handled
by
guests
to
make
sure
we
don't
forget
to
to
test
them.
Once
we
have
gas
enabled.
F
F
Do
we
so
that
was
something
that
we
were
we
have
been
discussing
yesterday
is?
Is
there
any
anybody
on
this
call
that
who
knows
whom
we
should
ask
when,
when
it
is
going
to
be
done,.
A
A
I'm
going
to
look
into
this
okay
yeah,
but
I
I
didn't
have
enough
time
to
to
look
because
I
didn't
spend
enough
time
I
would
say
not
to
have,
but
it
seems
like
most
of
this
was
addressed
so
finish.
It
today.
F
Okay-
okay,
that's
great!
I
don't
know
who
who's
also
been
doing
anything
on
the
on
the
evm
side.
G
I
can
give
some
update
so
last
friday
I
was
all
take
a
day
off.
So
from
last
friday,
I
did
figure
out
the
deploy.
Contract
volatility
is
because
we
didn't
account
for
both
evm
gas
and
the
contact
size,
because
the
deploy
step
is
actually
both
a
stage
change
that,
depending
on
the
size
of
the
contracts
and
also
calling
the
constructor
of
the
contract.
G
This
part
is
reflected
by
the
evm
gas
used
by
the
uvm
crate
and
continue
both
to
and
do
a
linear
regression.
G
G
But
if
the
size
of
the
state
is
pinned
the
all
the
cost
is
now
very,
very
stable,
and
the
final
thing
is
oh,
so
I
also
try
to
measure
the
pre-compiled
functions
like
shar
256
and
the
other
tool
like
identity.
The
the
cost
should
show
the
cost,
as
earlier
point
out
should
not
be
fixed
either
instead,
depending
on
the
size
of
the
input.
G
G
Since
our
event,
guest
provider
is
a
integer,
so
it's
often
just
zero.
So
I'm
I
want
to
discuss
here
whether
we
still
want
that,
since
we
we
will.
We
would
expect
like
this
hash
of
an
identity.
Input
of
the
function.
Input
is
usually
less
than
1
kb
and
that
multiplies
the
uvm
guess
per
byte
for
that
precompiled
function
is
still
like
very
trivial.
D
So
the
reason
why
we
needed
to
scale
is
exactly
what
eugene
mentioned
is
because
if
you
pass
a
100
megabyte
long
input,
then
the
data
like
or
like
there's
a
l
there's
a
links
like
you,
don't
even
need
the
input,
long
input.
You
just
say
I
want
you,
know
a
hundred
billion
terabytes
of
of
input
like
there's
like
multiple
multiple
functions
that
take
lengths
as
one
of
the
parameters
specifically
to
calculate
gas,
and
so
the
reason
to
use
it
is
be
is
to
pretty
much
make
sure
you
will
not
be
charging.
D
You
will
not
be
trying
to
allocate
hundred
tera
gas
to
100
byte
of
memory.
Now
I
don't
think
we
should.
We
would
need
like
if
we
already
mapped
our
our
evm
gas
to
our
stuff.
Why
don't
we
just
use
the
gas
calculations
of
the
evm
of
the
ethereum
in
the
first
place
like?
Why?
Don't
we
just
take
the
price?
The
gas
pricing
cost
calculations
from
the
open,
ethereum
parity,
client
and
just
use
that.
G
So
evm
precompile
functions
is
measure.
Its
cost
is
deducted
in
a
different
way.
It
doesn't
have
been
exposed
by
the
grid.
Instead,
it's
run
a
uvm
interpreter,
integrator,
evm
contact,
and
when
it
see
each
of
the
precompiled
functions,
you
will
try
to
deduct
calc
them
as
a
guest.
G
You
give
you
know
config.
So
once
we
once
I
fund
the
near
cost
of
this
precompiled,
I
need
to
convert
it
back
to
evm
gas
and
the
letter
that
the
interpreter
deduct
the
evm
guess
when
it
integrated
this
contract.
G
F
So
sorry,
I
I
probably
missed
this.
What
what
is
the
solution
here
to
this
problem.
D
A
A
Let's
say
we
are
faster
than
ethereum
gas,
then
the
limits
might
be
higher
and
depending
on
how
much
more
gas
we
can
process
within
a
second
and
that
will
affect
our
limit
and
it
might
blow
up
the
memory
just
because
if
it
was
not
possible
in
ethereum
and
it
was
limiting
to
100
max
before-
and
let's
say
we
can
do
10x,
then
it
will
be
one
gig
on
us.
A
So
we
need
just
to
make
sure
that
limits
are
dec
adequate
for
the
max
ram
that
it
can
allocate
and
stuff
like
this.
A
G
Yeah
yeah
the
same
you'll
and
me
doesn't
know
we
are
not
sure
it
will
be
compatible
with
our
near
to
evm
gas
conversion
rate,
because
different
uvm
gas,
like
on
the
different
circumstances,
it
has
different
conversion
rate
for
deploy,
contact
and
co
function.
It
might
be
different
than
this
pre
pre-compiled
function
cost
so
the
the
way
we
did
is
probably
more
accurate,
as
we
calculate
the
number
of
in
cpu
instructions
and
use
that
to
reflect
how
much
uvm
gas
it
should
be.
D
G
Okay,
so
it's
not
a
trivial
like
link
like
cost
per
byte.
Instead.
D
F
Okay,
so
so
the
calculation
so
once
again,
just
just
to
wrap
this
up,
we
can
use
the
existing
implementations
that
are
used
by
parity
in
their
rus
client
that
are
doing
the
calculations
of
the
gas
in
the
air
in
the
evm,
and
we
just
need
to
to
use
the
same
thing
and
then
rescale
it
to
to
the
near
gas.
Is
this
correct.
G
F
Okay:
okay,
great
any
any
additional
updates.
C
F
F
Let
me
share
with
you
the
approach
that
I
I
have
in
my
mind
right
now,
I'm
going
to
to
share
it
in
some
kind
of
I'm
going
to
put
it
in
the
document
and
share
it
with
all
of
you
and
then
you
would
be
able
to
comment
this,
and
it
is
just
like
the
stage
staged
approach
without
without
the
dates.
Let's,
let's
first
agree
on
the
stages
that
we
will
have
there
and
then
we
would
be
able
to
to
discuss
the
dates
and
maybe
set
some
some
milestones.
F
So
we
can.
We
can
do
this,
so
it
is
already
it
is
already
created,
so
I
actually
shared
with
elia
today
earlier
yeah,
so
I'm
going
just
to
send
it
to
everybody
and
let's,
let's
see
whether
it
fits
our
vision,
how
we
need
to
how
we
need
to
perceive
okay.
C
C
Right
seems
that
we
also
need
to
add
an
action
items
to
sync,
with
bowen
on
rolling
out
official
flex
with
evm
yeah
and
like
exactly
what
kind
of
coordination
we're
gonna
have
there.
F
F
Not
for
me,
okay,
great
yeah,
okay,
great
well,
then
thanks
everybody
for
for
joining
this
call
and
yeah
see
you
in
a
week.