►
From YouTube: 🖧 IPLD Every-four-weeks Sync 🙌🏽 2023-06-19
Description
An every four weeks meeting to sync up on all IPLD (https://ipld.io) related topics. It's open for everyone and recorded. https://github.com/ipld/team-mgmt
A
Welcome
everyone
to
this
ipld,
sync
and
community
meeting
is
June
the
19th
2023,
it's
the
middle
of
the
Year
hot
in
Germany
code
in
Australia
and
yeah
as
every
four
weeks.
We
talk
about
random
things
about
ipadi,
but
also
that
yeah.
If
people
have
worked
on
things,
we
go
over
those,
but
also
we
tend
to
have
presenters.
We
will
see
if
we
have
some
today,
but
also
if
you
are
interested
in
presenting
something
about
iPad.
A
So
it
does
really
have
to
be
about
like
developing
core
ipld,
but
can
be
just
like
uses
of
iPad
because,
like
it
is
from
my
perspective,
I
also
see
two
not
enough
examples
of
people
actually
using
it.
So
it
would
be
great
yeah
if
you
just
build
some
application
with
iPod
and
you're
happy
to
present
it.
Also,
it
doesn't
have
to
fill
the
meeting.
A
It
can
be
just
like
five
minutes
or
half
an
hour
just
as
you
bike
all
right,
so
I
need
to
pull
up
the
notes
document,
but
yeah
I
certainly
have
an
update
so
because
there
was
so,
as
this
meeting
is
only
every
four
weeks,
there's
more
things
that
happen.
So
there
was
a
rust
multi
hash
release
and
version
019,
and
it's
kind
of
a
big
release,
not
so
much
in
regards
to
Features
but
really
into
refactoring.
A
So
it
was
basically
someone
from
the
w2p
community
came
in
Thomas
and
he
wanted
to
reflect
the
model
from
rust
multiformis
in
a
way
that
there
are
less
breaking
changes
on
the
core
parts
of
it.
A
So
we
kind
of
now
have
the
system
split
into
several
crates,
and
so
we
have
the
core
part,
which
was
really
only
the
multi-hash
implementation
without
the
hashes.
So
in
case
you
have
been
with
the
humidity
for
five
years.
You
might
remember
that
we
had
something
similar
in
that
JavaScript
world,
where
we
had
JS
smarty
hash
and
JS
Martin
hashing.
A
So
we
kind
of
have
this
thing
as
well
in
the
rust
world
now
kind
of,
and
so
our
goal
is
or
also
like
how
it
works.
Is
that
if
you
similar
to
what
we
do
with
JavaScript,
is
that
you
don't
bundle
all
the
hashes
with
implementation,
because
then
you
have
a
huge
dependency
fee,
but
what
they
really
do.
Is
you
define
your
own?
A
What
we
call
codec,
cable
and
really
just
only
use
the
HS
that
you
really
want,
and
we
have
some
predefined-
you
can
import
them
and
so
on,
but
generally
I
should
use
your
own
Code
table.
This
was
the
idea
already
before
the
refactoring,
but
now
it's
kind
of
like
more
obvious
and
one
of
the
features
that
some
people
or
many
people
are
excited
about.
We
finally
removed
the
identity
hash
from
rust,
multi
hash.
A
So
basically
we
have
apis
for
the
hashes,
but
for
if
you
want
to
use
an
identity
hash,
you
kind
of
have
it
to
wrap
it
manually.
So
we
have
basically
a
function,
call
which
says
take
those
bytes
and
put
the
identity
thing
in
front,
but
it's
not
like
nicely
integrating
to
the
API,
because
it
costs
a
few
problems
and
it's
kind
of
like
now
more
obvious
that
you
probably
shouldn't
use
it
or
if
you
use
it,
you
should
really
know
what
you're
doing
foreign.
A
Also
one
of
the
reasons
because
like
what
we
do
in
Rust,
is
that
we
do
for
the
hashes.
So
we
basically
stack
allocate
the
things
and
put
on
the
Heap.
So
we
already
had
a
problem
that
identity
hash
had
basically
a
fixed
size
that
you
need
to
Define
upfront
and
this
then
the
problem
was,
if
you
didn't,
have
an
identity
which
is
bigger,
you
might
end
up
with
panics
and
so
on,
because
it
doesn't
fit
in
and
so
on.
A
So
now
like
this
is
kind
of
worked
around,
because
now
it's
like
more
explicit
that
yeah,
you
really
need
to
do
it
manually
explicitly
and
yeah.
So
we'll
see
how
many
like
complaints
about
reports
will
get
about
this
Behavior,
but
yeah,
we'll
see,
and
also,
as
always
with
the
rust
commodity
formats
and
I
put
this
stuff.
It's
really
open
for
contribution,
so
it's
nothing
set
in
stone.
A
So
if
you
have
better
ideas
about
apis
or
anything,
just
yeah
chime
in
and
report
on
the
repositories,
all
right,
yeah,
that's
what
I
had
in
my
world
I
guess
that
Rod
might
want
to
talk
about
frisbee
a
bit
I
guess.
B
Yeah,
so
it
always
is
a
long
time
yeah.
So
go
back
over
the
stuff
that
you've
done
so
the
the
main
thing
I've
been
working
on
it
is
still
Lassie.
So
Lassie
is
our
our
retrievable
client
that
that
speaks
all
the
different
IPL
lead
protocols
and
the
majority
of
the
work.
B
Well,
a
lot
of
the
work
around
retrieval
at
the
in
the
last
couple
of
months
has
been
around
the
the
new
trustless
Gateway
HTTP
protocol.
So
it's
been
a
big
push
across
multiple
teams
to
get
an
HTTP
protocol
landed.
That
is
is
good
for
ipld
and
it's
sort
of
it's
it's
it's
for
trustless
communication.
B
So
we
shouldn't
have
to
trust
that
the
the
other
party
that
you're
talking
to
you
should
be
able
to
incrementally
verify
the
data
that
you're
getting
and
it's
going
to
be
rolled
out
and
or
maybe
it
already
is
roll
down
on
the
ipfs
main
ipfs
gateway.
But
then
it
flows
all
the
way
through
the
back
end
of
the
whole
Gateway
architecture,
which
is
being
rebuilt
in
this
distributed
architecture
that
we're
all
currently
working
on
in
the
across
various
teams,
not
all
sorry
but
I'm
I'm.
B
On
one
team,
one
of
the
four
teams-
I
think
that's
that's
working
on
this
distributed
Gateway
thing
and
and
the
trustless
HTTP
protocol
is
something
that
is
being
used
by
various
backends,
including
web3
storage,
who
they
have
a
lot
of
popular
content
that
gets
used
by
the
Gateway.
So
so
now
the
flow
from
the
Gateway
will
will
be
going
through
this
sort
of
distributed
chain
of
nodes
that
are
running
a
trustless
way
that
are
able
to
fetch
fire
bit
swap
graph.
B
Sync
are
now
HTTP
and
so
went
through.
Storage
supports
HTTP.
Now,
so
you
can
fetch
you'll,
be
fetching
data
via
HTTP
to
go
back
up
the
chain
and
eventually
be
served
through
HTTP
from
the
Gateway,
but
there's
there's
some
potential
for
it
to
become
a
a
it's,
a
more
prominent
protocol
in
the
the
transfer
stack.
B
It
has
its
share
of
problems
just
like
every
other
protocol.
These
things
are
hard,
but
you
know
we're
working
through
all
the
problems
that
that
we
sort
of
know
from
all
the
various
other
protocols
that
we've
implemented.
B
B
So
if
you're
serving
in
HTTP
provider,
like
filecoin
providers
can
will
soon
be
able
to
switch
on
HTTP
mode,
you
can
front
your
HTTP
endpoint
with
nginx
or
various
other
services
to
do
all
sorts
of
things
that
are
easier
in
the
HTTP
world,
because
it's
tooling
for
it,
you
can
do
rate
limiting.
You
can
do
caching,
all
sorts
of
things.
B
These
things
already
exist.
We
don't
have
to
build
them
so
enable
the
protocol
and
then
plug
into
an
ecosystem
of
tooling,
which
is
kind
of
nice
when
a
lot
of
what
we
do
in
this
part
of
the
world
in
the
web,
3
world
is
just
invent
new
stuff
and
have
no
tooling.
You
have
to
build
it
yourself.
So
this
is
an
interesting
Adventure.
B
So
that's
from
the
retrieval
side,
but
the
other
thing
I
was
I've
been
working
on
in
my
my
spare
time
is:
is
the
flip
side
of
Lassie
and
a
little
project
I
released
over
the
weekend
called
called
Frisbee?
And
it's
a
it's
a
data
provider
that
provides
iPod
data
on
the
HTTP
trustless
Gateway
protocol
and
it's
I
built
it
because
one
of
my
frustrations
has
just
been
how
hard
it
is
to
just
serve
data.
Just
I've
got
ipld
data
I
want
to
share
it
between
two
nodes.
B
B
Does
a
lot
and
it's
sort
of
heavyweight
and
a
lot
of
stuff
gets
opaque
really
quickly
and
with
the
trustless
Gateway
protocol,
I
I
realize
that's
an
opportunity
to
do
something
that
was
really
minimal,
basically
lean
on
a
lot
of
other
pieces
to
get
the
work
done
and
write
a
really
simple
tool
to
do
it,
and
so
one
of
the
things
that
leans
on
is
the
the
network
indexes.
B
So
we
have
an
indexing
service
now
that
the
Distributors
a
distributed
set
of
indexes
that
index
content
across
the
ipfs
network,
and
so
instead
of
having
to
do
the
whole
Discovery
thing
you
know
advertise
on
the
DHT
and
all
that
stuff,
which
is
you
know,
I'm
I,
love
doing
that
stuff.
It's
really
not
what
I
really
want
to
be
doing
so
so
now.
B
I
can
hand
that
off
to
the
indexer
and
say,
hey,
here's,
my
content
come
and
get
come
and
get
it
and
then
have
clients,
talk
on
the
HTTP
protocol
and
come
and
fetch
content.
So
anyway,
maybe
I'll
do
a
talk
about
that
in
the
next
time.
Unless
we
have
folks
come
and
other
folks
come
and
do
it
might
might
do
a
talk
about
the
trustless
Gateway
protocol
and
maybe
frisbee
as
well.
B
But
if
you
want
to
check
out
frisbee
it's
in
the
ipld
org
on
GitHub,
the
name
is
Frisbee
with
a
double
I
on
the
end,
frisb
Double
I
and
you
can
run
it
as
a
CLI
app.
All
you
do.
Is
you
point
it
at
car,
Files,
full
of
data
and
it
will
go
and
just
advertise
and
then
serve
that
data
and
be
serving
you
know,
requested
dags
as
cars
via
HTTP,
nice
and
simple
and
and
right
now,
I've
got
one
I've
got
an
instance
running.
B
That's
serving
the
latest
snapshot
of
Wikipedia
that
that
is
being
shared
across
the
ipfs
network,
so
I
have
Wikipedia
being
served
off
a
a
single
node
off.
It's
like
500
car
files,
280
gigs
of
data
at
a
time
away
at
one
gig
of
memory
serving
these
car
files
into
the
distributed
Gateway
system.
So
working
quite
well
and
it's
it's
quite
fun
to
to
see
IPL
lead
data
as
HTTP.
So
that's
that's.
What
I've
been
up
to.
A
Cool
I'm
so
excited
to
see
again
like
more
like
small
ipld
pieces
coming
together.
C
A
Like
I,
always
like
everybody
see
five
years
ago,
I
was
like
oh
I
need
to
do
something
like
pure
ipld,
without
the
ipfs
kind
of
all
those
things
and
yeah.
So
finally,
we
get
there.
So
it's
super
cool
to
see
yeah
all
right.
Does
anyone
else
have
any
update
or
wants
to
share
things
about
iPod.
C
I
just
wanted
to
say,
hi
and
also
that
Rod
I'll
probably
be
pushing
A
couple
more
experimental
updates
to
that
PR
we
were
discussing
I,
think
I'm
getting
a
handle
and
where
I
want
it
to
go
so
I'll
I'll
have
something
for
you
soon.
C
Yeah
yeah,
so
it's
been
a
while
that
I've
been
working
on
a
way
to
be
able
to
update
my
ability
nodes.
It's
currently
kind
of
difficult
to
update
nodes.
Most
logic
creates
new
nodes
that
replace
older
ones
and
that
can
get
very
complicated
and
non-optimal
with
a
with
a
nicer,
more
efficient
way
to
update
nodes.
It'll
really
make
a
lot
of
the
logic
easier,
more
efficient
and
simpler.
C
B
So
that
that
PR
is
going
in
the
goibly
prime
repo,
if
anyone
wants
to
to
have
a
look
and
participate
and
if
they
have,
if
they
have
a
use
case,
if
you
have
a
use
case
where
you
are
you're
mutating
iPod
data
through
ipld,
Prime
and
and
you're
interested
in
how
that
could
be
made
more
efficient,
then
that
might
be
a
place
to
go
and
participate
and
check
out
some
of
the
benchmarks
and
the
apis.
A
A
A
So
we
give
a
talk
at
geospatial
conference,
the
phosphogene
and
so
it's
kind
of
ipd
related
because
of
what
I
really
so.
What
I
always
wanted
to
do
is
that
I,
you
can
do
browser
to
browser
sharing
of
data
without
internet
connectivity.
So
if
you
are
in
the
same
Wi-Fi,
you
can
just
with
web
RTC
connect
to
the
browser
directly
without
any
server
infrastructure.
So
that's
the
idea
and
of
course
I
will
then
obviously
I
don't
really
want
to
use
it
for
files.
A
I
really
want
to
use
it
for
like
structured
data,
so
this
is
kind
of
like
the
hyperl
D
relationship,
but
the
release
tiny
Panasonic
iPod
because,
like
I,
spend
so
much
time
on
Twitter
getting
it
work.
So
the
idea
is
that
it
basically
scan
QR
codes
of
each
other
and
then
you'll
make
the
connection,
and
it
ended
up
in
like
a
huge
Rabbit
Hole.
A
My
JavaScript
thing
with
like
like
lift
copy
and
everything
is
around
like
one
and
a
one
and
1.5
megabytes,
so
I
had
to
strip
down
the
the
image
of
it
and
then
I
found
out
that
there's
also
a
Rudy
which
I
know
Port
is
where
you
can
plug
in
a
USB
drive.
So
basically,
I
would
then
want
to
serve
map
data
because
it
should
be
a
memory
application.
A
So
you
serve
up
your
map,
data
from
a
USB
drive
if
the
application
on
the
router
and
basically
you
just
plug
in
the
router
and
then
you
connect
to
it.
And
then
everyone
gets
like
this
Linux
page
and
then
you
have
you
enable
your
webcam.
You
have
a
QR
code
and
then
you
just
scan
each
other
and
then
you
are
connected
and
can
then
map
things
on
the
map
and
once
you're
done,
you
can
then
duplicate
the
data
back
and
everyone
has
the
data.
A
So
that's
the
idea
and
the
hard
part
really
is
getting
it
right
across
browsers
and
I've
spent
so
much
time
on
getting
it
worked
with
Safari
I
think
I
I
once
managed
to
connect
my
Android
device
with
Firefox
to
my
friend's
iPhone,
but
I'm,
not
sure
if
it
was
really
true
or
if
I
dreamed
it
up,
because
I
just
can't
get
it
working
anymore.
A
So
I've
spent
like
last
weekend
a
long
time
trying
to
get
Safari
on
Macos
work
with
directly
with
something
else
and
my
conclusion
what
which
so
clearly
so,
if
anyone
wants
to
help
or
hasn't
touched
about
webrtc
on
make
OS
and
so
on,
please
contact
me
what
I
found
out
what's
interesting
is
so
I
can
so
the
code
works
with
one
Safari
window
to
the
other
Safari
window
in
the
same
browser
works.
A
I
can
make
the
connection
it
works
with
fire
folks
and
Safari
on
the
same
machine,
but
it
doesn't
work
from
Safari
on
Mac
to
a
Linux
machine
in
the
network
with
Firefox.
It
doesn't
not
work,
but
what
works
is
Firefox
and
Mac
OS
over
the
network
to
a
Firefox
in
Linux
that
one
works.
What
also
does
not
work
is
Safari
to
a
chromium.
A
You
see
on
the
same
machine
it
does,
but
on
other
machines
it
does
not
yeah.
So
I
had
hope,
basically
for
the
talk
with
everything
to
say:
okay
put
audio
phones
and
make
it
work.
So
basically
now
people
with
iPhones
are
excluded,
but
it
I
really
think
it's
a
safari
bag
I'm,
not
sure,
and
also,
of
course
like.
Basically,
it
led
to
like
so
it
so
as
I
base
it
on
dp2p
and
so
on,
and
on
the
new
web
RTC
transport.
A
So
also
yeah,
thanks
to
Alex
aching
brain
who
also
fixed
bugs
that
I
discovered
when
I
did
like
my
stupid
experiments,
so
they
were
actually
bugs,
but
now
it
seems
to
be
pretty
solid
but
yeah.
Those
details,
I
just
yeah
I-
have
no
idea.
I've
checked
really
like
the
low
level
webrtc
stuff,
the
the
the
ice
candidates.
It
gets
the
ice
candidates,
I,
don't
know
and
yeah
anyway,
so
yeah
if
anyone's
interested
in
those
things
so
I
will
so
I've
pushed
already
on
GitHub.
But
it's
really
like
just
a
broken
progress.
A
So
basically,
next
meeting
I
have
something
proper
because,
like
for
the
conference,
I
need
to
publish
it
properly
and
I
might
also
like
yeah,
give
a
talk
in
this
meeting
or
some
or
something
about
it,
and
but
yeah
I
think
it's
really
cool
if
it
really
works
and
to
conclude
so
that,
basically,
why
I
built
this
is
it's
also
an
idea.
I
had
to
read
for
years
is
I,
see
it
as
I
come
from
a
true
spatial
world.
Really
in
the
these,
that's
the
response
loose
case.
A
Where
I
mean
you
don't
have
internet
because,
like
it's
I
mean
getting
up
a
router
with
some
HTTP
Statics
HTTP
files
on
it
is
super
easy
way
easier
than
getting
some
internet
productivity
on.
So
that
would
mean
media
basis.
You
then
have
some
like
headquarter
where
you,
basically
everyone
in
the
morning,
connects
with
their
phones.
Then
they
do
the
mapping
or
collecting
data
on
their
phones
and
then
they
get
back
replicated
and
then
you
have
all
the
data.
So
that's
kind
of
what
I
want
to
build.
A
A
Webrtc,
yes,
yeah
and
also
like
I,
do
like
I
want
to
make
it
a
bit
simpler.
So
basically
I
also
do
the
what's
called
managing.
If
people
are
familiar
with
it,
so
you
basically
you
menu
play
the
STP
responses.
You
get
so
you
have
one
less
Roundup
of
the
data,
so
it's
kind
of
like
outside
of
the
standard,
but
it's
but
it's
supported
by
all
browsers
because
it
used
to
work
and
so
on
so
kind
of
like
this
might
also
be
the
problems,
but
it
doesn't
work
on
Safari
and
so
on.
B
There
was
some
Safari
things
that
were
fixed
in
the
latest
announcement.
The
latest
OS
announcement.
Wasn't
there?
Yes,.
A
Yeah,
so
this
do
they
touch
on
this,
maybe
but
yeah
still
like
still
doesn't
work,
so
I've
tried
the
latest
version
and
yeah
but
yeah.
So
yes,
so
the
RTC
is
certainly
tricky
and
yeah
and
yeah
and
I
was
wondering.
Let's
see
why
why
didn't?
No
one
else
did
I
mean
people,
that's
just
similar
basic
things
with
QR
codes
and
web
RTC,
but
nothing
like
more
elaborating
and
I
was
wondering
why,
but
now
I
know
so.
A
Yeah
yeah
we'll
see
yeah.
This
was
my
fun
project,
all
right,
I
guess.
If
no
one
else
has
any
updates
or
anything
to
say.
B
But
for
the
benefit
of
the
of
those
who've
tuned
in
and
for
our
advertised
content,
we
did
have
a
couple
of
people
lined
up
to
do
some
presentations
and,
but
unfortunately,
I
think,
we've
hit
a
US
public
holiday
and
for
people's
calendars
are
not
being
checked
and
so,
unfortunately,
we're
without
our
presenters.
But
hopefully
next
week
and
next
month
we
will.
We
will
bring
some
more
content.
Yes,.