►
From YouTube: libp2p weekly sync - February 10, 2020
Description
A
Alright,
hello,
everybody
welcome
to
the
February
10th
flip
p2p
sink
go
ahead
and
get
kicked
off.
Crib
pad
is
in
the
chat,
feel
free
to
add
up
your
a
sink
notes
and
any
agenda
items
that
you
want
to
discuss.
I,
don't
see
any
agenda
items
so
we'll
go
ahead
and
kick
off
the
weekly
updates.
It
looks
like
I
am
at
the
top
of
the
list,
so
I'll
get
started.
A
Yeah
this
week
in
last
week
are
primarily
focused
on
the
getting
noise
working,
so
I've
been
working
with
note
factory
and
then
coordinating
with
youssef
so
that
we
can
start
getting
the
PDP
Interop
tests
between
go
and
j/s
versions
running
so
I've
got
some
open,
daemon,
pr's,
they're,
starting
and
then
I'm,
coordinating
with
node
factory.
For
that,
aside
from
noise
on
primarily
just
working
on
content,
routing
coordination
for
the
jailer
PTP
release,
we
did
some
metrics
collection
spent
some
time
profiling
last
week.
A
The
nice
thing
about
the
giant
refactor
is
that
like
performance
is
pretty
much
stayed
the
same,
so
we
didn't
break
anything.
But
memory
improvement
actually
saw
like
a
30
to
40
percent
improvement
on
memory
usage,
both
heap
and
RSS,
so
that
was
that
was
good
to
see
so
but
yeah
other
than
that.
Let
me
continue
on
noise
and
content.
Writing
any
questions
for
me.
B
B
Oh
and
I
noticed
that
in
the
docs
website
we
currently
have
the
getting
started
tutorial
for
Jessup
HP,
which
is
updated
with
new
release,
so
I
just
created
a
PR
to
make
a
notice
that
that
tutorial
is
from
my
older,
older
release
and
I
will
also
work
on
updating
the
release
content
after
the
Stardust
effect
are
being
finished,
and
that's
it
for
me.
Any
questions.
C
Yeah,
let's
see
so
I
did
the
last.
You
know,
hopefully
last
API
changes
for
a
P
records
and
fix
the
overly-complicated
pierced
or
implementation.
So
I
think
at
this
point
most
of
it's
done.
I'll,
look
at
some
2d
comments.
I
left
for
myself
didn't
address
in
the
as
he
was
going
through
this
morning,
but
I
think
we'll
be
ready
to
merge
today.
C
The
so
that's
the
plan
merge
that
I
did
a
big
noise
refactor
thing
just
to
clean
up
a
lot
of
the
code
and
make
it
easier
to
read
and
that
flushed
out
a
bug
in
go
noise
implementation
that
might
also
be
present
in
the
jeaious
implementation.
I
haven't
looked
closely
enough
into
their
code
yet,
but
if
they
were
basing
it
off
of
go,
it's
likely,
they
get
the
same
error
case
so
I'm
going
to
check
their
if
they
fix
errors,
I
can
still
there
to
fix
either
way.
C
I'll
coordinate
with
them
and
tag
raise
an
issue
on
that
repo.
If
it
looks,
there's
also
house
the
same
bug
so
I'm
gonna
try
to
fix
that
I'm.
Also
writing
today.
Writing
a
benchmark
to
make
sure
that
my
refactor
didn't
make
things
worse
and
just
generally
nice
to
have
going
forward
to
make
sure
that
we're
getting
reasonable
performance
and
we
change
things
yeah
I
can
watch
good.
Let
me
know
if
you
want
to
help
with
the
getting
started
guide.
C
I
can
take
on
as
well
like
you
know,
or
if
you
want
me
to
help
review
it
or
any
input
at
all.
Just
let
me
know
otherwise.
I'm
probably
gonna
focus
on
getting
noise
as
good
as
I
can
get
it,
and
that's
probably
most
my
week
is
just
benchmarking
and
making
sure
that
knowing
is
actually
works.
The
way
we
hope
it
does.
D
One
thing
that
I
wanted
to
add
it
to
that.
To
add
to
that
update
is
that
peer-to-peer
records?
We
are
aiming
to
merge
it
today
step,
but
I
don't
want
to
make
a
release
until
somebody
from
the
controlling
afraid.
Possibly
a
Dean
has
had
the
chance
to
go
through
the
interface
to
go
through
the
implementation
and
see
and
verify
that
it
actually
integrate
able
in
the
in
the
state
than
it
is
we've
used,
Eva's
or
has
already
integrated
at
upstream
and
gossip
sub
in
the
peer
exchange
PR.
D
A
C
Cool
I
feel
like
the
problem,
this
specific
problem
I'm
having
with
the
Bob
I,
think
the
spec
might
be
kind
of
ambiguous
there
as
to
what
you're
actually
supposed
to
do
to
reinitialize
the
state.
So
when
I
fix
this
I'm
gonna
try
and
push
this
back
up
there
as
well,
but
if
you
find
it
confusing
it
all
in
yeah,
okay,
it's.
D
E
D
D
Okay,
yeah,
so
I
think
the
highlight
from
last
week
is
that
we
should
test
on
to
your
point.
One
mm-hmm
and
besides
that
I've
been
doing
intersection
so
been
reviewing
our
shows
PRS
on
all
across
the
board
for
to
incorporate
the
the
intersection,
the
intersection
words.
So
there's
a
PR
and
go
to
be
to
peak
or
to
be
the
obstructions
on
the
intersection
types
and
and
the
schema.
Then
there
is
a
PR
in
go
alleppey
to
be
so
there's
a
new
record
that
we
created
go
the
p2p
interest.
Factor.
D
You've
probably
received
a
notification,
it's
right
now,
there's
that's
not
a
good
read,
be
and
so
on,
but
the
code
is
there.
We
paid
a
few
releases
already
and
basically
this
is
this:
has
the
initial
default
implementation
of
the
interest
factor,
which
is
the
the
the
component
that's
responsible
for
aggregating
the
data
in
the
system
in
serving
and
serving
the
introspection
data?
D
And
then
there
is
the
introspection
endpoint,
which
can
be
anything
particularly
the
the
the
implementation
that
we
have
now
is
a
WebSockets
endpoint
that
serializes
to
protobufs
so,
but
in
the
future
it
can
be
anything
else.
It
can
be
eg,
RPC
can
be
lipids,
it
can
be
in
Libby
to
the
endpoint
as
well.
So
that's
why
we
wanted
to
decoupled
these
two,
these
two
components
and
that
work
is
there,
the
repo
exists
and
it's
very
things,
I'm
urge
to
master
and
with
a
few
releases
then
we
made
as
well
PR.
D
There
was
a
PR
to
go
to
be
TP
core
s,
re
going
to
be
to
be
sworn
to
source,
introspection,
data,
about
connections
and
streams
from
the
swarm
and
then
there's
another
PR
in
go
Libby
to
be
to
add
constructor
constructor
options
to
inject
an
introspection
and
enable
the
intra
Spector,
and
also
the
changes
in
basic
hosts
to
actually
support
and
instantiate
an
introspection
and
expose
the
intro
Spector.
So
this
is
all
the
work
of
art.
So
let's
stop
for
him.
It's
like
his
big.
D
F
D
And
it's
gonna
be
like
that.
All
of
this
is
it's
gonna
be
living
for
now
in
a
feature
branch
we
haven't
merged
to
master
yet
because
yes
yet
has
to
stabilize-
and
that's
probably
a
bunch
of
changes
that
are
gonna,
come
down
the
pike
so
yeah
besides
that,
when
working
on
security
fixes
and
what's
up
for
me
this
week
is
more
reviews
and
unblocking
stuff
I'm
gonna
do
a
sprint
and
gossip
table.
There
is
a
bunch
of
things
that
be
so
breast
that
lab
I
and
a
few
others
from
the
popcorn
team.
D
E
Finally
back
to
being
able
to
devote
some
time
to
that,
and
also
we
have
this
advisory
board,
the
PDP
project,
Advisory
Council,
it's
like
some
people
from
protocol
labs
and
some
of
the
advisors
who
was
just
are
gonna
meet
with
me,
enrolled
in
actually
not
for
two
weeks,
but
I'll
be
going
next
week,
so
I'm
gonna
be
preparing
for
that
and
I'll
sure.
What
comes
out
of
that
yeah.
F
So
my
side
number
one
I
reviewed
now-
will
give
me
a
master
class
on
colon.
By
reviewing
his
introspection
code.
I
took
the
don't,
join
the
DHT
when,
behind
a
network
of
completion,
there
wasn't
much
to
do
a
lot
of
the
code.
Was
there
I?
Just
you
know
what
the
spread
across
pulled
a
quest,
so
there
has,
to
you,
know,
put
it
all
together
and
fill
in
the
gaps,
so
I
think
Jacob.
Based
on
today's
discussion,
we
can't
deem
the
dev
work
for
this.
F
Epic,
complete
and
I
had
a
discussion
with
the
Dean
and
he
is
going
to
test
this
along
with
the
test.
Ground
test
is
that
is
already
running
and
look
mean
when
you
know
required
after
that,
I
reviewed
some
resources
to
come
up
with
a
proposal
to
separate
routing
table
membership
from
connectivity.
I
think
we
pretty
much
locked
down
the
approach
today,
based
on
the
marathon
discussion
we
had
with
Rowell
yeah
so
now.
My
next
major
goal
is
to
finish
the
work
to
you
know
separate
this
membership
from
connectivity.
That
will
be
my
focus
this
week.
F
A
If
you
need
more
time
like
these
are
all
just
rough
estimates,
so
if
you
need
some
more
time,
that's
fine!
Just
let
me
let
me
know
sooner
rather
than
later
and
I
can.
We
can
add,
like
a
couple
more
days
to
that
it
shouldn't
be
a
problem,
so
I
don't
think
it's
not
actually,
while
it's
in
the
timeline
I
haven't
actually
marked
it
for
the
milestone
that
we're
trying
to
hit
this
week.
So
there
is
some
flexibility
there.
F
Sorry,
just
to
give
some
more
detail
on
this
issue.
What
we
decided
in
the
discussion
is
that
basically,
any
strategy
that
you
use
to
separate
connectivity
from
membership
as
a
very
generic
scaffolding,
around
selecting
peers,
everything,
piers,
giving
scores
to
peers,
etc.
So
we
are
trying
to
come
up.
We're
trying
to.
You
know
like
make
this
as
much
pluggable
as
we
can.
So
if
you
want
to
quickly
change
the
implementation
tomorrow
to
run
some
tests
and
get
you
know
some
stats,
we
should
be
able
to
do
so.
H
D
H
I
Nothing
directly
specific
to
you.
Libby
2
p.m.
I've
been
drafting
the
outputs
of
research
intensive
workshop
into
RFC's.
All
of
them
are
going
to
land
by
well,
hopefully
more
than
a
couple
of
them
are
going
to
learn
by
the
ipfs
team
week
so
that
you
guys
have
a
look
and
we
try
to
formalize
as
much
as
possible.
I
What
we
want
to
expect
out
of
that
and
move
forward.
I
have
done
a
bit
of
a
brainstorm
on
this
v5
and
got
some
colleagues
are
interested
to
contribute
with
ideas.
We
haven't
yet
done
any
sinkhole
with
Felix
yeah.
It's
on
the
plan
to
happen
depending
on
availability
and
yeah
yeah.
That's
it
basically
I've
been
reading
a
couple
of
papers
on
pops
of
security
for
the
like
the
new
effort,
ro
mentioned
on
doses
of
security,
basically
not
nothing
tangible
to
date,
but
we'll
have
later
in
the
week.
J
J
You
know
we
kind
of
like
every
world
right
now,
because
security
is
and
we
don't
wanna
promote
early
all,
but
it
out,
if
you
guys
ask
me
or
Raul-
and
we
can
give
you
access
it,
sir,
and
basically
the
idea
is
we're
looking
at
specific
attacks
and
we're
friends
or
regret
now
we're
building
the
list
of
attacks
and
then
we're
gonna
develop
the
mitigations
which
do
forward
paths.
So
one
path
would
be
actually
to
test
the
intimidation
of
a
condom,
ezra
litigations
in
simulated
environment.
J
A
K
Yeah
well,
thanks
for
everybody's,
been
really
really
warm
and
welcoming
so
everywhere.
I've
kind
of
stumbled
through
out
the
slack
channels
or
an
email
everybody's
been
really
hospitable
and
helpful,
so
appreciate
that
I
am
bite.
Radio
am
a
technical
project.
Manager
might
come
to
that.
You
know
from
coding
years
and
years
ago,
so
I
know
a
little
bit
of
what's
going
on
and
specifically
the
networking.
A
D
We
only
have
to
do
this.
In
fact
it
land
actually
I
commit
it
to
it.
Well,
we
the
task
bound
team
committed
to
it
in
the
shipped
emo.
We
just
need
to
carve
out
time
to
plan
this.
It's
probably
going
to
be
on
Anton
side,
so
I'll
definitely
definitely
brought
him
about
this,
but
it
should
be
should
be
this
week,
ideally
so
that
everybody
that
also
is
going
to
I
give
has
team
week
next
week
comes
with.