►
From YouTube: 🖧 IPLD Every-two-weeks Sync 🙌🏽 2022-05-09
Description
An every two 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
A
All
right
and,
of
course
like,
if
you
attend,
please
add
your
name
to
the
handys
list
this
week.
Even
I
have
an
update.
Finally,
because
I
worked
a
bit
on
rust
things
and
iple,
and
I'm
happy
to
announce
that
there
is
now
a
new
dexy
board
codec
and
it's
based
on
30
and
for
those
not
familiar
with
rust.
30
is
for
serialization
and
decibelization,
and
it's
basically
the
standard
that
you
use
in
the
roth
world
and
in
the
past
there
have
been
discussions.
A
If
you,
for
example,
want
to
do
things
like
dates,
which
is
not
part
of
the
data
model,
but
you
want
to
represent
dates
somehow
and
we
use
a
similar
hack
in
order
to
get
cids
working
because,
like
the
rest
of
the
data
model,
matches
quite
well
between
ipld
and
sorting,
but
cids
do
not,
and
we
found
a
good
hack
for
it
or
like
basically,
the
heck
that
everyone
is
using
in
the
ecosystem
for
such
cases
and
what
it
means
for
you.
A
If
you
want
to
use
it,
you
can
kind
of
just
yeah
use,
it
derive
your
civilization.
This
is
a
deserialization
automatically
and
get
zebra
out,
and
currently
it's
the
only
coding
we
have
so
hopefully
perhaps
someone
from
the
community
comes
up
with
a
dac
json
one,
because
what
you
then
can
do,
which
is
pretty
nice.
A
You
could
then
directly
convert
back
back
keyboard
to
json
and
back
without
a
real
intermediate
step.
So
you
don't
really
serialize
it
into
a
the
ipl
data
model,
but
you
stay
within
the
survey
data
model,
which
would
be
quite
nice.
What
you
need
to
do
is
so
what
we
need
to
do.
First,
for
sieber
was
that
you
need
to
fork
the
an
existing
library
to
add
support
for
the
cids,
but
that's
mostly
it
so
compared
to
the
upstream
version
of
this
library,
there's
not
that
much
difference
normally,
but
especially
for
dachshunds.
For
example.
A
A
So
the
underlying
zebra
parsing
is
done
by
us
library
called
keyboard
for
ii,
and
I
want
to
mention
it
because
the
author
was
really
great
and
helping
with
the
development
and
like
when
I
had
questions
super
responsive
and
like
it's
a
super
nice
open
source
project
and
I'm
happy
to
help
out
there
and
yeah.
A
It
was
really
good
good
working
with
them
and
together
on
this,
and
they
also
did
releases
when
we
needed
them,
and
so
because,
like
this
project
was
also
like
it's
critical
or
was
critical
for
the
flightcard
virtual
machine,
it
was
great
to
basically
have
support
from
the
open
source
world.
Basically,
someone
is
responsive
and
helps
out
yeah,
really
great
work,
so
shoot
out
with
your
author.
B
Oh,
yes,
I'm
next,
all
right,
so
I've
been
doing
some
work
on
webassembly,
plus
ipld
things
which
I've
mentioned
previously,
where
we
less
laptop.
Where
we
last
left
off
was
we
had
adl
support
for
large
bytes,
and
that
was
it.
B
I've
been
working
on
on
sort
of
reworking
the
apis
to
be
reuse,
more
components
and
be
a
little
more
like
less
tailored
for
each
individual
function.
So
you
can
sort
of
share,
share
pieces
and
include
things
like
returning
error
codes
instead
of
just
panicking
and
dying
and
working
on
a
bittorrent
directory
adl.
B
Basically,
I
wanted
an
example
of
a
recursive
type
as
opposed
to
a
plane
type
like
like
bytes
or
strings
or
integers,
or
something
because
then
there's
there's
more
complexity.
There
and
yeah
there's
still
a
little
bit
to
go
there
of
me
fighting
with
rust
things
that
I'm
still
new
at,
but
next
up
will
be
finishing
that
reworking
the
go
bindings
so
that
they
can
then
handle
the
new
adl
types
and
then
we'll
be
able
to
see
a
nice
demo
where
my
koala
loads
from
a
directory.
B
Instead
of
just
as
the
root
object
of
a
thing,
there
are
sort
of
some
interesting
things
that
we
get
to
now.
You
know
revisit
or
re-examine
as
part
of
this
too,
which
includes
and
there's
probably
like
aside
from
all
the
adl
stuff,
there's
some
like.
Let's
look
at
what
happened
with
bittorrent
post-mortem
things
to
to
look
at
whether
it's
like.
Where
do
we?
You
know?
Where
do
we
decide
to
draw
the
codec
versus
adl
boundaries
or
is
the
unix
fs
style
where
you
enter
the
unix
fs
adl?
B
And
then
you
just
live
there
forever
right,
as
you
recurse,
through
directories
and
into
files?
Is
that
the
strategy,
or
do
you
go
like
one
layer
at
a
time
and
then
reinterpret
and
then
one
layer
at
a
time
and
reinterpret
right
like
do
you
assume
from
now
on,
everything
is
a
as
a
bittorrent
directory
or
is
it
like
just
the
first
thing's
a
bit
torrent
directory
and
then
who
knows?
B
I
followed
the
unix
fs
model
for
now
because
it
seemed
natural,
also
the
directory
structure
but
yeah
some
interesting
things
that
will
emerge
from
that
and
hopefully,
once
that
looks
a
little
more
complete,
then
we
can
start
I'll
start
rolling
out
more
specs
around
yep.
Okay,
this
thing
actually
works
on
more
than
just
the
one
example
that
I
used
because
now
it'll
have
support,
for
you
know
recursive
types
and-
and
you
know
things
like
bytes
and
that's
all
for
me.
A
Thank
you
next
is
move.
C
Okay,
so
I
I
did
lots
of
little
things.
Last
couple
weeks,
I
took
little
things
around
ipod
and
multicode
multiformats.
The
main
thing
that
I
got
over
well
nearly
got
over.
The
line
is
the
unified
ci
stuff
for
guy
ipl,
prime,
so
kind
of
working
on
mac
os
windows
got
all
the
vet
and
static
check.
C
Fixes
done
figured
out
a
strategy
for
getting
the
tests
working
in
ci,
with
the
the
way
that
the
unified
ci
wants
to
work
and
the
main
barrier
there
has
been
because
it
wants
to
use
coverage,
the
plug-in
tests
don't
work,
and
so
the
strategy
is
to
find
a
way
to
disable
the
the
plug-in
tests,
and
I
was
yeah.
There's
we've
come
up
with
a
couple
of
ways
of
doing
that,
but
I
think
we've
got.
C
We've
got
a
good
way
that
doesn't
require
upstream
changes
to
the
unified
ci
pipeline
and
then
introducing
a
second
lot
of
tests.
C
You
know,
like
a
second
action
that
just
that
that
runs
them
without
coverage
just
on
linux,
doing
the
plug-in
tests.
So
we
actually
get
proper
coverage
in
ci,
so
there's
a
couple
of
hoops
to
jump
through,
but
it
means
we
get
some
pretty
good
coverage,
so
I
think
I'll
be
able
to
merge
that
and
get
that
fixed
today.
So
that's
good.
It
feels
nice
to
get
that
done.
I
had
a
hack
week
for
my
team
not
last
week.
Oh
was
the
last
thing
I
don't
even
know.
C
I
think
it
was
two
weeks
ago,
two
weeks
ago,
it
was
two
weeks
ago
and
I
spent
most
of
my
time
lots
of
little
things
around
iple,
prime
and
some
other
places,
but
I
spent
most
of
my
time
on
a
packing
go
bitcoin,
bitcoin
codec,
just
it's
just
a
niche.
I
wanted
to
scratch
and
I
had
a
lot
of
fun
doing
that
not
finished,
but
it's
I
keep
on
going
back
to
it
in
some
spare
time
to
tinker
with
it
to
try
and
get
over
the
line.
C
I
really
like
to
get
a
new
a
new
codec
in
that
uses,
bind
node.
That's,
I
think
what
has
emerged
as
one
of
the
goals
so
using
buy
node
to
do
a
simplified
form
of
a
codec
and
demonstrate
it,
because
we've
got
other
people
doing
codecs
and
doing
the
whole
schema
code.
Gen
thing:
it's
really
verbose
and
would
be
nice
to
have
alternative
way.
C
People
to
do
this,
and
we
also
have
this
question
around:
are
we
maintaining
which
one
are
we
maintaining,
are
maintaining
both
the
code,
gen
or
and
and
and
by
node,
or
only
one
of
them
and
yeah.
A
D
Thanks
yeah,
so
I
just
recently
pushed
my
prototype
for
ipld
update
and
it's
kind
of
based
on
the
code
that
eric
wrote.
I
did
I've
changed.
It
is
a
similar
approach
with
some
tweaks
and
basically,
what
it
does
is
that
it
it
accumulates
updates
as
they're
applied.
So
it
lets
you
accumulate
as
many
updates
as
you
want
before.
Any
node
objects
are
constructed
and
the
other
thing
it
does
is
that
it
provides
a
lens
lens
to
encode,
which
will
just
see
a
node
object.
D
A
Thanks
can.
A
E
Excellent,
so
actually,
following
up
on
the
ipld
operations
stuff,
I'm
actually
finishing
up
the
ipld
patch
spec,
so
eric
or
warpwork
started
putting
together
some
tests
and
an
initial
implementation
in
go
ipld,
prime,
so
I'm
just
going
to
be
finishing
up
whatever
needs
to
be
done
to
get
it
merged,
as
my
next
steps,
so
hopefully
that'll
all
just
be
upstream
in
a
few
weeks
other
than
that
some
stuff
I
was
working
on,
got
merged,
so
we
now
have
a
new.
E
What
is
ipld
image
and
some
new
descriptions
on
the
website
and
the
ipld
slash
reviewers
team
is
up
and
running.
So
if
you're
on
the
go
ipld
prime
repo
or
the
ipld
ipld
repo,
you
can
at
ipld
reviewers
to
tag
a
bunch
of
people
that
can
review
your
code.
So
this
is
probably
going
to
be
more
useful
when
we
have
more
capacity
or
more
people
just
contributing
to
the
repos,
but
for
right
right
now,
that's
there
to
start
playing
around
with
for
people
actually
working
on
these
repositories.
E
Also,
I'm
planning
on
merging
the
exploration
report.
I
did
initially
for
ipold
urls.
This
isn't
like
a
formalized
spec
yet,
and
I
think
a
dean
had
a
lot
of
useful
comments
for
stuff
that
would
be
useful
to
actually
differentiate
it
from
existing
features
in
the
gateway.
E
So
I
think
it'll
be
useful
to
like
strategize
more
about
like
how
we
can
actually
make
a
spec
and
see
if
we
can
get
buy-in
from
other
parts
of
the
ecosystem
to
actually
use
the
spec.
E
This
initial
one
is
like
bare
bones:
just
like
getting
and
putting
stuff
from
ipld
urls
kind
of
like
using
ipfs
or
ipns
urls,
so
we'll
kind
of
see
how
that
goes.
E
Finally,
I've
been
working
on
putting
together
a
dev
grant
for
formalizing
chunky
trees
into
a
spec.
So
I
think
I've
mentioned
it
a
few
times,
but
we're
starting
to
have
calls
now
and
we're
going
to
be
doing
a
call
with
some
of
the
people
that
originally
made
this
spec.
E
And
so
the
idea
is,
we
want
to
make
it
easier
to
create
indexes
and
stuff
using
ipld
data
structures.
So
if
you
have
a
large
data,
set
it'll
be
easier
to
index
it
and
then
search
through
large
chunks
of
data
without
having
to
figure
out
how
to
do
that,
every
single
time
you
want
to
do
an
application
so
right
now
we're
mostly
figuring
out
what
exactly
we're
going
to
do
and
we've
got
one
person
that's
interested
in
committing
on
the
research
side
of
things,
but
it
would
be
useful
if
there's
any
go
developers
out
there.
E
That
would
want
to
jump
on
and
kind
of,
like
you
know,
sink
their
teeth
into
some
ideal
development.
So
this
is
open
for
pretty
much
anyone
in
the
community
to
join
in.
So
if
you're
interested
hit
me
up
and
move
on
matrix
or
on
discord
in
probably
whatever
channels
you're
already
in
cool,
that's
all
I
got.
A
Thank
you
any
more
updates
from
anyone.
A
If
not,
then
I
have
an
agenda,
a
quick
one.
So,
yes,
those
not
familiar.
We
have
chat
rooms,
several
ones
so
on
several
platforms
and
the
question
came
up
is
so
so
I
guess
the
most
used.
One
is
probably
on
matrix,
the
ipld
channel,
which
is
bridge
to
discard,
and
in
this
chord
it's
called
ipld
chatter,
and
then
it's
also
a
thinx.
A
I
don't
know
it's
curly
bridge,
but
it
will
be
rich
or
it's
bridged
also
to
fly
coins
like
somewhere.
I
don't
know
where
and
we
have
another
channel
called
ipld
death
and
like
originally.
A
I
guess
because
like
there's
also
not
that
much
traffic
currently-
and
I
don't
see
really
the
point
of
having
like
a
semi-internal
one
for
developers
because,
like
yeah,
so
in
my
opinion
we
should
have
a
single
channel
which
is
bridge
to
discord
matrix
and
slack
and
that's
it
and
then,
if
it
gets
too
noisy,
we
can
branch
off
but
yeah.
So
I
just
want
to
hear
from
others
like
what
they
think.
If
we
should
have
two
channels
or
one
channel
or
yeah.
F
Cool
thanks
for
raising
vulgar,
so
yeah
totally
agreed.
Whatever
happens,
we'll
make
sure
it
gets
bridged
to
all
the
places
so
that
that
will
definitely
happen.
That
may
not
be
the
case
yet,
but
that
will
happen
and
what
we
had
happened
a
few
months
ago
with
ipld
in
terms
of
their
at
least
in
ipfs,
discord
there
being
an
ipld
chatter
and
an
ipld
dev
that
was
kind
of
following
the
convention
of
ipfs,
where
at
least
it's
very
useful
to
separate
out.
You
know
why
isn't
my
node
working
and
all
that
kind
of
chatter?
F
Why
doesn't
this
install
etc
from
like
those
who
are
actually
doing
ipfs
implementations,
but
I
can
totally
see
how,
like
maybe
that
doesn't
make
sense
in
the
ipld
world,
but
that
was
kind
of
you
know
wanting
to
make
sure
there
was
a
high
signal
place
where
those
who
are
working
on
implementations
are
not
distracted
by
like
support
requests.
If
ipld
doesn't
need
that
like
great,
I
think
I'm
all
for
combining
it,
but
that
that
was
the
that
was
the
mental
model,
but
so
anyway,
I'm
not
really
adding
anything.
B
One
sounds
pretty
good
to
me.
I
think
I
think
two
was
had
some
utility
when
we
nobody
really
wanted
to
use
threads
in
the
matrix
channel,
because
it
would
break
everything,
but
now
that
you
can
use
threads
at
least
from
the
matrix
side,
and
they
still
makes
its
way
everywhere.
I
think
it's
probably
fine.
I
think
the
one
time
I
you
know
accidentally
the
the
discord
only
channel
and
then
for
like
let's
go
move
it
to
the
other
channel,
and
then
I
was
like
okay
great.
B
E
For
what
it's
worth,
I
think
both
channels
are
fully
bridged
up.
I
think
when
I
first
started
working
on
ipld
stuff,
one
of
them
wasn't
and
then
I
like
repeatedly
prodded,
discordant
and
bridges
got
set
up,
but
yeah.
I
could
see
it
being
confusing.
Like
myself,
I
wasn't
aware
which
one
was
really
for
which
it's
just
like
my
pld
channels,
I'll
just
post
in
one,
so
I'm
into
it.
E
It
would
be
nice
not
to
lose
the
history,
though,
for
searching.
I
don't
know
if
deleting
channels
will
destroy
the
history
or
if
we
can
mark
them
as
archived
or
read
only
or
something.
A
F
Yeah,
I
mean
definitely
dealing
with
making
sure
andres
rehomes
into
file
coin
slack,
and
so
obviously
part
of
that
commitment
was
that
we
have
public
bridging
out
into
matrix
and
also
to
discord.
I
think
ipld
is
the
only
case
where
we're
going
to
like
collapse,
a
channel
but
yeah.
I'm
fine
for
that
to
be
for
to
take
that
or
to
find
an
owner
for
it.
A
Yeah
thanks
all
right
yeah,
I
guess.
Then
it's
decided
that
we
could
lapse
it
to
a
single
channel
cool.
Alright,
anything
else.
A
No
cool,
then
this
week
it
was
a
quick
meeting
and
as
every
week.
So
this
is
the
end
of
the
live
stream.
But
if
you
want
to
hang
out
it's
now
the
time
for
also
like
off-topic
things
or
secret
things,
you
can't
share
publicly
so
yeah
so
feel
free
to
hang
out.
But
I
say
goodbye.