►
From YouTube: 🖧 IPLD Every-four-weeks Sync 🙌🏽 2023-09-11
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
Now
welcome
everyone
to
the
ipld
community
and
syn
meeting
it's
great
to
see
that
the
streaming
is
working.
Finally
again
so
yeah,
it's
November
11th.
B
A
And
yeah,
as
every
four
weeks
now
we
go
over
stuff
that
have
people
have
worked
on,
but
also
discuss
things
or
if
people
are
excited
about
yeah
project
projects,
they've
bu
with
ipld
feel
free
to
join
and
tell
us
about
it
and
I've
even
I,
even
see
that
we
are
now
more
people.
That's
great
hello,
that's
great
yeah,
so
with
updates
I
will
I
will
start
with
my
updates.
I
have
a
bit
news
from
the
rust
ipld
ecosystem
and
IP
and
multiformat
ecosystem.
A
So
there
has
been
a
multi
has
buck
fix,
release
which
is
kind
of
like
an
interesting
one.
Therefore
I
mention
it
it's
about
no
standard,
so
you
can
basically
in
Russ.
You
can
build
with
without
the
standard
library-
and
this
was
broken
wither
and
yeah.
The
fix
needed
some
unsafe
code,
which
is
unfortunate,
but
it's
hecking
around
con
generics
so
yeah,
it
ended
up
with
a
nice
hack
and
yeah
I.
Think
I'm
happy
with
the
result.
A
A
But
I've
also
made
a
PR
with
the
serializing
code
path,
because
one
thing
which
is
really
weird
about
cabore
is
that
the
negative
integers
are
the
full
46
bit
range,
but
with
a
negative
sign
and
the
fun
part
is
it's
even
bigger
than
the
maximum
positive
unsign,
64-bit
integer
and
because
the
maximum
maximum
size,
64-bit
integer,
is
two
to
the
power
of
64
minus
one
and
the
negative
one
even
can
go
to
the
not
minus
one,
but
to
the
H
one,
which
is
yeah
from
the
implementation
perspective
insane
and
to
me
it
seems
like
yeah
seore
is
pretty
theoretic
and
not
really
built
by
people
doing
practical
things,
but
anyway
in
R.
A
We
soon
support
this
yeah
again,
I
F
back
to
this
cover
or
find
out
about
and
yeah
work
around
yeah.
So
that's
pretty
much
it
for
the
items.
I
have
so
I
see
that
people
are
still
typing,
but
does
anyone
want
to
go
next?
I'll.
C
Have
a
quick
go:
it's
it's
another
one
of
those
months
of
lots
of
little
things
hard
to
pull
them
together
and
describe
all
all
of
the
all
of
them,
but
I'll
try
and
pick
off
the
memorable
bits.
So
there's
been
various
in
the
JavaScript
ecosystem.
There's
been
releases
of
various
M
sort
of
secondary
components
like
sh,
3
and
Blake
2
that
have
been
not
brought
up
to
speed
with
the
rest
of
the
stack
for
because
we
did
a
the
multi
JS
multiformat
release
a
while
back.
C
That
was
esm
only
and
we
didn't
we
didn't
haven't
brought
the
others
up
to
speed
with
that
and
there's
been
some
breakage.
So
it's
been
difficult
to
use
some
of
these
extra
components
together,
so
those
have
been
updated
and
released.
There's
another
one
coming
down
the
pipe,
the
core
cabore
and
JS
paing
engine,
that's,
but
that
that's
been
able
to
sit
the
background
without
without
breaking
anything.
C
But
there's
been
a
a
little
bit
of
work
going
on
there
and
a
new
release
of
that
and
that'll
bubble
through.
The
stack
probably
won't
be
noticeable
for
people,
though,
there's
been
various
updates
and
fixes
to
the
JavaScript
Unix
FS.
We
found
we
found
a
bit
of
functionality,
it
wasn't
able
to
handle
and
manag
to
fix
that
and
then
there's
two
other
teams
that
are
working
on
some
traversal
and
verification
Utilities
in
JavaScript
through
Unix
FS
for
the
trustless
Gateway
spec.
C
So
that's
really
interesting
work.
The
web,
three
storage
people
with
their
trust,
Gateway
that
goes
through
Cloud
flare,
posted
through
cloud
cloud,
flare,
they've,
got
a
traversal
thing
and
then
there's
the
satin
team,
which
is
doing
a
in
browser
verifier
that
can
take
a
a
car
and
then
verify
it
walking
the
same
traversal
in
JavaScript.
Those
two
efforts
haven't
quite
been
unified,
but
they
hopefully
don't
need
to
walk
too
much
of
the
same
paths,
but.
B
C
That's
lots
of
interesting
stuff
in
JS
at
the
moment
in
go
the
main
items
I
can
think
of
this
goar
has
has
had
a
few
new
releases
with
some
some
minor
features
and
some
fixes
there
was
a
there
was
a
release
of
go
020
that.
C
0121,
no
I
think
it
might
020
that
that
tried
to
fix
something,
but
then
we
realized
that
it
actually
changed
the
functionality
of
something
and
so
that
ended
up
getting
fixed.
The
latest
is,
is
sorry,
went
up
to
two
so
2120
was
this
version,
and
231
is
the
latest
one
that
we
would
recommend
people
be
on.
It's
a
minor
feature
for
skipping
over
cars
to
sort
of
index
them.
C
If
you
wanted
to
do
them
outside
I
I
I'm,
not
aware
of
any
other
uses
other
than
boost
uses
it
directly
and
we
broke
boost
in
a
branch,
but
that's
about
it
so
goar
is,
has
been
updated.
There's
a
new
package
called
go
trustless
utils
in
the
ipld
org.
That's
got
a
bunch
of
trustless
Gateway,
passing
in
traversal
utilities
and
various
types
for
describing
trustless
Gateway
my
requests
and
responses.
C
That's
an
extraction
of
code
that
was
being
used,
reused
across
a
couple
of
projects
and
I
if
I
had
if
I
had
have
had
time
and
if
I
had
have
remember
this
meeting
was
coming
up.
I
would
have
presented
a
presentation
on
some
of
this,
because
one
of
the
interesting
things
in
here
is
the
I
think
for
this.
This
meeting
is
that
I
think
trust.
C
This
Gateway
has
defined
a
a
pathing
spec
for
dags,
that
is
a
a
decent,
a
minor
extension
to
standard
pathing,
but
a
decent
subset
of
selectors
that
or
stive
functionality,
that
is,
that
is
fairly
use
useful
and
if
youed
in
that
way,
then
I
think
this
is
a
the
trustless
Gateway
stuff
is,
is
a
good
development
for
describing
dags
in
what
common
ways
that
people
want
to
describe
them,
and
so
one
of
the
things
that
go
trustless
utils
does
is
extracts
out
some
of
the
traversals
that
we
do
so
if
you,
instead
of
doing
a
traversal
with
Co
Prime
directly,
you
can
do
it
through
this
utility
and
it
it
essentially.
C
Simp
simplifies
the
traversal
API
and
that's
one.
C
One
of
the
things
I've
been
interested
in
working
on
is
better,
better
interfaces
to
traversals,
perhaps
skipping
around
the
complexities
of
selectors,
because
I
think
it's
one
of
the
sort
of
the
the
the
the
gnarly
edges
in
go
in
go
build
Prime,
particularly
of
traversal,
is
this
selectors
and
the
messiness
and
the
complexity,
and
so
often
you
come
to
IP
do
traversals,
and
you
want
to
do
something
really
simple
and
you're
made
to
jump
through
all
these
hoops
to
do
something
simple,
and
it's
very
rare
that
you
want
to
do
the
complicated
stuff
that
all
the
complexity
is
there
for
so
providing
simplified
apis
I
think
is,
is
a
a
good
thing
to
do,
and
it's
sort
of
one
of
the
things
that's
come
out
of
this
and
it'
be
interesting
to
to
see.
C
If
we
can
extend
that
work
and
make
it
more
useful
outside
of
just
this
trustless
Gateway
work,
can
we
can
we
provide
something
that
is
more
generically
useful
for
he's
a
dag
he's
a
way
I
want
to.
You
know,
select
part
of
it
so.
B
C
There
some
thoughts
around
that,
but
that's
the
main,
IP
specific
stuff
that
I
can
I
can
list
there
there's
a
lot
of
other
little
things,
but
probably
not
worth
listing.
A
Yeah
thanks
I
see
that
Adina
also
has
some
updates.
So
please
go
ahead.
C
B
I
I
I
agree
p
a
good
this
Shan.
This
kind
of
didn't
go
very
far,
but
I
feel
like
we're
missing
pathing
with
like
how
to
do
anything
other
than
Unix
FS
as
a
multi-block
data
transform,
but
I
feel
like
if
you
had
that
you
would
be
in
in
pretty
good
shape,
yeah
yeah,
so
I
guess,
I
I've
haven't
been
to
this
meeting
in
a
little
while.
But
since
then
I've
kicked
a
couple
of
hornets
M.
B
B
The
result
was
usable
for
things
like
I
just
want
to
basically
show
hash
and
a
block
Explorer,
but
not
particularly
good
at
transferring
data
around
and
so
as
a
result.
There
are
two
proposals
for
how
to
do
this
better.
Basically,
instead
of
treating
this
as
a
Merkel
tree,
which
it's
not
a
very
good
Merkel
tree,
you
you
just
treat
it
as
like.
B
A
big
block
of
data
like
you,
would
for
a
Blake
three
hash
or
something
instead
of
trying
to
represent
the
the
internal
tree
structure,
because
it's
it's
not
a
very
good
tree.
There
are
sort
of
two
proposals
for
this.
One
is
sort
of
the
the
bare
minimum.
That's
required
to
to
be
a
useful
CID
and
the
other
one
is
more
featureful.
It's
like.
B
Can
this
just
be
used
as
an
arbitrary
hash
function
that
lets
me
reference
arbitrary
data,
so
people
are
interested
in
that,
whether
from
like,
because
they're
interested
in
sort
of
the
the
file
coin
stuff
or
because
they're
like
this
sort
of
pushes
some
of
our
boundaries
around
like
Cs
and
multi
hases
and
whatever
like?
What
is
it
doing?
B
Why
is
it
doing
might
be,
might
be
interesting
to
take
a
look
I
guess,
maybe
the
two
most
obvious
ones
are
like
what
what
are
CS
for
which,
in
my
opinion,
is
like
for
for
verifying
the
data
right,
if
it's
just
like,
if
you
just
liked
multibase
and
you
could
have
put
anything
on
the
other
side
that
would
be
using
multibase
and
CS
are-
are
content
identifiers,
which
means
that
they
have
to
be
has
to
be
like
a
verifiable
mapping
between
the
bytes
and
the
the
c,
and
the
other
is
what's
reasonable.
B
To
put
in
a
multi
has
digest?
Is
it?
Is
it
anything
that
you
can
use
to
verify
that
a
stream
of
bytes
is
the
stream
of
bytes?
You
were
looking
for,
or
is
it
something
more
restrictive
than
that?
So
some
of
the
I
wish
there
was
more
engagement
on
GitHub
instead
of
through
other
channels,
but
the
you
know
one
of
the
more
feature,
the
more
featureful
options
that
sort
of
looks
like
a
regular
hash
function
relies
on
shoving.
More
data
in
the
multi
has
digest
right.
It's
like
is
this
a
good
idea.
B
Is
it
sad
that
once
we've
exceeded,
like
you
know,
once
we
get
a
little
bigger
than
a
256bit
hash,
That
Base
36
in
coding
won't
fit
it
into
a
subdomain
for
or
like
IPS
gateways
and
sub
domains
are
sort
of
a
useful
security
property
on
the
web
right.
So
those
those
sorts
of
things
kind
of
pop
up
there.
A
Share
thanks,
yeah
I
will
certainly
I,
think
I
haven't
to
the
to
the
more
featureful
version.
Yet
so
I
I
will
definitely
have
a
look
at
it.
So
because
yeah
I
mean
I'm
also,
certainly
interested
in
like
like
yeah
what
arti
hes
what
they
are
used
for
or
CAD
and
like
yeah,
so
I
mean
we.
We
also
like
have
discussed
this
like
whenever
we
be
in
person.
We
I'm
happy
to
discuss
it
or
we
did
in
the
past.
So
yeah
I
have
to
be
well
prepared
for
thatb
week.
A
C
So,
what's
the
path
for
these
two
FC's,
how
how
do
these
move
forward?
Is
it
just
discussion,
then
voting
or
what's
what
happens.
B
I
mean
they're
they're,
F
frc's,
so
they're
they're,
not
fips,
don't
seem
to
be
needing
votes,
I,
I
think
mostly
I.
Just
there
are
like
three
or
four
stakeholders
that
are
using
bad
CI
that
are
using
like
sort
of
unuseful
Cs
at
the
moment
like
they
are
not
useful
for
the
purpose
that
they're
using
them
for
and
they
either
they
need
to
change
how
they're,
using
their
apis
to
select
one
of
these
CS
or
do
something
else,
custom
that
fits
their
purpose.
B
B
First,
one
yeah
yeah
I,
didn't
realize
they
were
just
going
to
like
merge
them
in
once
they
were
PLL
requests.
I
was
sort
of
waiting
for
some
more
engagement
and
people
to
give
thumbs
up
and
they
were
like.
Oh
no,
no
we're
just
going
to
merge
the
draft,
and
then
people
can
discuss
or
make
changes
to
the
draft,
but
ultimately
I
feel
like
at
the
end
of
this.
There
should
really
only
be
one
F,
FRC
69.
B
There
shouldn't
be
like
two
different
ones
that
some
people
use
and
some
people
don't
use
because
either
way
you
need
to
merge
a
change
to
the
codec
table.
The
Code
table
to
say
this
multi
has
number
means
this
and
I
feel
like
adding
two
new
ones
for
the
same
filecoin
piece
thing
feels
kind
of
awkward
people
want
it
and
there's
demand
for
that.
Then
sure,
but
I
feel
like
right.
Now
we
just
need
to
decide
which
of
A
or
B
we're
going.
C
With
I
wonder
whether,
because
this
was
this
was
always
slightly
awkward,
a
slightly
awkward
fit
in
the
multic
multic
codic
table
from
the
beginning
and
I
wonder
if
clearing
some
of
this
stuff
up
actually
helps
us
a
bit
with
that
precedent,
because
we
did
we
we're
always
setting
precedence
in
the
multic
codic
table
when
we
compromise
or
we
we,
you
know
we
we
let
something
in
because
we
can't
say
no
or
whatever
and
a
lot
of
time.
That's
fine,
but
there
are
some
cases
where
it
it
pushes
us
towards
this.
C
This
this
space
of
meaninglessness
and
compi
was
was
a
little
bit
close
to
that
and
I
wonder
if
this
helps
us
get
out
of
that
bind
where
injecting
meaning
back
into
it.
It's
like.
B
Well,
yeah
I
mean
to
some
extent
right,
like
the
I
I,
think
it's
it's
actually
great
that
so
much
of
the
discussion
there
was
on
GitHub
that,
like
I
wasn't
involved
in,
but
I
could
go
see
some
of
what
was
going
on.
B
And
what
what
was
what
was
settled
on
was
like
was
was
just
very
awkward
right.
It
was
like
tagged
with
the
filecoin
serialization
thing
because
it
was
like
this
isn't
really
any
this
isn't
really
either.
But
people
seem
uncomfortable
with
the
large
block
thing
and
just
calling
it
one
big
block,
because
it
is
internally
a
Merkel
tree
and
it's
like,
but
it's
a
bad
tree
and
it
turns
out
it's
like
a
bad
Tree
in
practice
and
so,
and
so
we
should
just
use.
B
Right
and
so
the
fact
that,
like
those
people
are,
you
know,
those
people
are
are,
are
accepted
right
I
mean
there
are
trade-offs
to
using
large
blocks
in
places.
Please
don't
make
me
load
to
one
terabyte
Jon
object
into
memory
before
I,
try
and
decode
it
right,
or
something
like
that.
That
sounds
bad,
but
but
like
there
are
very
good
reasons
to
use
that
in
places.
B
I
also
think
it's
interesting.
It's
like
okay,
we're
reserving
yet
another
code
for
roughly
the
same
data,
but
we're
doing
it
because
there
was
some
user
experience
that
didn't
feel
good
about
the
last
one
and
to
some
extent
the
point
of
the
Code
table
is
to
like
be
able
to
evolve
past
mistakes,
which
means
we
gotta.
Let
people
make
mistakes.
B
Yeah
right,
like
people
did
things
they
made
mistakes.
We
try
and
move
on.
We
try
and
move.
We
try
and
limit
mistakes,
so
we
don't
do
things
like
maybe
encode
like
a
hundred
or
more
elements
for
for
the
various
Blake,
two
hashes.
Nobody
is
using
right.
So
we
try
not
to
do
those
again,
but
but
reserving
a
few
numbers
that
you
might
need,
you
might
have
problems
with
later
yeah
seem
seems
not
so.
A
Bad
yeah
and
as
we
already
talk
about
multi
and
multi
formats,
I
just
want
to
mention
which
I
forgot
to
mention
that.
Currently
there
is
a
push
in
the
ITF
to
form
a
multiformat
working
group
and
there's
email
discussion.
So
if
anyone
is
interested
so
I
think
it's
public
emails
and
especially
if
you
use
M,
you
can
chime
in
and
say
like
how
you
use
it
or
and
but
to
be
specific.
C
is
really
about
multi
and
multi
base.
A
So
we're
not
talking
about
CS
yet
to
keep
the
scope,
small
and
yeah
because,
as
you
can
see,
we
are
still
discussing
CAD
details.
So
it's
good
to
start
with
those
ones
first
yeah.
But
if
anyone
interested
Bas
search
for
m
in
ITF-
and
you
should
find
everything
you
want
to
know-
yeah
or
of
course
like
paying
us
if
B,
even
if
you're
interested
getting
even
more
deeper
into
it
and
yeah,
feel
free
to
ping.
Me.
B
A
Yeah
I
will
redirect
you
to
the
right
people
all
right.
Anything.
C
Say
about
the
multi
formats
and
multi-
based
stuff,
but
it's
it.
It's
just.
There's
been
a
lot
of
chatter
in
those
repots.
There's,
probably
some
discussions
that,
if
folks
are
really
interested,
they
can
go
and
check
out,
particularly
in
that
the
big
one
that
where
was
it
was
update,
update
the
docs
ready
for
ETF,
something
or
other
there's
a
big
pull
request
that
merged.
C
But
in
that
pull
request
there
was
a
lot
of
interesting
discussion,
just
sort
of
delving
into
history
and
why
things
are
the
way
they
are
and
what
do
we
do
about
base
two
and
all
that
sort
of
stuff
I
think
that's
a
valuable
discussion
to
be
kept
on
guub
being.
B
There
I
actually
kind
of
thought
it
was
I.
I
I,
like
the
some
of
the
the
push
back
and
the
multi-base
repo
someone
basically
saying
hi,
I
I,
don't
like
somebody's
base.
Can
we
remove
it?
People
are
like
it's
a
silly
base
but
like
why
remove
it
it's
a
valid
base.
That's
kind
of
you
know
that
that
that's
kind
of
how
this
goes
people
are
going
to
make
hashes
you
don't
like
you,
don't
have
to
include
them.
People
are
going
to
make
bases.
You
don't
like
you,
don't
have
to
include
them.
C
A
C
It
toys
with
your
the
edges
of
your
sensibilities,
about,
what's
going
on
here
and
and
not
I,
don't
think
in
in
ways
that
I
think
it's
been
useful
for
things
like
the
discussion
about
Unicode
Cod
point
at
the
at
the
beginning
that
actually
clarifying
that
as
a
a
thing
does
give
us
some
interesting
room
to
move
so
maybe
not
essential,
but
anyway.
A
Yeah
I'm
also
like
I,
also
what
like,
when
it
came
up
with
the
when
the
Emoji
based,
including
came
up
I,
also
wasn't
really
a
fan
of
it.
I
have
to
say,
but
now
looking
back,
I
think
yeah.
It's
a
good
idea
to
really
have
it,
as
as
a
good
example
or
like
yeah,
pushing
the
boundaries
and
then
yeah
explaining
to
people
why
it
is
there
or
yeah.
So
it's
kind
of
like
yeah,
an
interesting
piece
so
yeah,
but
I
also
would
say
like
if
people
Implement
multipase,
they
don't
necessarily
need
to
be
implement.
B
A
C
There's
a
multibase
on
there
that
that
none
of
us
Implement,
which
is
proquin
I've,
had
this
tab
open
my
computer
for
a
while
I've
been
staring
at
the
the
spec
for
this
thinking.
This
might
be
really
interesting
to
implement.
B
So
it's
actually
sort
of
implemented
in
go
it
just
sort
of
got
ripped
out
the
the
history
there
is
that,
for
like
a
really
long
time,
there
was
a
I,
don't
know
what
to
call
it.
Let's
call
it
an
Easter
egg,
because
you
could
basically
only
discover
its
existence
via
error
code
or
finding
a
random
medium
post.
Where
you
could
identify.
You
could
use
slns
some
proquin
to
identify
an
ipns,
key
or
I
guess
DNS
link,
but
there
was
no
docs
for
it.
B
All
that
would
happen
is
that
if
you
basically
typed
in
like
a
DNS
label
wrong,
it
would
just
say
error,
not
a
valid
proquin,
because
that
was
the
order.
It
was
checking
things,
and
that
was
the
only
reason
anyone
had
ever
heard
of
this
is
because
they
would
be
like
what
is
they
typed
in
their
thing
is,
like
you
know,
do
czz
and
they'd
see
like
not
a
valid
proquin
or
something,
and
then
they?
B
What
is
this
and
they'd
find
and
they'd
find
this
medium
post
telling
you
about
the
proquin,
so
we
were
like
this
is
I
mean
this
thing
is
a
Bas
en
coding.
Let's
just
delete
this
from
here
nobody's
using
it.
It
was
basically
an
Easter
egg.
If
somebody
would
like
it,
though
this
seems
like
a
reasonable
multi-base,
because
it's
just
a
binary
encoding,
there's
nothing
ipns
specific
about
it.
B
It's
just
roquin
was
already
in
was
already
was
already
a
binary
encoding
format
before
it
was
added
there
in
the
first
place
that
was
sort
of
the
the
compromise
like.
Let's
remove
it,
and
if
anyone
wants
it
here
is
where
you
would
go
want.
Here's
where
you
would
put
in
poll
request
to
go
re
it
nobody.
Nobody
has
yet.
A
Good
all
right,
cool
yeah,
then
what's
left
to
say,
is
that
so
the
next
meeting
will
be
four
weeks
again,
so
it
will
be
on
October,
9th
yeah
and
yeah,
also
like
what
we
usually
would
like
to
have
is
perhaps
some
presentations.
So
if
you
have
anything
to
present
or
any
cool
stuff,
you've
built
feel
free
to
reach
out,
also
or
also
like
on
the
IP
channels,
in
slack
Discord
or
Matrix,
and
then
yeah
we're
happy
to
yeah.
A
Have
you
speaking
at
this
meeting
all
right
see
you
all
in
four
weeks,
then
bye.