►
From YouTube: libp2p weekly sync - December 9, 2019
A
All
right,
hello,
everybody
and
welcome
to
the
Libby
to
be
weekly
sink
II
crib
pad
is
in
the
chat.
Please
add
your
name
to
the
attendees
list.
Add
your
a
sink
updates.
If
you
have
any
agenda
items,
add
those
and
we
will
get
them
at
the
end
of
the
call.
Otherwise
we
will
just
get
things
kicked
off
with
go.
B
That
was
one
of
the
main
issues
we
asked
like
community
and
we
we
had
I
think
when
we
can
offer
dog
someone
who
created
an
issue
about
it,
and
so
I
decided
that
now
that
we
are
in
the
refactor
phase
to
just
create
a
proper
dog
for
exempting
from
the
creation
of
just
a
beta
P,
we
were
Jacob
already
reviewed
it
in
Yusuf
and
I
also
involved
that
person
who
created
issue-
and
we
got
really
good
feedback
in
the
new
dog
for
it.
So
I
think
it
was
a
really
good
movement.
B
Then
I
also
removed
Jacob
in
the
in
the
final
things
for
the
release,
and
we
basically
a
shift
last
week,
the
interoperability
between
edges
and
go.
We
have
still
in
PRS
the
demon,
work,
demon,
client
work
in
the
interrupt,
but
we
will
hopefully
this
week
global
everything
up
and
merge
everything.
Then.
Finally,
I
also
have
been
working
on
a
factoring.
The
our
examples
to
to
the
new
API
with
testing
stuff
I,
already
created
PRS
for
some
of
them,
but
I
mean
like
I,
think
in
the
mural.
A
Nope
all
right
next
up
is
me:
yeah,
so
fix
some
bugs
with
discovery
logic
in
the
refactor,
so
that
should
be
all
good
to
go
on.
That
was
part
of
the
Interop
in
daemon
work,
cleaned
up
a
bunch
of
code
for
that
I
started,
working
on
stats,
refactor
and
then
going
through
that
realize
that
the
metrics
collection
in
je
s,
WP
and
monitoring
was
super
super
non
performant.
A
We
were
like
a
vending
emitting
insane
amount
events,
so
that's
gonna,
get
all
cleaned
up
as
part
of
this,
so
we'll
get
performance
boost
there
and
we
originally
looked
at
making
that
optional,
because
some
people
just
didn't
want
it.
But
right
now
connection
manager
relies
on
it.
So
we
might
see
like
what
we
can
do
in
the
interim
to
get
like
some
basic
connection
management,
but
we'll
likely
just
use
what
we
have
so
far
and
then
worry
about
the
connection
management
stuff.
A
When
we
work
on
a
connection
manager
B
other
than
that
yeah
kind
of
finish,
a
three-factor,
we
are
going
to
do
a
pre-release
this
week,
come
hell
or
high
water,
it's
coming
out
and
then
we'll
start
integrating
that
into
JSI
p
of
s
and
as
part
of
that
create
the
migration
guide.
Any
questions
for
me
all
right.
Next,
up,
roll.
C
C
That
hiring
carry
thesis
been
writing
a
lot
of
post
network
called
I
post
up
my
notes.
There
was
a
bunch
of
thoughts
for
everybody
I'm
talking
about
the
e
theorem,
2.0
Network.
All
this
is
new
series
that
got
started
just
last
week.
It
was
super
productive
and
I'm
very
happy
that
we
have
the
series,
because
it
has
different
cadence
to
the
standard.
C
The
theorem
2.0
calls
which
are
bi-weekly,
and
the
networking
course
is
just
to
deal
with
networking
and
it's
gonna
happen.
Every
three
weeks
are
so
this
is.
This
is
great
for
me.
I
don't
have
to
participate
in
the
wider
courseware
networking.
It's
just
like
you
know
a
10-minute
point
for
that
average
I'd
like
to
always
so
that's
great
I've,
been
writing
a
job
description
for
a
technical
project
manager.
I
think
we
have
a
really
good
sense
now
of
what
we
want
there.
C
Next
I'm
gonna
be
in
Madrid
Evans
Erick
for
the
test
ground
hack
week,
I'm
producing
a
lot
of
design
blueprints
for
test
ground,
I'm
gonna
be
taking
on
the
project
any
duties
while
Mike
is
away
he's
gonna
be
away
for
it.
It's
kind
of
like
an
extended
period
of
six
weeks
where
I
think,
like
after
one
month
he's
gonna
be
working
party
for
bridge
was
just
to
just
to
keep
the
recruitment
going.
I'm
gonna
I
need
to
look
at
deep
metrics
for
it
from
the
b2b
and
deep
tracing.
C
D
But
you
know
this
is
you
could
start
using
it
in
this
that'd
be
like.
Why
are
completely
compliant
with
the
spec
and
a
wire
protocol
level
I
feel
is
like
the
marker
there
and
then
after
that,
we'll
want.
Like
you
know,
this
is
stable
and
we're
probably
not
gonna
change
it
very
much
and
that
ultimate
goal
is
like
well
have
included
in
:
TP
by
default.
So
I
still
haven't
really
like
figured
out.
What's
in
those
two
buckets
yet,
but
it's
it's
been
pretty
fun
like
hacking
on
it
and
I.
Think
it's
coming
along.
D
Okay,
I
mean
the
other
thing
I
did
sort
of
in
the
earlier
part
of
the
week
was
just
like
a
lot
of
head-scratching
about
how
to
go.
Move
forward
with
the
sign
routing
records
and
row
help
me
clear
up,
though
we
don't
really
need
a
bunch
of
big
interface
changes.
We
can
sort
of
hide
the
implementation
details
for
like
whether
you're
going
to
be
using
signed
records
or
not
inside
the
DHT,
and
then
users
won't
really
notice
a
difference.
D
They'll
just
they'll
get
addresses
they
will
come
from
a
signed
source
if
you're,
like
you
know,
if
you
say,
I
only
want
signed
records
and
nobody
like
a
little
transition
period,
where
you're,
probably
getting
both
until
everybody's
upgraded,
so
that
I
haven't
done
much
with
that
today.
I've
mostly
been
thinking
about
noise
things,
but
I'd
like
to
get
the
open
PRS
that
I
have
like
for
the
just
enabling
signed
records
at
all
sort
of
into
in
the
can
or
whatever,
because
that'll
unblock
visa
who
wants
to
use
the
signed
record
exchange
and
gossips.
D
So
I
think
I've
been
looking
over
that
this
morning.
There's
a
couple
things
that
you
know:
I,
don't
well,
there's
things
that
we
could
improve,
but
I'm
not
sure
that
we
can't
do
it
later
like
imma.
Do
we
have
to
do
it
right
now
before
we
merge
the
man,
since
that
the
big
I
mean
the
reason
I
was
hesitating
at
all
is
because,
like
the
if
we're
gonna
make
braking
interface
changes,
we
should
break
them
all
at
once.
Maybe
but
I
think
that
this
might
be
the
only
braking
interface
change.
D
We
need
so
I'm
just
gonna
go
through
and
make
sure
that
everything
is
decently
clean
and
then
I'll,
probably
just
send
up
a
flare
like
hey.
Can
we
please
review
and
merge
this
in
the
next,
like,
maybe
tomorrow,
just
to
kind
of
get
that
stuff
and
there's
a
good
minister
building
on
it
and
then
the
look
at
the
DHT
and
how
we
can
actually
integrate
that
throughout
the
reach?
So.
C
The
one
thing
that
I
was
thinking
about
I,
don't
know
what
your
thoughts
are
is
how
we
going
to
encapsulate
the
API
sand,
the
primitives
to
actually
generate
these
records
so
removing
the
usage
of
these
records
by
the
DHT.
We
know
that
there
are
several
subsystems
that
are
going
to
use
these
records
right,
so
DHT
being
one
of
them,
so
don't
so
I
think.
C
D
So
far,
the
only
creation
method
I
have
it's
just
like
a
free
function
that
takes
a
host
and
will
like
use
the
private
key
to
scientic
and
and
whatever
addresses
the
host
has.
But
that's
kind
of
like
hung
up
on
that
at
the
moment,
because
we
probably
don't
like
we
want
to
filter
those
out,
maybe
like
filter
out
the
local
addresses
and
stuff.
So
that's
what
I'm
trying
to
like
on
the
fence
about
do.
We
need
that
right
now
or
not
yeah.
C
D
D
A
We
have
one
thing
so
note
factory,
you
know
this
this
week
earlier
this
last
week:
it's
not
it's
already
this
week,
so
I
think
they
have
like
an
initial
implementation
of
it.
I'm
not
sure.
What's
testing
this
for
that,
but
in
theory
we
couldn't
try
to
figure
out
what
we
need
to
do
to
get.
Maybe
Interop
testing
hooked
up
to
that
for
the
suite.
So
we
connection.
D
A
Yeah,
they
touch
basically
all
message:
the
repo.
Let
me
see,
if
I
think
from
the
real
quick
but
yeah,
it
would
be
cool
to
see
what
we
can
do
to
get
like,
interrupts
testing.
What
that
looks,
a
message
that
in
chat,
that's
the
Reba
fur
for
jeaious
so
and
that
conforms
to
the
new
async
refactor
API
they've
been
using
that
API.
D
That's
a
good
question:
I
I,
don't
know
I
mean
I
feel
like
for
me.
It's
mostly
because
we
can
get
by
and
I'm,
not
sure
like
that.
I
currently
see
it
like
a
better
enough
situation
like
advocate
for
breaking
on,
but
it
is
possible
that
you
know
the
interfaces
could
use.
The
read
like
a
bigger
like
sort
of
review
and
we
could
figure
out
if
they're
actually
working
like
I
know
that,
like
there
are,
the
peer
store
is
like
a
giant
sort
of
mecha
interface.
C
So
one
reason
why
they
can't
be
exposed
only
in
the
ps4
and
they
need
to
be
exposed.
You
say
event
bus
is
that
they
would
change.
So
as
some
where
address
has
changed
and
we
discover
new
addresses
for
ourselves-
and
you
know
what
that
does
its
thing
as
the
magic
of
hint
and
that
happens,
they
won't
change.
So
we
can't
become
the
Pierce.
Troy
is
not
the
right.
The
piercer
is
established
right,
it
doesn't
have
some
some,
no
it
doesn't,
it
doesn't
have
any
subscription
facility,
so
it
can't
be.
That
can
be
the
place.
C
D
C
D
C
D
C
Let's
do
one
thing:
I'll
put
put
some
sketch
out
some
notes
on
everything
that
I
think
needs
to
be
touched,
just
an
outline
and
then
a
little
bit
just
soon.
So
I
can
unblock
you
for
now
and
I'll
evolve
that
into
a
just
a
fully
fledged
assigned
blueprint
over
the
next
few
days,
but
yeah.
So
you
know
this
changes
in
the
eventbus.
That's
gonna
be
changes
in
the
PS
chart.
Those
appear
so
changes
don't
need
to
be
I,
don't
need
to
be
interference,
interface
breaking
and
we
got
a
like
tested
against
the
DHT.
D
A
sort
of
similar,
my
related
note,
I,
feel
like
my
naming,
could
use
like
a
little
like
consistency
pass
or
whatever,
like
in
some
places,
I'm
calling
them
signed
records
and
other
places
certified
records
and,
like
whatever
I
mean
we
might
as
well
just
figure
out
what
we
want
and
clean
up
before.
We
merge
things.
G
G
F
Yeah,
so
the
baby
pub/sub
router
thing
seems
pretty:
pretty
Merville
should
be
happening
soon.
I'm
would
like
to
do
a
little
more
testing,
but
unfortunately,
I
can't
do
the
testing
without
a
fuss
because
of
some
interface
shenanigans
around
the
asynchronous
data
stores.
All
those
pairs
are
like
merged
and
usable
within
the
pupae.
It's
just
in
order
to
put
them
in
an
application.
You
gotta
get
everybody
doing
the
same
things.
At
the
same
time,.
F
F
Until
we
can
expose
and
pub/sub
events
that
say,
hey
and
we
dropped
a
message,
but
as
soon
as
we
do
that
we
can
remove
the
rebroadcasting.
So
that's
that's
nice
and
then
I'm,
just
gonna
try
and
start
being
useful
with
tht
things.
Roll
asked
me
to
review
a
PR,
so
I'm
gonna
go
I'm
gonna.
Do
that
the
other
ones
I
was
gonna.
Look
at
were
John
and
Stevens.
F
Pr
is
about
querying
multiple
paths
at
the
same
time
and
then
also
allowing
paths
to
terminate,
because
you
sort
of
need
to
do
both
of
those
things,
not
just
one
of
them.
If
they're
gonna
be
useful
at
all
another
one
I
was
just
sort
of
interested
in
and
was
running
by.
Some
of
the
like
JX
like
AFS
folks,
is
whether
it's
feasible
to
tackle
the
content
routing
by
a
multi
hash,
or
at
least
not
make
it
so
that
what
if
we
switch
to
CID
v1
by
default,
we're
not.
F
We
don't
have
to
do
two
DHG
queries
to
look
for
the
same
data
and,
if
it's
possible
to
do
this
without
like
turning
all
the
interfaces
inside
out
by
maybe
just
having
everyone,
rebroadcast
ocid
be
zeros
or
some
preset,
cid
v1
or
something
like
that
and
I.
Don't
know
like
I
guess:
I
wanted
to
gauge
whether
this
sounds
like
really
big
thing:
lots
of
design
decisions
or
whether
it's
yeah.
As
long
as
you
don't
break
the
interfaces,
we
can
push
the
design
decisions
down
to
like
a
month
or
two
from
now.
C
A
Okay,
we
are
done
with
the
list.
We
have
a
couple
of
people
on
the
calls:
let's
go:
alphabetically
Jim,
okay,.
H
I'm
mostly
saved
my
updates
for
the
Tuscan
meeting
tomorrow,
but
I
am
sort
of
embarking
on
a
new
test,
doing
some
conductivity
testing,
which
is
the
obviously
going
to
use
some
whippy-tippy
there
so
absorbing
information.
Yes,
there's
a
whole
process
going
to
be
like
figuring
out
what
I'm
actually
going
to
bake
into
these
tests.
I
think
parts
of
it
will
will
will
develop
some
parts
of
test
ground
which
will
be
reusable
by
anybody
who
wants
to
build
some
custom
for
the
p2p
software,
you're,
writing
and
hopefully
that
will
be
useful
to
people.
I
Alone,
so
yes,
it
have
been
carrying
on
with
some
work,
the
gossips
up
front,
and
mainly
the
writing
off
that
they
can
color
report
that
it's
going
to
be
useful
for
people
in
and
out
of
Pio,
so
I've
been
yeah,
be
so
quite
a
beach
a
little
bit
with
Raul
and
I.
There
is
a
first
draft
of
this
ready
he's
not
exactly
readable
to
share
with
outside
before.
If
anyone
wants
to
have
a
look,
just
Bing
me,
I
can
give
you
to
another
leaf
la
tech
document.
I
So
if
anyone
wants
to
have
a
look
already
just
give
me
a
shout
out,
they
can
give
the
link
I
hope,
that's
I'm,
going
to
have
a
version
of
that
ready
before
Christmas
might
be
a
little
bit
up
to
me,
sick,
but
I'll
do
my
best
to
have
it
yeah.
The
least.
This
will
not
include
results,
but
Raul
said
that
VC's
fine,
they
are
not
looking
for
results.
I
Of
these
poems
results
is
going
to
be
good
for
the
more
general
like
community,
but
we
can
live
without
it
for
now
so
I'm
going
to
try
and
focus
on
what
are
the
main
features.
The
other
value
comparison
with
related
work
and
what
is
bad,
what
is
also
likely
I'm
going
to
have
some
more
one-to-one
catch-up
with
visa
to
get
an
understanding
of
you
know
what
had
in
his
mind
when
designing
that
role.
Is
that
enough
to
have
in
mind
when
producer
yeah.
I
C
That's
fantastic,
I
think
the
technical
description
and
the
comparison
with
the
state
of
the
art
is
honestly
like
would
go
agree
to
great
extent
right
now
to
do
to
what
we
need.
The
these
things
keep
coming
up
right,
especially
Leahey.
Is
there
like
what
are
the
other
options?
Besides
gossip
serve
right,
this
keeps
coming
up.
C
No,
not
because
people
are
do
not
trust
call
cell,
but
because
they
want
to
know
what's
what
else
is
out
there
and
what's
gay,
let's
hide
cheap
right,
so
it's
completely
understandable
so
yeah
that
would
that
would
help
it's
even
for
me
right
like
even
even
for
us,
for
this
team.
We
are
all
behind
gossips
now
than
we
are.
I
I
treated
my
husband
so
yeah
that
that
would
be
really
useful.
Okay,
yeah
I'll
try
to
focus
on
that.
I've
done
some
quite
a
bit
of
related
work.
Reading
a
couple
of
months
ago,
when
I
was
preparing
the
first
one
I
need
to
brush
up
a
little
bit
and
talk
to
visa
in
in
greater
detail,
but
I
hope
I'm
going
to
put
most
of
next
week
on
that
and
yes.
C
I
would
say
that
we
can
also
release
this
in
circles
right.
So
if
you
don't
feel
like
super,
you
feel
like
there's
still
some
work
to
be
done
to
together
to
a
stage
where
you
feel
it's
complete
to
like
be
released
to
to
a
wider
audience.
Then
we
can
release
her
to
like
select
people
as
well
right
to
to
to
gather
feedback.
Yeah.
I
So
so
my
plan
with
this
is
that
eventually
we
discussed
this
before
you
know
it
would
be
good
to
have
a
kind
of
like
published
paper
out
of
that.
So
in
order
to
get
there
we'll
need
results,
first
of
all
which,
with
death
ground
and
five
going
to
implementation,
it's
going
to
be
there
it's
going
to
take
a
few
months.
That's
fine!
So
the
first
version
should
be
yeah.
It
could
be
a
living
document.
That's
why
I'm
saying
so.
I
C
Thing
it
do
this,
this
print
see
where
you
get
to
before
Christmas
and
then
we
get
together.
We
assess
you
know
what
what
we've
achieved,
what
you've
achieved
and
what
else
like.
What
other
points
where
you
feel
less
confident
or
whatever
we
decide
do
we
release
it
to
an
audience?
Do
we
like
send
it
to
a
few
people
only
and
we
can
we
can
have
a
tech
yeah.