►
From YouTube: Filecoin Core Devs Biweekly #8
Description
Recording for: https://github.com/filecoin-project/tpm/issues/20
For more information on Filecoin
- visit the project website: https://filecoin.io/
- or follow Filecoin on Twitter: https://twitter.com/Filecoin
Get Filecoin community news and announcements in your inbox, monthly: http://eepurl.com/gbfn1n
A
All
right
good
morning,
good
evening,
everyone
today
is
thursday
december
17th,
and
this
is
our
eighth
core
devs
bi-weekly
meeting.
Our
agenda
is
somewhat
compact
today,
we'll
start
by
talking
about
flip
nine,
which
we
kind
of
expedited,
along
with
the
process,
and
then
we'll
hear
updates
from
the
various
teams
and
discuss
anything
else
in
whether
that's
fips
or
other
issues
that
people
want
to
raise
so
jumping
into
fit
nine.
A
So
this
came
about
because
there
were
concerns
about
the
sustainability
of
submitting
window
posts
during
periods
of
high
base
fee.
It
was
particularly
affecting
smaller
miners
miners,
with
somewhat
limited
financial
resources
and
just
minors
who
didn't
want
to
be.
You
know
paying
that
much
to
prove
storage,
that
they'd
already
committed
to
the
network.
It's
easy
to
say
that
if
the
base
fee
is
spiking,
now
is
not
a
good
time
to
onboard
new
storage,
but
you
do
still
have
to
submit
your
window
first
messages.
A
So
the
proposal
which
is
very
simple
and
was
discussed
in
the
film
implementers
channel-
and
we
kind
of
already
had
this
conversation
got
buy-in
but
put
it
on
the
table.
The
proposal
is
simply
to
exempt
window
post
messages,
so
the
submit
window
post
message,
in
particular
from
the
network
base
fee,
burn
sorry.
So
this
way,
this
way
the
cost
of
submitting
your
window
post
message
is
largely
divorced
from
what
the
base
fee.
A
What
the
current
base
b
is
you
still,
it
still
affects
a
little
bit,
because
any
overestimation
burn
that
you're
charged
will
be
will
be
tied
to
the
base
fee,
but
it
should
be
orders
of
magnitude
less
it
just
does
not
change
chain
throughput
in
any
way
and
is
very
much
a
a
stop
cap
solution
that
that
doesn't
necessarily
address
the
problem
itself,
but
but
you
know,
helps
helps
network,
especially
as
we
were
going
into
the
holidays
when
there
might
be
a
lower
or
a
longer
turnaround
time.
A
So
I
was
wondering
if
people
had
thoughts
about
this.
I
know
some
people
here
had
basically
proposed
this
already
steven
you
in
particular
had
essentially
proposed
a
while
ago.
So
how
do
people
feel
about
this.
B
Okay,
I
would
think
that
yeah,
all
the
matters
work
and
you
know-
will
welcome
this
kind
of
change
and
and
banners
and
want
to
have
more
to
do
after
the
whole
day,
perhaps
to
actually
improve
the
tps.
But
this
is
the
first
step
yeah.
It
may
not
be
able
to
improve
the
tps,
but
it
will
protect
the
small
miners
and
other
manners
actually
in
this
holiday
yeah.
I
think
it's
yeah.
It's
really
working
currently.
A
Are
there
specific
concerns
that
anyone
had
with
the
idea
I
had.
C
Oh
all
right
yeah,
so
I
mean
I
just
had
like
in
terms
of
like
you
know
all
the
technical
stuff
about
this
fip.
I
I
think
it's
pretty
it's
a
pretty
good
idea,
but
I
think
it's
important
to.
C
I
guess,
make
it
clear
whether
this
is
actually
a
stopgap
solution
and
if
it
is
then
give
like
a
better
timeline.
So
people
know
when
to
expect
and
adjust
of
things
or
if
this
is
a
solution,
that's
going
to
be
built
upon
and
say
so
because,
like
you
know,
I've
heard
that
this
is
a
stopgap
solution,
but
I've
also
heard
that
this
is
a
first
step
to
a
bigger
solution
and
it's
like
kind
of
confusing.
C
Confusing
messages
yeah
so
just
wanted
to
put
that
up.
There.
A
Yeah,
thank
you.
Thank
you
very
much
for
flagging
that.
I
think
I
think
that
you're
right,
so
our
hope
right
now
is
that
this
essentially
becomes
the
first
step
in
what'll
and
the
the
second
step.
The
bigger
piece
of
it
will
be
what
we've
been
calling
optimistic
window
posts.
This
is
issue
number
42
in
the
fips
repo.
The
fib
itself
hasn't
been
opened.
A
Yet
I
think
so,
but
once
once
optimistic
window,
post
basically
means
we
will
be
taking
the
verification
of
window
posts
off
chain,
at
which
point
the
the
actual
gas
cost
of
submit
window
post
itself
will
be
significantly
lowered,
and
so
that's
kind
of
the.
What
we're
hoping
will
be
the
the
longer
term
and
that's
the
actual
solution
that
actually
allows
you
to
increase
gps.
So
that's
that's.
What
we're
that's
what
we're
thinking
right
now.
A
We
feel
we're
feeling
pretty
confident
about
optimistic
when
at
least
internally,
it's
it's
looking
good.
We
haven't.
No
major
rate
red
flags
have
been
raised
yet
so
that
that's
that's
the
timeline
word,
that's
what
we're
hoping
will
happen
and
we'll
talk
about
optimistic
window
post,
maybe
a
bit
more
later
on.
D
But
I
also
want
to
say
one
thing
eric,
which
is:
maybe,
if
you
feel
like
this,
we
should
make
it
more
explicit
in
our
communication,
because
this
is
a
stop
gap,
meaning
it's
unclear
if
this
is
going
to
stay
as
a
final
solution
for
long,
because
new
solutions
may
supersede
this,
but
there
is
an
unlikely
but
possible
scenario
where
fees
are
going
to
be
higher
than
right
now,
but
everyone
is
working
no
higher
than
with
this
change,
and
so
so
what
I'm
saying
is
that,
but
but
the
goal
here
is
to
make
them
lower
than
what
they
are
today.
D
So
so
what
I'm
trying
to
tell
you
is
that
there
may
be
in
the
future
a
solution
where,
instead
of
costing
one
thousandth
of
thousand
cents,
it
will
cost
one.
I
don't
know
one
one
hundredth
or
one
150th
or
something
like
this.
I
don't
want
to
give
like
exact
numbers.
What
I'm
trying
to
tell
you
is
that
this
one
may
make
it
exceptionally
cheap,
and
maybe
it
will
not
be
exceptionally
cheap
for
long,
but
the
goal
is
to
is
whatever
is
going
to
come
after
this
is
going
to.
D
It
could
be
substituting
this,
but
without
making
things
drastically
worse,
that's
so
we
don't
want
an
expensive
window
post
in
the
next
upgrade,
no
matter
what
we
do.
A
Yeah
nicola,
I
think
I
think
the
point,
but
we
can
make
it
more
explicit
in
our
communications
that
this
is
in
at
least
partially.
A
stop
cap
would
be
good
other
thoughts
that
people
have
about
fib,
nine.
A
Okay
sounds
good
yeah,
so
also
I
wanted
to
quickly
thank
everyone
here
for
kind
of
moving
on
fifth,
nine,
fairly
quickly
kind
of
vetting
and
thinking
about
it
and
giving
us
giving
us
a
thumbs
up
on
the
idea.
It
allowed
us
to
move
at
the
pace
that
we
needed.
So
with
that
maybe
we'll
jump
into
updates
from
the
teams.
I
can.
A
I
can
go
first,
given
that
our
large
update
is
around
fifth
nine
as
well,
so
we've
prepared
lotus
one
three
zero,
which
is
the
third
post
liftoff
network
upgrade
all
it
does
is
introduce,
is
implement
flip
mine.
So
it's
a
really
really
small
change
code
wise
and
we
feel
like.
We
feel
confident
that
we've
that
we've
tested
it
and
we're
happy
with
it.
The
network
is
scheduled
to
upgrade
on
monday
december
21st
or
tuesday
december
22nd,
depending
on
where
you
are
in
the
world.
A
So
yeah,
that's
kind
of
the
the
big
update
from
lotus
is
relevant
to
this
group.
Apart
from
that
we're
continuing
to
as
we
as
we
discussed
in
the
past
couple
meetings,
our
focus
is
just
kind
of
on
making
lotus
better.
So
we're
really
just
improving
improving
the
performance,
improving
the
ux
of
various
aspects
of
lotus
from
the
from
the
actors
side,
I'll
quickly,
speak
to
that
as
well.
A
lot
of
the
majority
of
their
work
has
been
on
the
optimistic
window.
A
Post
idea,
we've
sketched
out
what
an
implementation
will
look
like
and
we've
been
continuing
to
kind
of
vet
that
internally,
so
we're
hoping
to
introduce
that
in
the
first
network
upgrade
in
2021,
which
will
probably
be
late
january
early
february
and
things
are,
things
are
generally
going
well
there.
So,
let's
hear
from
stephen,
do
you
want
to
take
us
through
network.
B
Okay,
yeah,
in
the
last
two
weeks,
we
have
done
yeah
some
sense
implant.
The
first
thing
is
that
we
had
wrong
venus
in
calibration
network
and
we
yeah
we're
running
one
node
in
the
calibrating
network
and
with
the
sync
cap
and
all
the
and
all
along
the
chain
and
yeah
well,
and
keep
moving
so
and
apparently
yeah
the
synchronization
and
yeah
is
very
good
and
yeah.
But
what
not
to
start
mining
right
now
is
because
we're
still
working
on
the
mining,
interface,
development
and
integration.
B
All
the
interfaces
are
almost
done,
and
there
yeah,
and
only
a
few
of
them
still
an
in
progress.
We
yeah
we
we
expect
that,
and
that
could
be
done
in
a
few
days.
And
another
thing
is
that
what
working
on
the
message,
poor
implementation,
including
the
invasive
distinction
and
also
the
proverbial
addresses
and
also
the
coding,
is
done.
But
some
testing
is
ongoing
and
we
still
expect
that
will
be
down
here.
A
few
days.
B
And
and
still
we
we
are
facing
some
challenges.
The
first
thing
is
that
the
chance
organization
is,
the
performance
is
not
so
good
and
currently
the
venus
yeah.
Currently,
the
block
verification
is
yeah.
Kind
of
slow
is
much
store
than
lotus.
Sometimes
it
will
take
about
10
seconds
or
an
8
second
to
prevent
and
block.
So
it's
very
slow
and
we
think
it
is
due
to
a
few
things.
B
One
thing
is
that
actually
venus
actually
take
about
a
double
of
the
yeah,
the
space
utilization
and
then
the
notice
and
also
the
memory
usage
is
not
so
good.
So
we
need
to
optimize
those
and
we
need
to
use
some
catching
and
yeah
to
do
you
have
to
make
it
better.
That
is
one
thing
we
need
to
do
in
next
week
because
we
think
yeah
it's
kind
of
hard,
but
we
have
to
do
that
and
based
on
with
this
and
we
we,
we
also
plan
to
do
some
other
things.
B
The
one
thing
is
that
we
we
need
to
complete
the
main
interfaces
and
also
to
integrate
and
the
binding,
yeah
and
positive
and
post
with
with
the
chance
of
relation
together
and
so
that
we
could
have
a
minor
and
actually
in
the
network.
So
the
plan
is
that
we
would
do
it
in
calibration
network
first.
The
plan
is
to
have
that
in
two
weeks,
so
yeah.
B
Let's
do
that
and
and
some
smaller
things
that
we
we
will
also
have
also
to
add
some
user-friendly
serial
comments
and
to
make
it
better
to
to
to
use
okay,
yeah
and
that's
all.
A
Yeah
one
decision
yeah
how
how
far
along
calibration
net
are
you
currently
able
to
sync
to
you
getting
to
the
top
or.
B
Yeah
as
a
collaboration
they
and
essentially
the
chance
of
relation,
is
very
good
and
we
are
yeah
we
and
already
synchronized
and
we
are
in
keep
moving
with
it.
Yeah
we
talk
right,
yeah,
that's.
A
Fantastic
cool-
maybe
let's
hear
from
someone
from
sorry,
meets.
E
E
I'm
not
sure
I
am
into
all
the
details,
so
if
everyone,
if
anyone
from
the
team
wants
to
celebrate,
feel
free,
but
a
little
bit
later
so
currently
we
are
working
on
resolving
this
issue.
Basically,
it
blocks
us
from
syncing
with
the
lotus
main
net
in
somewhat
fast
way.
E
And
I'm
not
sure
we
have
a
solution
for
it
right
now,
but
we
are
looking
for
it.
First
of
all,
we
need
to
identify
what
is
the
problem
exactly
so.
As
for
the
minor
work,
we
test
against
the
local
test
net
with
lotus
with
the
big
sectors,
our
miner,
it
seems
to
be
working
just
fine,
so
the
next
step
will
be
to
test
again
the
maintenance
yep
and,
as
for
our
worked
on
our
work
on
implementation
of
c
plus
plus
actors,
we
have
recently
implemented.
E
Reward
actor
and
this
fight-
remember:
storage,
my
interactive
v2,
so
we
are
continuing
working
on
that.
Not
a
lot
of
hurry
for
this
one
as
we
are
trying
to
focus
on
the
minor
plus
block,
both
changing
parts
of
the
protocol.
A
Sounds
good!
That's
like
some
great
progress,
great
yeah,
let's
hear
from
someone
from
train
safe.
F
Hey
so
yeah,
our
main
goal
has
been
to
work
towards
being
like
audit
ready,
that's
sort
of
what
we've
been
focusing
on
and
the
I
mean
like
there's
no
point
where
we're
really
finished.
So
the
real
question
is:
where
do
you
actually
start
this
audit,
and
so
our
our
goal
is
to
like
basically
have
like
three
things
working
before
we
start
auditing
and
one
is
like
interop
with
the
original
mainnet
network
version,
and
so
there's
like
two
parts
to
this.
F
One
is
like
being
able
to
have
upgradeability
for
actors
so
being
able
to
call
a
different
version
of
actors
depending
on
the
the
height
and
that
that
part
is
done
now
and
matching
actors
v2,
which
is
in
progress.
We
are
almost
done.
Only
the
minor
actor
is
left
so
making
great
progress
on
on
that
goal.
F
The
second
part
to
getting
audit
ready
is
the
storage
miner
integration,
so
we
can
now
initialize
the
storage
miner
with
a
forest
daymond
and
start
creating
tip
sets,
which
is
awesome
and
we're
working
on
getting
the
full
minor
life
cycle
to
work.
I
think
just
before
this
call
eric
message
saying
that
we
were
able
to
get
a
sector
committed,
which
is
awesome
so
making
great
progress
there
and
then.
F
Lastly,
we
want
to
be
able
to
stay
in
sync
with
the
local
devnet
of
forest
and
and
lotus
nodes,
and
so
the
last
step
here
before
testing
is
basically
getting
the
logic
for
switching
the
syncing
state
from
chain
sync
to
chain
follow
and
once
that's
done,
then,
basically
we'll
test
and
see
how
it
goes.
So.
The
goal
is
to
wrap
all
of
these
things
up
by
the
early
new
year,
hopefully,
and
then
start
thinking
about
an
audit
and
then
all
the
other
things
that
need
to
happen
after
that.
A
Nice,
fantastic
yeah,
it
sounds
like
everything
you
listed
towards
getting
as
goals
where
it's
getting
audit
ready,
you're
pretty
close
to
to
have
it
done,
which
is
great,
yeah
cool,
so
raul
had
to
dip,
but
he
left
a
little
update
on
on
his
work.
So
I'm
just
gonna
read
that
out
for
the
benefit
of
the
recording.
A
Regarding
regarding
test
factors,
he
says
you
now
have
the
ability
to
extract
full
tip,
sets
from
live
chains
or
to
extract
a
range
of
tip
sets
such
as
one
two,
three,
four,
five.
Six,
all
the
way
to
one
two,
three,
four,
nine
nine.
The
result
can
be
saved
as
one
tip
set
per
vector
or
squashed
into
a
multi-tip
set
vector,
which
is
great.
A
A
The
differences
in
state
he's
also
adding
a
feature
to
execute
a
callback
after
applying
every
tip
segment
vector
which
allows
you
to
capture
things
from
the
state
that
you
might
want
to
chart
and
analyze
further
as
an
example
he's
using
this
feature
to
save
the
balances
of
all
actors
at
every
chipset.
This
allows
you
to
compare
how
the
current
protocol
version
applies.
A
Gas
pricing
as
compared
with
a
new
version
like
class,
so
this
is
a
way
to
back
test
a
candidate
fork
against
chain
history,
which
is
something
we've
wanted
for
a
really
long
while,
and
he
says
it
needs
to
be
cleaned
up
a
lot,
but
it's
a
good
starting
point
and
a
good
building
block.
A
So
that's
exciting
for
all
of
us.
I
think
cool
anyone
else
in
the
call
would
like
to
provide
an
update,
whether
from
the
other
teams
or
from
pl
folks.
A
A
So
fit
eight
was
opened
somewhat
recently.
I
don't
think
we've
discussed
this
yet
nicola.
Would
you
like
to
talk
about
tip
eight?
A
A
Okay,
assume
he's
left
us.
It
does
give
me
a
second.
A
A
Go
there,
we
are
okay,
yeah,
so
the
other
thing,
the
other
thing
that's
kind
of
moving
more
quickly
that
we
definitely
want
to
get
ready
for,
for
our
first
network,
upgrade
in
january,
is
the
optimistic
window
post
idea.
This
is
something
we've
discussed
somewhat
extensively
already
in
in
the
implementers
channel.
I
was
wondering
if
anyone
had
any
new
thoughts
or
insight
about
the
idea,
because,
if
not,
I
think
we're
pretty
close
to
saying
we.
A
We
want
to
do
this
and
opening
the
fib
for
it
any
new
thoughts
about
it.
A
Cool
yeah,
we
we
discussed
it
two
weeks
ago
as
well.
It's
I
think
I
think
yeah,
it's
it's
moving
in
the
right
direction.
We've
already
gotten
implementation
sketched
out.
So
that's
something
we
can
look
forward
to
in
the
new
year,
which
I
think
yeah
will
stack
well
with
with
what
was
implemented.
What
will
have
been
implemented
in
fifth,
nine
by
then
so
so
that'll
be
good
to
have
cool.
Is
there
anything
else
that
people
wanted
to
talk
about
any
issues?
Any
questions.
B
Yeah
just
one
thing:
yeah:
do
we
have
a
detailed
plan
for
our
next
units,
for
example
the
your
favorite
age
or
any
others?
People
could
refer
to
or
an
attachment
or
a
project
plan,
or
something
like
that.
A
Yeah,
that's
a
good
question
short
answer
is
no,
unfortunately,
we
haven't
we
because
we're
kind
of
still
scoping
that
out
ourselves
and
but
we
haven't
put
that
we
haven't
displayed
that
publicly
anywhere.
I
don't
think
we
should
right
now,
yeah,
so
maybe
I'll
make
a
note
too
to
create
that
and
get
get
input
on.
B
A
Yeah,
I
think
we're
kind
of
missing
that
right
now
yeah.
I
can
quick
for
now
stuff
that
we
intend
to
get
into
the
next
release
is
well
actors,
v3,
which
includes
fip7,
which
is
the
amp
enhanced
optimizations,
and
the
optimistic
window
first
changes.
So
those
are
kind
of
the
the
two
big
pieces-
okay
good,
but
no
thanks
for
that.
We
should
definitely
have
a
place
for
the
community
to
see
that
information.
C
Yeah,
I
just
have
some
questions
actually
now
that
I
think
about
it
with
some
of
just
like
these
fips
in
general
and
like
these
network
upgrades
and
in
particular
like
the
optimistic
window
post
stuff
is,
are
there
any
efforts
going
on
right
now
to
do
like
simulations
on
different
types
of
attacks
because,
like
for
example,
there
were
some
for
for
the
optimistic
window
post?
There
were
some
potential
concerns.
C
C
Heads
I
was
just
wondering
if
there's
any
plans
to
do
some
sort
of
simulations
and
testing
that's.
A
A
good
question,
nicole,
are
you
here,
he's
he
kind
of
leads
all
of
our
internal
security
checkings,
but
he's
not
yeah.
So
for
the
most
part,
I
think
we've
been
we've
kind
of
been
thinking
about
these
things
and
kind
of
theoretically
sketching
them
out
and
seeing
what
what
attacks
are
feasible.
There
are
a
couple
that
we've
that
we've
agreed
is
maybe
possible,
but
we're
not
too
concerned
about
the
impact
of
it.
G
C
A
Cool
sounds
good
yeah.
I
think
I
think
we
can
wrap
up
here.
Thank
you,
everyone
for
joining,
I
think
yeah.
We
won't
be
meeting
two
weeks
from
now,
because
it'll
be
in
the
holidays
or
right
after
so
talk
to
you
at
some
point
in
2021
it's
been
a
great
year.
Thank
you
for
all
your
work.
A
lot
has
happened
so
look
forward
to
what
q1
2021
will
bring
us
so
bye.
Everyone.