►
From YouTube: 🖧 IPLD Every-two-weeks Sync 🙌🏽 2022-06-06
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
And
I
think
we're
streaming
now
on
youtube,
so
we're
starting
our
meeting.
Welcome
everyone
to
our
two
weekly
ipld
sync.
We
have
a
smaller
group
today.
Welcome
is
not
with
us
today
due
to
public
holiday
and
yeah.
We're
just
gonna.
Do
a
quick
run
through
of
any
sync
items
that
anyone
who's
on
the
call
has
if
anyone
is
watching
on
youtube,
you're
welcome
to
join
the
zoom
and
tell
us
about
what
you're
working
on
otherwise
feel
free
to
listen
in
and
yeah.
A
So
my
updates
first
I'll
do
my
lists
on
first
on
the
notes.
A
So
yesterday
I
was
working
on
a
reviewing
and
testing
out
a
pull
request
that
goes
on
has
had
against
the
js
model
formats
repo
for
a
while
now
and
it's
just
it
hasn't
been
done
because
it's
initially
it
was
a
breaking
change
to
the
cid
class.
But
we
talked
him
off
the
ledge
and
we
turned
it
into
a
backward,
compatible
change
which
is
really
interesting.
So
the
the
main
work
was
looking
at.
Is
it
actually
backward
compatible?
A
A
You
can
use
typescript
or
typescript
annotations
to
to
specify
the
interface
wherever
you
expect
a
cid
to
be
given
to
you,
rather
than
the
concrete
class,
and
that
sort
of
brings
us
closer
towards
detaching
from
all
these
concrete
implementations
we
have
everywhere.
Cid
is
the
the
main
concrete
implementation
we
have
that
is
used
by
everything,
and
so,
if
you,
if
we
turn
that
into
an
interface,
then
we
get
to
more
loosely
couple
things.
So
that's
in
the
js
module
format,
three
pokeball
request
number
161..
A
I
one
of
the
ways
I
tried
it
out
was
using
it
in
the
was
updating
the
the
bitcoin
codec
to
it,
which
is
makes
a
lot
of
use
of
cids
and
doesn't
really
yeah.
It
makes
it
it
has
to
parse
and
pass
around
and
decode
cids
and
do
some
other
things
with
them.
So
pull
request.
Number
seven
on
the
js
bitcoin
repo
in
the
ipl.org
uses
the
interface
version
of
it
where
it
can
and
the
interesting
thing
about
about
upgrading.
A
So
it
means
that
it's
not
breaking,
but
you
can
opt
for
using
an
interface
by
changing
your
imports
a
bit.
So
it
seems
to
work
so
a
little
bit
more
work
on
that
today
and
then
that
will
likely
get
in
this
week.
I
think
yeah.
So
the
next
thing
I
did
was
merge.
The
thing
that
I
had
up
for
a
little
while,
which
is
a
proposal
to
make
the
go
ipld
primes
codex
error.
If
you
try
and
encode
a
graph
that
has
a
link
node,
that's
got
an
undefined
cid.
A
So
this
is
the
case
where
you
have
where
the
cid
types
zero
value
is,
is
called
undef.
It's
got
no
bytes
in
it
and
go
out
with
the
prime
doesn't
currently
check
for
that.
The
problem
is
that
it'll
encode
invalid,
mainly
invalid
c-ball,
when
you
do
diagonal
and
when
you
would
do
it
with
jason
it'll,
it
won't
do
a
round
trip,
so
it'll
actually
technically
be
invalid,
jason
too.
So,
instead
of
coercing
that
was
an
option.
A
We
could
coerce
it
to
null
that
sort
of
changes,
the
kind
which
we
probably
don't
want,
and
especially
if
your
field
is
not
nullable,
we're
opting
for
erroring,
so
it
just
puts
the
owners
on
the
data
author
to
make
sure
that
their
cids
aren't
undefined,
which
is
probably
good
practice
anyway.
A
I'm
not
aware
of
any
places
where
this
is
a
problem.
I
I
think
those
who
have
stumbled
across
this
have
done
through
error
through
programming
error.
That's
pull
request.
Number
four.
Three
three
on
go
up
with
e
prime
another
one.
Four
three
two
on
go
up:
ld
prime,
was
just
it
was
improvement
to
bind
node's
schema
inference,
so
it
can
infer
schemas
from
your
go
types
a
little
bit
more
flexibly.
A
Now
it
can
detect
a
few
more
kinds
and
then
just
that's
432
and
then
a
bunch
of
other
minor
cleanups
and
fixes
and
reviews
that
I'm
not
going
to
list
but
been
trying
to
keep
an
eye
on
things
and
make
sure
that
things
get
reviewed.
And
lastly,
I
want
to
drop
a
note
that
there
is
a
there's,
an
interesting
proposal
from
geropo
over
on
the
multi
multi
codec
repo.
So
in
the
multicode
table
portal
across
number
272
over
in
the
multi
coding.
A
Repo
has
a
proposal
for
what
he's
calling
an
m
set,
which
is
essentially
just
an
identity
multi-hash,
but
with
some
compression,
which
sounds
a
bit
weird.
But
the
use
case
is
interesting
in
there
and
it
looks
like
it's
been
some
discussion
since
I,
since
I
looked
at
it
yesterday.
So
that
might
be
interesting
for
some
folks
to
look
at
and
possibly
engage
in,
and
that's
it
for
me
and
then
it's
over
to
mo.
B
Yeah
hi
I've
been
kind
of
preoccupied
with
some
other
stuff
lately,
but
thanks
to
rod's
help,
we
got
the
jstorm
patch
spec
that
bore
pork
was
working
on
in
a
place
where
we
can
merge
the
initial
version.
So
I'm
gonna
merge
that
today,
if
no
one
has
any
other
comments
or
concerns,
so
it's
still
kind
of
experimental.
We
haven't
test
stuff
like
traversing
links
when
doing
patches,
but
it's
kind
of
working.
We
have
some
tests.
B
People
are
welcome
to
start
playing
with
it,
we've
also
merged
the
json
or
sorry.
The
ipld
url
schemes
exploration
report
that
I
made
a
while
ago.
So
at
the
moment,
I'm
not
sure
where
the
next
steps
are.
There
might
be
cool
to
talk
about
that
if
there
should
be
net
next
steps
and
then
there's
just
some
small
docs
things
that
I've
been
helping
out
with
as
well
working
on
ipld
checklist
or
for
implementations,
which
there's
a
lot
to
it.
B
That
needs
to
be
rewritten,
but
now
patch
is
going
to
be
in
there.
That's
about
it.
B
Oh
also
been
slowly
progressing
on
putting
together
a
grant
proposal
with
the
can
ken
labs
folks
that
I
think
showed
up
last
week
to
get
ipld
prolly
trees
standardized,
so
we're
kind
of
figuring
out
what
we're
even
gonna
do
with
that.
And
then
I'm
gonna
have
a
proposal
together
and
I'm
probably
gonna
talk
to
dietrich
from
whatever
team
he's
on
in
protocol
labs
to
push
that
through.
B
But
there's
a
lot
of
potential.
For
one
thing
we
were
thinking
of,
including
is
having
an
ipld
indexed
database
that
people
could
build
off
of,
as
just
like
one
of
the
examples
that
we
build
on
top
of
prolly
trees
and
we're
just
kind
of
figuring
out
whether
that's
going
to
be
part
of
the
grant
or
whether
they're
going
to
do
that
as
their
own
thing.
A
Cool
then
immersion
gonna
update
for
us.
C
Yeah
so
I
cleaned
up
and
and
pushed
the
the
map
type
amanda
implementation.
All
that
map
tests
are
now
passing
the
the
list.
C
Amendment
was
a
little
more
complicated
and
I'm
almost
done
with
that
as
well.
Have
one
more
test
passing
or
left
to
pass.
Once
that's
done,
I'm
gonna
clean
it
up.
C
Some
more
add
more
comments,
and
then
I'm
thinking
of
adding
more
fixtures
to
to
cover
more
corner
cases
will
help
test
both
implementations
and
then
might
have
to
work
with
you
below
or
rod
to
figure
out
a
way
to
do
the
benchmarking
and
and
maybe
y'all
can
recommend
what
would
be
the
next
step
once
this
implementation
is
complete,
so
yeah,
that's
where
I
am.
A
B
B
B
So
originally
we
were
even
thinking
of.
I
was
also
talking
to
the
stewards
team
to
lydle
about
maybe
getting
some
of
these
ipld
things
into
the
ipfs
gateway
code
base,
because
right
now
they
have
like
some
ipld
stuff,
but
it's
mostly
for
like
downloading
blocks
from
ipld-ish
urls.
But
it's
it's
very
rough
and
it
feels
very
hacky.
B
So
part
of
the
goal
here
was
to
make
it
kind
of
more
of
like
here's,
ipld
here's,
the
data
structures,
here's
how
you
can
reference
them
to
get
data
out,
but
also
how
we
can
have
a
simple
api
for
putting
data
in,
and
so
that
would
work
with
gateway,
enabled
things
say:
ipfs,
companion
down
the
line
or
aggregator
or
whatever
other
peer-to-peer
web
browsers
out
there.
There's
not
a
lot.
Well,
actually,
there's
a
bunch
now,
there's
capilou!
B
It's
based
on
firefox
os
we're
a
great
grandchild
of
firefox
os
that
has
ipfs
integrated,
so
it's
just
like
files
are
cool,
but
ipld
is
also
cool
and
it'd
be
nice
if
it
got
like
recognition
as
its
own
thing.
B
Sorry,
that's
about
it.
A
I
think
lidl
is
definitely
the
one
that's
been
pushing
us
ahead.
The
most
he's
he's
been
working
on
things
like
car
exports
and
using
content
types
to
signal
some
of
these
things,
because
it's
it
really
is
a
retrofitting
exercise.
We
have
the,
I
think
we
have
the
slash
ipld
namespace.
I
think
that
might
be
in
the
gateway.
A
I
know
it
is
on
some
of
the
the
ipf
the
go
ipfs
commands,
but
I
don't
think
it's
useful
on
the
gateway
and
there's
been
plenty
of
threads
about
how
do
we?
How
do
we
make
this
thing
useful
and
use
content
types
and
stuff,
and
I
don't
think
anything,
anything's
matured
like
your
exploration
report
has
so
I
think
you
know.
Lydell
is
certainly
the
one
to
push
it
forward,
there's
particularly
for
the
gateway.
A
Where
else
would
it
was
there
any
anywhere
else?
We
could
apply
it,
or
is
it
really
just
a.
B
B
Well,
I
was
thinking
it
could
be
useful,
even
in
go
ipld,
prime,
as
a
way
to
work
with
data,
because
like
right
now
go
ipld.
Prime,
is
really
great
for
like
working
with
data
structures
and,
like
you
know,
we
have
all
of
these
nice
type
systems,
but
I
think
there's
opportunity
for
having
just
like
a
really
really
high
level
api,
which
is
just
like
urls,
and
it's
like.
B
A
Yeah,
it
could
fit
nicely
in
the
traversal
package.
I
think
patch
has
gone
in
there
and
it
is.
It
is
a
form
of
traversal,
so
yeah
it
makes
sense.
The
update
side
is
less
so,
but
I
think
we're
putting
those
you
know
patch
and
update
things
in
there
as
well.
So
it
does
make
sense,
and
I
wonder
whether.
A
Trying
to
think
through
whether
the
gateway
touches
go
up
really,
prime
for
these
things,
because
because
we've
got
this
in
between
state
in
the
go
ibfs
code
base,
where
we've
we've
migrated,
some
of
it
to
google
brian,
but
some
of
it's
still,
some
of
the
more
raw
block
handling
stuff
is
still
using
some
of
the
older
tools
tools,
but
maybe
maybe
adding
this
kind
of
functionality
would
be
a
way
of
getting
some
of
those
things
in
to
the
code
bases.
B
Yeah
like
if,
if
lytle
is
interested
or
if
there's
like
buy-in
from
the
gateway
team,
I
don't
know
who's
on
it
right
now
and
having
like
a
slash,
ipld
path
that
we
can
like
document
and
make
official
that
could
be
cool
like
there
is
also
the
writable
gateway
spec
in
the
gateway
code
base,
which
we
could
work
off
of
to
make
writable
ipld
down
the
line,
because
I
personally
I
find
the
developer.
B
Ux
can
be
really
nice
when
it's
just
like
links
and
you're,
either
you're
doing
just
crud
operations
on
them,
because
it's
like
a
lot
less
to
learn
and
a
lot
less
to
kind
of
like
have
to
figure
out
compared
to
like.
Oh,
I
have
like
all
of
these
different
apis
and
like
a
billion
ways
of
doing
things.
I
guess
this
is
a
a
billion
and
once
a
way
of
doing
things
like
that
xkcd
comic,
but
it's
a
billion
and
one
way
that's
like
unifies
a
bunch
of
different
ways
of
doing
it.
A
B
B
Obviously
I'm
super
biased,
but
I
think
urls
are
just
like
hey
here's,
the
thing
that
exists
and
a
lot
of
devs
know
about,
and
you
just
like
learn
that
and
like
there's
potential
for
integrating
with
other
tools
like
I
know,
there's
that
ffmpeg
integration
with
ipfs.
B
B
Or
even
as
a
meta
thing
like,
I
was
talking
to
eric
about
this
a
while
ago,
and
one
of
the
things
we
were
thinking
is
like:
oh
could
we
like
specify
lenses
as
like,
cids
or
urls
to
something
like?
Could
I
link
to
a
schema
from
within
a
url?
B
D
Well,
yeah,
not
maybe
some
of
this
is
all
just
a
few
data
things,
so
so
dietrich
and
u.n
and
lydle
are
have
been
well
so
so
mark
is
who's
been
kind
of
taking
on
integrating
ipfs
into
different
places
like
curl,
ffmpeg,
etc,
and
I
think
that
grant
and
relationship
is
managed
by
dietrich.
D
But
you
know
lytle
is
heavily
involved,
so
I
know
obviously
you're
talking
with
both
of
those
guys
anyway,
but
you
can
you
know
if
you
have
other
ideas
there
on
how
to
broaden
like
they're
like
they
would
be
the
ones
to
talk
to
for
that
particular
zone.
As
I
put
in
chat,
you
know
we
did
get
like
official
gateway
specs
actually
out
for
pr.
I
don't
know
if
you've
seen
that
yet
or
not
mav,
but
I
I'm
you
know.
D
Lyle
just
got
that
out
and
I
know
he's
gonna
be
like
coming
around
and
pinging
different
people
to
give
feedback
on
it.
So
that's
just
one
thing
in
general
about
gateway
specs,
then
I
you
know
there
is
a
ipfs
we're
calling
it
ipfs
implementers
all
thing.
It's
very
focused
on
ipfs
implementers
in
july.
I
don't
know
if
an
invite
has
been
sent
out
to
you
yet
or
not
on
that
mob,
but
it's
a
it's
more
of
a
private
event.
It's
not
this,
isn't
the
ipfs
like
big
public.
Sorry!
D
This
is
an
ipfs
camp
which
was
intended
to
be
like
a
bigger
production
that
we
were
you
know
which
is
going
to
get
postponed
to
later
on
in
the
year.
This
is
more
of
a
like
a
very
targeted
focused
on
implementers
kind
of
like
the
workshop
that
you
participated
in
a
couple
weeks
ago,
but
we're
doing
a
gathering
in
iceland
in
july.
For
that-
and
I
know
for
that
gathering
that
lydle
in
particular
wants
to
show
like
we
do,
have
a
specs
process
and
a
few
cases
have
run
through
it.
D
You
know
next
thing
he's
planning
to
add.
Besides
just
these
basic
gateway,
specs
is
like
how
we
add
selectors
into
the
gateway
spec,
and
so
it
wants
to
show
a
couple
of
use
cases
where
it's
actually
run
through.
So
I
say
all
of
that,
and
maybe
it's
worth
you
syncing
up
with
him
in
case
you
want
to
put
out
a
spec
for
more
ipld
things,
because
that
would
kind
of
fit
in
his
in
the
narrative
that
he
wants
to
show
of
like
hey.
D
A
B
A
Well,
thanks
everyone
for
tuning
in
get
these
notes
onto
the
ipld
team
management
repo
afterwards
and
anyone
else
that's
watching
is
welcome
to
join
and
share
your
ipld
updates
or
join
in
with
questions.
Usually,
we
pull
questions
from
youtube
comments,
but
I
don't
have
the
access
to
that
right
here,
unfortunately,
but
yeah
anyway.
So
thanks
everyone
for
joining
and
goodbye.