►
From YouTube: libp2p Bi-Weekly 9/10/18
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).
B
B
Attempts
were
just
wrong,
like
I
was
trying
to
test
it
swap
over
localhost,
which
was
not
the
right
way
to
test
this
I
think
we
reduced
the
number
hackers
are
sending
both
sniffing
reasonable
amount
of
back
grab
and
which
we
have
for
the
DHT
and
the
speed
for
their
site,
though
speed
in
which
we
can
transfer
gate
across
the
open
Internet
where
we
actually
drop
packets.
So
if
any
wants
to
pick
this
up,
there
is
an
issue
in
Goa
fest
called
better
buffering.
B
C
Next:
okay,
roll
yeah,
hey
guys
so
I've,
been
working
on
the
pier
store
right
now.
I'm
gonna
be
focusing
this
week
and
hopefully
we'll
get
it
rounded
up
this
by
Friday
on
the
caching
aspects
on
the
caching
layouts.
We
want
to
make
the
caching
eerie
through
cache
such
that
when
you
get
a
hit
on
the
cache.
You
know
you
have
all
the
multi
address
for
the
pier,
so
you
don't
need
to
scan
the
database
and
get
a
essentially
do
a
prefix.
Can
there
I
think
that's
gonna,
give
us
some
pretty
good
improvement.
C
C
What
we're
finding
solutions
for
that
and
and
yeah
I
put
up
a
proposal
about
an
analysis
that
I
did
on
the
DHD
and
the
traffic
that
we
generate
and
the
way
that
we
perform
lookups
and
how
the
routing
table
the
cadena
routing
table
is
actually
populated
and
the
utilization
there
I
think
a
few
of
you
in
that
issue.
So
if
you
can
feel
free
to
just
contribute
to
the
discussion
there,
we
consider
an
alternative
data
structures,
and
that
could
be
an
interesting
direction.
To
take
from
aware
need
to
take
a
word.
D
F
Guys
so
last
week,
I
almost
finished
all
the
PRS
regarding
I
pianist
were
pops
up
on
Jesslyn.
They
will
hopefully
be
reviewed
this
week
and
afterwards
merge
and
I.
Meanwhile,
I
will
also
make
interrupts
tests
for
this
afterwards.
I
will
start
probably
looking
at
the
DHT
interrupts
and
I
pianist
over
the
HD
also
on
GSI,
and
it
will
be
it
for
me.
Any
questions.
F
G
A
H
So
last
week
I
was
pretty
much
finishing
up,
delegator
routing
stuff.
They
done
with
that
and
IP
if
it's
all
hands,
but
it
did
not
get
recorded.
So
I
will
probably
try
to
do
a
recording
at
some
point
of
that.
I
just
need
to
get
the
demo
up
there
and
Lars
is
working
on
finishing
up
the
glide
PFS
implementation
of
that.
Hopefully
this
week
or
next
week
and
I'm,
currently
working
on
updating,
J
slippy
to
be
switch
to
make
that
whole
thing,
including
connections,
all
state
machine
driven.
H
A
D
Yeah
I'm,
just
really
starting
the
very
first
thing,
I'm
looking
at
I,
actually
am
not
sure
whether
it
works
this
way
down
to
lid
p2p
land
I
know
that
little
but
I'm
working
on
faster
directory
listing
of
sharded
directories,
which
is
evidently
an
issue
on
some
very
large
directories,
which
is
kind
of
right.
In
the
realm
of
where
we
get
data
back
from
the
DHT
and
how
we
decided
to
proceed.
D
So
it
seems
like
a
good,
a
fruitful
place
to
begin
learning
about
liquor
and
I'll,
probably
be
pairing
with
Michael
today
and
probably
say:
looping
Hanna
in
tomorrow,
all
stare
at
screens
and
be
like
wow.
This
is
crazy
and
hopefully
we'll
learn
and
I'll
have
more
interesting
stuff
to
report
next
next
week-
or
maybe
this
is
every.
A
I
Oh,
hey
everybody,
Steve
Richmond
here
I
saw
a
lot
of
you
on
the
ipfs
call
I'm
really
here
for
a
few
different
reasons.
One
is
just
to
learn
more.
You
guys
are
up
to
incredible
things
and
I'm,
seeing
some
of
you
for
the
first
time
the
other
is
to
plug
lab
OS,
which
is
a
group
here
at
protocol
labs
that
I
helped
lead,
and
one
of
our
core
functions
is
helping
do
planning.
I
J
D
I
Also,
you
know
we're
looking
our
our
mandate
is
really
to
support
all
the
teams
in
the
organization,
so
our
repos
open.
We
have
an
open
issue
there.
How
can
we
support
you
better
as
ideas
come
to
mind,
please
hit
that
we're
gonna.
Do
some
work
at
Team
week
to
kind
of
prioritize
and
really
start
implementing
some
these
things
and
you
can
see
what's
in
flight
and
what's
what's
been
proposed,
but
would
love
any
feedback
there
as
well?
So
thanks,
everybody.
A
K
K
What
it
does
is
that
it
creates
a
service
that
the
that
you
can
request
to
dial
you
back
through
a
different
cost
object
so
that
you
can
detect
your
network
bootable
disk
space,
or
you
know
in
a
very
practical
way,
and
it
also
adds
an
ambient
or
the
alternate
discovery
service.
So
basically,
when
you
discover
providers,
please
dial
back
service,
not
not
service,
then
you
can
use
this
ambiently
to
start
trying
the
hell
box
every
now
and
then-
and
basically
this
gives
you
there
you're
not
status.
K
A
J
A
B
B
Much
maybe
announced
with
the
I
prefer
him,
but
he
to
be
cat
is
a
simple
net
cat-like
utility
for
I
favor
for
them
to
be,
but
a
cat
is
that
a
small
utilities
for
decoding
and
encoding
protobufs
I
can
show
my
screen
and
do
a
very
short
demo.
This
is
really
simple.
If
I
can
share
my
screen
there,
we
go
okay,
so
see
if
it
works.
B
We
need
a
classroom.
Okay,
so
the
first
thing
you'll
need
to
do
is:
go
get
the
cat
and
proto
cat
and
run
the
command
here
run
this
command.
It
should
install
both
of
them.
You
have
to
put
your
go
pass
in
in
your
go
pad
next
for
this
command.
We're
actually
are
going
to
go
to
identify
a
server,
so
we
need
to
make
sure
we
have
the.
B
Okay,
show
this
I
need
to
make
sure
we
actually
have
the
right
part
of
us.
So
I'm
gonna
go
get
these
si
protobuf.
Next,
we
can
decide
whichever
we
want
to
talk
to
in
this
case.
Let's
talk
to
a
server
called
Mars,
so
I'm
going
to
set
the
multi
editor
to
that.
This
just
gonna
make
the
command
simpler.
Finally,
we'll
decide
which
program
you
want
to
use.
In
this
case,
I'm
gonna
use
the
idiot
or
I
guess
I
heard
of
us.
B
So
this
index
here
is
a
head
of
the
dove
shortcut
where
I've
added
go
pass
to
the
protobuf
path.
So
you
can
give
heard
about
definitions
relative
to
go
pass.
You
can
also
give
them.
You
know
just
give
absolute
proto
definitions,
but
this
is
basically
just
like
pass.
The
parameter
position,
drop
the
doc
koto
and
add
the
message
type.
So
this
is
from
the
identify
little
file.
Well,
they
identified
it.
Just
simple
I
think
that
set
that
up
and
then
the
command
is
quite
simple
here
it
is.
B
F
B
So
don't
have
to
manually.
Put
me
at
the
address
matter
is
the
multi
editor
we
specified
up
here
what
the
protocol
is.
Ib
first
ID
equal
point,
therefore
zero
and
then
I'm
here,
I'm
just
closing
standard
in
because
we're
not
actually
said
anything
finally
I
pipe.
The
output
to
proto
cat
L
means
that
we
using
a
link
delimited
stream.
That
means
every
prototype
is
prefixed
by
a
length,
they're,
sorry,
buddy,
apparently,
and
dash
D
means
decode
and
then
I'm
specified
the
Puerto
Rico's.
B
B
B
Okay,
so
this
should
allow
me
to
try
to
use
the
trick
of
relay
so
connected
now.
I
could
just
send
an
empty
request
in
this
case.
So
that
view,
of
course,
and
say,
I
can't
like
I
can't
do
this.
You
could
also
say
type
stop
and
then
it's
gonna
form
me
so
I
can't
use
channel
pretty
no
one.
So,
let's
see.
B
A
A
Basically,
if
you
look
in
the
doc
I
linked
to
the
q3,
so,
okay,
the
quarter
of
the
attending
in
the
next
few
weeks,
we're
now
going
to
go
into
a
planning
process
where
we
define
the
goals
for
the
fourth
quarter,
and
what
I
want
to
try
to
do
is
a
couple
things.
First,
83
she
became
extremely
and
I'm
gonna.
Try
to
sort
of
minimize
that
complexity,
I
think
reached
a
point
was
hard
to
follow.
We're
looking
at
retro
on
the
q3.
It
goes
how
he
did
so.
A
I
Yeah
so
it'll
be
about
it's
gonna,
be
quick
at
the
project
level,
though,
for
you,
like
you,
know
about
five
minutes
on
kind
of
your
retro
and
what
you
discovered
on
q3,
even
though
we
won't
quite
be
at
the
end
of
q3
by
then
we're
gonna
kind
of
try
to
push
that
forward
and
the
rest
of
the
call
ten
minutes
or
so
I'm.
Just
talking
about
planets,
Ricky,
four,
okay,.
A
Thanks:
okay,
so,
in
addition
to
the
spreadsheet
being
too
complicated,
I
think
there
were
two
other
problems
with
how
we
did
q3
okay,
ours.
So
first
off
they
ended
up
getting
done
in
a
room
in
Berlin
and
it
was
sort
of
like
whoever
happened
to
be
in
that
room
at
the
time
was
was
dictating
the
okay
ours.
What
I'd
like
to
have
this
time
is
a
more
open
collaborative
process
where
you
know
we
do
it
over
github.
Everyone
participates
and
we
just
saw
something
a
little
more
open
than
do.
A
You
know
the
twelve
random
people
who
happen
to
be
in
Berlin.
The
second
thing,
I
want
to
say
is
part
of
why
that
happened
was
I
made
a
number
of
attempts
to
get
folks
to
engage
with
the
okay,
our
process
and
the
JavaScript
team.
Jit
did
like
a
plus
work
like
they're
excellent.
The
go
team
was
a
little
more
kind
of
ignored
my
pings
and
and
seemed
more
reluctant
to
engage.
A
So
what
I
would
like
to
suggest
is
I'm
gonna,
try
to
schedule
an
actual
synchronous
meeting,
I
first
I'm
gonna
post,
a
doc
async
on
github
and
I'm
gonna
reach
out
to
folks
and
ask
people
to
propose
items
that
that
should
be
in
the
okay
ours,
I'll
define
with
de'vide
what
the
overall
project
okay.
Ours
are.
A
So
that's
like
the
first
tab
of
the
spreadsheet,
but
then
for
the
individual
languages,
I'm
going
to
ask
the
team's
to
propose,
based
on
by
looking
at
what
the
overall
project
goals
are
and
then
I
want
to
have
a
synchronous
meeting
like
this
one
Ted
for
us
to
discuss
that
and
deaths
will
be
open
to
the
public
and
I'm,
hoping
that
if
we,
if
we
do
all
that
right,
we
won't
end
up
in
a
situation
like
last
time
where
we
were
past
the
deadline.
There
were
no
okay,
ours,
and
so
there
happened
to
be.
A
The
last
thing
I'll
mention
is
what
I
really
like
to
do
is
so
so
meet
me
and
daveed
will
define
the
overall
project
LePew
project
goals
for
the
quarter.
My
hope
is
that
individual
people
might
step
up
and
take
responsibility
for
go.
Jas
and
rust.
I
mean
people
from
those
teams,
so
we
don't
have
to
sorted
out
right
now,
but
I
would
love
it
if,
if
other
people
stepped
up
yeah,
not
not
so
much
that
I
mind
doing
the
work,
but
I
think
that
the
I
think
it
may
bubble
up
things
that
as
a
p.m.
A
I'm,
not
really
aren't
really
on
my
radar
but
are
actually
important
so
I.
Might
they
reach
out
to
folks
to
volunteer
if
you're,
if
you're
willing
so
yeah?
That's
all
I
wanted
to
say
about
Opare
is
I,
just
want
to
warn
you
it's
coming
and
yet
that's
the
sort
of
good.
Can
we
talk
about
Lupita?
Be
demon,
cool?
Oh
sorry,
are
there
any
questions
right
now.
E
That
sounds
great
yeah,
okay,
awesome,
yeah,
so
I
wanted
to
open
a
little
discussion,
I
guess
about
the
PDP
daemon.
This
has
been
kind
of
requested
in
a
couple
of
contexts.
The
original
context
that
came
across
my
plate
was
from
some
of
the
people
working
on
a
theory
in
2.0
which
or
sorry
working
on
the
sharding
implementation
for
etherium
and
they're
operating
in
Python
and
wanted
to
be
able
to
abstract
out
the
peer-to-peer
networking
layer
from
their
implementation
without
fully
implementing
limb
p2p
themselves.
E
E
So
ideally,
we
can
have
a
single
specification
for
like
command
specification
where
we
could
send
yeah.
We
could
spawn
up
a
limb,
p2p
Damon
and
then
send
commands
to
say,
set
up
a
host
attach
a
DHT
to
it.
Execute
this
query
over
it
expects
this
result.
Something
like
that.
I
think
Steven
made
a
good
point
in
the
last
meeting
that
it
probably
makes
sense
to
treat
them
as
completely
separate
use
cases.
I'm
not
worried
necessarily
about
huge
design
overlap,
the
more
I
yeah
I,
don't
know.
E
I
guess
I
I'm
very
curious
if
people
have
like
initial
thoughts
on
like
general
design,
the
other
thing
that
I
wanted
to
bring
up
is
that
I
was
like
looking
into
the
ipfs
Damon
and
Steven.
You
might
be
a
good
person
to
give
me
a
little
insight
on
that.
It
looks
like
it's
really
heavily
married
to
the
go
ipfs
codebase.
It
also
looks
like
it's
kind
of
just
wrapped
over
the
default
net.
Http
library,
so
I
was
just
wondering
if
there
is
from
like
a
reuse,
standpoint
or
extraction
modularity
standpoint,
so
yeah.
E
E
K
K
Maybe
we
can
set
up
a
call.
The
other
thing
is
definitely
so
glad
service,
it's
something
you
can
do
by
simply
opening
e
by
using
the
steam.
Your
face,
I
mean,
if
you
want
to.
If
you
want
to
do
your
own,
complete
implementation,
perhaps
of
protocols,
you
can
do
that
with
the
same
interface,
but
you
should
also
get
some
high
level
interface
and
something
that
does
RBC's
know
like
just
like
you
guys
suggest
it
with
in
VHD.
Queries
will
be
great
if
we
say:
ok,
please,
you
know
like
connect
to
this
operator.
K
B
The
way
Bastion
it's
like
the
first
step
would
be
allowing
users
to
register
custom
vertical
handlers
and
like
next
to
custom
verticals.
So
it's
basically
just
dream,
that's
it.
So
this
allows
us
to
share
connections
between
multiple
applications,
but
it
wouldn't
do
too
much
like
fancy
stuff.
The
next
step
would
be
like
allowing
used
efficient
like
sub
peers
behind,
like
they're
Damon.
So
again
they
can
share.
They
can
have
their
own
Prezi
their
own
system.
Mister
like
this
separate
application,
but
it's
sharing
the
connections
from
main
Damon
yeah.
B
That's
basically,
an
idea
so
like
we
can
also
have
like
will
also
want,
like
external
services,
so
we
actually
nice
to
break
out
the
DHT
and
stuff
like
that
like
separate
Damon,
but
all
connect
to
the
new
test
David.
But
that's
a
bit
tricky
we're,
not
goodbye.
Sorry
I
didn't
go
to
be
given
Authority
III.
E
Think
you've
both
kind
of
hit
on
something.
That's
definitely
useful
in
that,
like
of
a
simple
low-level
core
that
we
could
wrap
with
the
Damon
would
just
be
pure
peer
and
or
content
routing,
and
then
just
kind
of
opening
up
streams.
On
top
of
that,
yeah
I
think
that
sounds
great
I.
Think
for
the
so
I
guess,
I'm
I'm
envisioning,
then
something
where
you
could
kind
of
interact
with
the
pure
store
and
then
interact
with
the
host
for
opening
connections
and
then
I
guess
you
could
kind
of
have
some
sort
of.
E
B
B
E
K
K
My
can
you
be
to
peel
an
and
you
can
specify
how
it
connects
and
basically
just
kept
an
interface
for
saying,
hey,
open
the
same
on
this
protocol
to
this
fear,
so
just
that's
level
zero
so
that
we
can
even
implement
you
know
like
these
de
clients
in
other
languages,
so
that
should
be
our
fundamental
functionality
and
then
we
can
think
about
adding.
You
know
they
call
subsystem
interfaces
but
having
the
in
to
be
directional
streams
and
they're
registering
stream
handlers.
That
would
be
a
great
starting.
K
E
For
yeah,
I
think
I
think
that
it's
good
and
also
keeps
it
very
focused
on
the
core
so
could
I
jot
down
that
essentially
like
pierced
or
interaction
for
for
the
sake
of
adding
pure
IDs
and
stuff
when
you
need
to
get
bootstrapped
and
stream
management.
Opening.
Reading
writing
and
closing
that
sound
like
a
reasonable
first
start:
cool.