►
Description
Meeting description here:
https://github.com/ipfs/team-mgmt/issues/992#issuecomment-599632334
For more information on IPFS
- visit the project website: https://ipfs.io
- or follow IPFS on Twitter: https://twitter.com/IPFS
Sign up to get IPFS news, including releases, ecosystem updates, and community announcements in your inbox, each Tuesday: http://eepurl.com/gL2Pi5
A
Okay,
go
go
ahead.
I
saw
yours.
B
Okay,
we
are
now
recording
we're
happy
for
all
your
recording.
That
was
fun.
Welcome
to
the
corpus
weekly
sync
of
march
16th,
I'll
start
off
with
upcoming
shift
releases.
This
is
actually
my
update.
Yeah,
the
gopro
rc
is
a
bit
late.
We're
trying
to
get
out
the
release
by
the
end
of
this
month,
so
we'd
hope
to
have
the
release.
The
rc
out
by
friday.
Coordination
is
a
bit
tricky.
B
We've
been
having
like
synchronous
meetings,
constantly
yeah
secret
attack
time,
but
it's
only
semi
semi-synchronous
because
we're
all
in
different
time
zones.
So
we
get
a
good
chunk
of
the
day
together.
But
it's
not.
I
don't
know
it's
not
like
getting
the
templates
so
yeah.
B
The
coronavirus
is
a
bit
tricky
here,
but
remove
it
or
doing
your
best
with
a
couple
patches
practice
still
in
flight
you're
working
on
the
keystore
factor,
to
avoid
issues
with
like
file
system
or
like
system
specific
names
and
that
key
names
affecting
ipvs,
which
currently
we
store
them
unencoded.
We
just
drop
it
in
the
file
system,
but
somehow
don't
like
certain
types
of
names.
So
we
are
changing
these
two
base.
Three
into
base.
32
code
that
names
supplement
gateway
should
be
landing
soon
as
well
yeah.
B
B
That
the
address
filters
aren't
quite
working
as
well
as
we
had
hoped
and
they
are
kind
of
broken.
So
we
have
to
fix
that.
It's
a
bug
we
introduced
in
the
pcb
recently.
B
Yeah
we're
still
finishing
up
some
final
dht
patches.
We
should
have
an
arc
at
the
end
of
the
week
if
you
have
any
last
minute
change
soon
to
go
for
this,
please
make
them
now,
but
please
only
make
them
a
few
actions.
Let
me
go
access,
otherwise,
please
hold
off
until
well
april,
when
we
actually
have
finished
release.
B
We
do
not
have
any
old
desk
around
okay,
nothing
has
happened
on
describe
this
week.
So
much
has
happened
to
tesco
this
week,
but
we
don't
have
any
updates.
I
guess
unless
I'm
gonna
touch
something.
B
Okay,
thank
you.
So
content
writing.
D
Yeah
chris
talk
about
json,
give
us
for
a
second
yeah,
so
we
wanted
to
ship
the
pin,
speed
improvements
in
the
next
version,
but
it's
looking
like
it's
going
to
take
a
little
bit
longer
to
resolve
that
so
probably
going
to
try
to
ship
a
bug
fix
release
next
this
week
at
some
point
as
soon
as
basically
as
soon
as
we
finish,
merging
mfs
into
the
core,
because
we've
got
a
bunch
of
yeah
bugs
have
been
fixed
and
it's
all
a
bit
smaller,
so
be
really
great.
A
Okay,
hydra
the
we're
almost
done
with
stage
one
and
I've
tagged
a
0.1
release.
I
guess
for
hydra:
we've
got
thanks
to
peter
we've
got
we've
got
balanced
peer
id
generation,
so
we
don't
just
generate
rando
peer
ids.
We
actually
generate
a
bunch
of
your
ids
that
are
equally
kind
of
distributed
amongst
the
the
bit
space.
A
I
guess
you
could
call
it
so
thanks
for
that
we've
got
david
is
working
on
providing
for
things
that
we've
asked
for
that's
still
in
flight,
but
you
know
that
should
well
might
land,
hopefully
soon,
because
it's
not
it's
not
too
big,
but
it's
not
blocking.
So
that's
good.
We've
got
prometheus
metrics.
They
are
collecting
some
bits
and
pieces
that
we
want
to
collect.
So
so
that's
that's
awesome
and
we
do
like
yeah
there's
a.
A
I
did
a
little
bit
of
work
on
the
ui
so
that
actually
just
pulls
the
it's
all
this
information
from
prometheus
metrics
endpoint
that
we
expose,
which
is
super
rare.
So
we
don't
have
a
separate
kind
of
stats
collecting
process
for
that
which
is
nice.
So
it's
been
tagged
and
we're
trying
to
well
I'm
trying
to
deploy
it.
A
I'm
just
figuring
out
google
cloud
kubernetes
things
and
like
dockers
and
and
stuff,
and
just
trying
to
get
it,
get
a
version
deployed
with
a
few
heads
and
then
ramp
it
up
from
there.
Basically,
so
we're
having
fun
and
and
yeah
it's
going
well
so
far.
E
Yeah,
so
thank
you
for
yeah,
like
you
pretty
much
covered
it
all.
I
guess
just
like
to
highlight
next
steps
so
other
than
getting
it
deployed
and
started,
seeing
some
sets
getting
out
of
it,
which
hopefully
gets
done
today
tomorrow.
The
next
big
chunky
items
are
the
shared
routing
table,
and
so
right
now
there's
like
four
proposals
on
how
to
do
it
on
table.
It
will
require
some
design
thinking
like
there's
a
lot
of
input.
So
it's
not
that
we
are
looking
for
feedback.
E
We
just
need
to
think
more
about
it
and
then
the
other
thing
is
the
shared
data
store
across
multiple
machines.
We
are
thinking
of
using
just
like
a
like
a
standard
database
like
postgres.
I
believe
that
jeremy
implemented
a
backhand
in
postgres
a
long
time
ago.
Okay,
that's
a
thumbs
up.
Is
it
still
up
to
date?
Steven
can
I
can
we
trust
it.
B
E
A
day
today
just
to
get
it
tested,
I
also
require
more
like
tinkering
with
internals
and
so
yeah.
We,
we
haven't,
made
a
decision,
yet
we
need
to
think
more
about
it.
F
Just
that
most
of
the
dht
changes
are
looking
looking
good,
just
writing
more
tests
now
and
trying
to
scale
the
test
so
that
they
we
can
somewhat
plausibly
emulate
the
network
and
what
it
will
look
like
on
day,
one
when
we
roll
out
the
changes
to
make
sure
that
we
don't
break
anything
and
make
things
perform
at
least
as
well
as
they
are
now
even
for
people
on
the
old
network.
B
I
remember
you
talking
about
some
unit
test
issues.
Have
you
resolved
this?
Yet
lars
is
looking
into
those
now
excellent.
Thank
you.
B
Okay,
then,
I
think
that's
content
routing.
Hopefully
we
will
be
able
to
land
enough
of
the
changes
from
the
next
day
or
two
to
like
start
testing
against
the
network,
but
that's
coming
along.
B
E
Question
like
with
regards
to
testing,
I'm
very
curious
to
understand
like
what
what
is
the
success
rate
like
using
test
ground
to
test
the
dht
changes?
Are
we
like
being
able
to
like
run
multiple
network
topologies
like
different
network
sizes
and
like,
is
it
being
like
stable
right
now
or
is
well?
Can
you
can
you
hear
that
thing.
F
There
are
yeah,
I
mean
running
multiple
network
topologies
requires
defining
multiple
topologies
and
assuming
that
they
make
sense
so
sure
we
can
do
that,
but,
like
there
haven't
been
too
many
proposals
of
different
topologies,
so
we've
just
been
using
the
few
that
we
have
the
the
tests
are
supposed
to
be
able
to
scale
more.
I
haven't
taken
the
newest
test,
which
runs
old
dht
nodes
against
new
dht
nodes
in
various
different
ways.
I
haven't
scaled
that
up
to
you,
know
thousands
of
nodes
on
kubernetes,
yet
I'm
just
running
them.
C
F
And
making
sure
those
changes,
work
and
running
into
some
issues
with
the
synchronization
service.
But
hopefully
we
will
get
through
those
in
the
next.
You
know
day
or
so
and
see
what
happens
when
we
throw
it
on
kubernetes
and
if
it
falls
over
or
not.
G
One
one
thing
we
should
think
about
a
little
bit
is:
if
there's
a
way
to
disconnect
the
various
network
topologies,
we
can
think
about
from
the
specific
tests
so
that
we
can
reuse
those
in
other
types
of
tests.
Right,
like
both
like
I've,
got
some
nodes
that
are
old
and
some
nodes
that
are
new,
and
then
I
have
network
connectivity
that
looks
like
you
know
whatever
that
seems
like
we
will
want
to
reuse
it
for
each
new
feature.
F
E
E
E
Sounds
good-
and
I
guess
just
like
you,
like
so
for
the
gossip
testing,
what
we
are
thinking
if
there
is
a
container
limit,
we
just
like
to
spin
multiple
nodes
within
the
same
container
and
and
do
some
magic
to
tell
other
nodes
to
not
use
the
ip
as
an
identifier.
But
to
do
something
else.
You.
F
B
B
Okay
supplement
gateways.
H
It's
ongoing
radio,
I
think
most
of
stuff
is
addressed.
I
just
look
at
it
and
there
are
two
topics
that
I
need
to
address.
H
One
is
the
way
api
is
handling
right
now
it's
mounted
on
the
subdomains
and
we
don't
want
that,
and
the
second
one
is
additional
validation
before
we
redirect
so
everything
else
seems
to
be
already
on
the
pr.
So
I
hope
to
do
that
like
today
tomorrow.
So
stephen
can
make
another
pass
on
this
and,
in
the
meantime,
adding
support
to
companion.
H
I
hope
to
make
a
release
this
week
to
the
beta
channel,
make
it
as
soon
as
possible
because,
like
for
chromium
users,
that
takes
weeks
so
the
sooner
we
land
that
the
sooner
like,
hopefully,
we
will
be
able
to
like
coordinate,
go
ipfs
release
with
like
companion,
which
is
actually
capable
of
supporting
that.
H
H
Oh
yeah,
the
cool
thing
is
like
for
chromium
web
store
where
we
have
like
25k
of
users.
You
know
we
we
can
publish,
but
we
can
select
only
like
for
the
update
to
be
published
to
only
like
five
percent,
okay,
so
yeah
yeah,
no,
no
like
I've,
usually
the
better
smoke
testing
on
better
channel
is
enough.
So
I
would
not
worry
about
that.
It's
mostly
like
time
consuming.
That's
why
I
want
to
paralyze
this
work
this
week.
C
Yeah
so
this
week
just
kind
of
closed
out
a
couple
of
last
issues
adding
in
some
documentation
and
cleaning
some
things
up.
So
I
think
we're
now
now
testing
against
staging
right.
Stephen.
B
I
Chunking,
yes,
I
actually
didn't
get
very
far
because
I
caught
something
so
last
week
was
very
sleepy,
but
we
did
finalize
the
design
of
how
actually
important
export
should
work
I
for
on
on
the
ipfs
side,
this
is
actually
part
of
the
of
the
dagger
tool.
This
is
how
they
communicate
between
each
other.
I
already
did
some
prs
in
the
dependencies
of
ypfs
and
getting
ready
to
push
the
final
pr
for
or
go
ipfs
itself.
I
I
know
steven
you're
busy,
I'm
not
really
blocked
on
anything,
because
I
know
what
the
what
the
design
is
like.
So
you
know,
if
you
have,
if
you
have
better
stuff
to
review
review
the
better
stuff
on
the
dagger
itself,
a
little
bit
of
progress
on
stacked
chunker
interfaces,
basically
how
to
be
able
to
specify
multiple
in
line.
It
still
doesn't
work
correctly.
I
The
way
envision
it
to
hopefully
another
couple
days
should
they
should
land,
and
this
pretty
image
will
conclude
the
0.1
version
of
dagger
being
able
to
put
stuff
into
ipfs
and
being
able
to
specify
multiple
changes
one
after
another
and
that's
all.
I
have
right
now.
C
B
Got
it
then
rest
ipfess.
B
Mark
we
can't
hear
you
muted.
J
See
is
that
better,
yep,
okay,
yeah
so
phase
1.0
of
the
grant
is
complete.
Pending
the
report,
the
milestone
report
that
I'll
be
submitting
on
tuesday
or
wednesday
of
this
week
and
we're
also
working
on
two
blog
posts.
One
will
be
on
the
ipfs.io
blog
and
one
will
be
on
from
us
at
equilibrium
to
go.
You
know
go
up
together
as
of
right.
Now
we
have
the
ipfs
path
set
up,
so
the
folder,
the
repo
folder,
the
id
api,
the
version
api
and
the
swarm
api
are
either.
J
J
J
What
we're
doing
now
is
just
keeping
things
as
idiomatic
to
the
rust
language
as
possible
and
keeping
things
consistent
with
stevens
and
volker's
recommendations,
but
we're
you
know
we're
going
to
move
on
and
basically
circle
back
to
whatever
the
decision
that's
made
there
unless
something
happened,
the
last
72
hours
or
so
that
I
hadn't
heard
about.
B
At
thank
you
you're.
Welcome
any
other
initiatives.
B
I
take
that,
as
I
know,
then
we
have
design
review
proposals
which
I
have
one,
which
is
what
we
just
talked
about.
So
I
don't
want
to
beat
this
too
much,
but
we
do
need
to
have
this
discussion.
B
B
So
we
should
have
a
meeting
where
we
discuss
the
trade-offs
here
and
decide
if
we
want
to
continue
going
forward
and
switching
to
cid's
in
the
box
store
or
if
you
want
our
site
to
multihatch
the
box
store,
or
we
want
to
do
something
like
switch
to
cidv1
in
the
block
store
just
to
like
make
sure
that
we're
all
on
the
same
page
in
terms
of
the
trade-offs
it's
kind
of
at
the
11th
hour.
But
we
need
to
have
discussion
just
to
make
sure
we're
on
the
page.
B
So
please
read
this
issue,
read
all
background
and
if
you're
interested
write
your
name
here
and
I'll
schedule,
a
meeting.
B
So,
yes,
that's
one
of
the
searches
proposed
here
where
we
could
have.
We
could
store
the
block
along
with
a
set
of
codecs
for
the
block.
B
B
You
write
the
block
and
then
separately
you
write
like
some
additional
codec
information.
Do
you
write
like
slash
codex,
slash,
multihash,
slash
the
actual
kodak.
E
B
B
So,
like
we
could
do
is
just
say
everything
switched
the
cidv1
inside
the
block
store
and
that
will
get
rid
of
99.9
percent
of
the
problems
is
still
the
problem
of
like
I
would
like
to
be
able
to
like
take
a
unix
of
a
stag
turn
it
into
roblox
and
build
my
own
custom
tag
on
top
of
this,
as
kind
of
like
a
special
like
almost
like
a
car
format
where,
like
I
can,
have
this
really
unbalanced.
Whatever
tag
that,
like
you,
don't
necessarily
have
to
understand
all
like,
and
I
don't.
B
What
codex4,
but
then
I
can
sort
of
like
take
all
those
blocks,
turn
them
into
a
sequence
of
like
raw
blocks.
Build
this
nice
even
like
easier
traverse,
stack
on
top
of
that
and
tell
you
to
fetch
that.
So
that's
a
feature
that
I
would
like
to
be
able
to
have,
which
would
like
require
something
like
like
this
but
yeah.
Basically,
we
have
to
walk
through
all
the
different
trade-offs
here.
B
This
is
a
totally
optional
meeting,
attend
it
if
you're
interested
in
the
problem
and
you
it
affects
you
otherwise,
like
yeah.
E
I
Stephen
the
thing
you
just
said
that
I
want
to
have
x.
Can
you
put
this
in
the
issue
like
put
it
in
words
because
yeah
it
will
it's
not
something.
Many
have
thought
about
myself
included,
so
it
will
help
kevin
get
to
think
about.
J
I
put
in
a
small
request
that
jonas
from
the
equilibrium
rust
ipvs
team
could
attend
this
meeting
as
well.
What
would
be
the
best
way
to
ensure
that
he
gets
the
information
of
like
where
and
when
that
meeting
would
be.
I
B
Yeah
no
like
yeah
this
is
this
is
mostly
around
there.
So
this
is
the
meeting
around
making
the
decision
so
like.
If
you
come
to
the
meeting
and
you're
asking
a
lot
of
questions,
we're
probably
not
going
to
answer
a
lot
of
questions
so
read
up,
read
through
very
carefully
all
the
context.
B
J
He
would,
I
don't
think
he'll
be
disruptive
or
anything.
I
think
he
just
wants
to
know
one
that
will
be
implementing
it,
so
I
think
it
would
be
good
to
have
him
there.
B
Yeah,
I
just
want
to
set
the
temperature
like
that,
just
the
way
we
run
these.
These
design
decision
meetings
are
basically
like.
Okay,
everyone
who
joins
tries
to
have
most
of
the
content
ahead
of
time,
and
then
we
like
basically
say
we
will
come
to
a
decision
by
the
end
of
this
meeting
and
it's
a
forcing
function
that
we
use
so
like
because,
like
async,
you
kind
of
just
drag
things
on
forever
and
ever.
B
B
Okay,
thank
you.
Blockers
asks.
G
Steven,
if
you
take
a
look
at
the
pure
filtering
pr
in
cad
dht,
we
should
make
sure
that
that's
where
it
needs
to
be.
Thank
you
for
my
turn,
write
that.
B
K
Just
the
heads
up
guys,
I'm
gonna,
be
playing
docs
together
over
the
next
week
for
the
release,
so
I'm
probably
gonna
be
pinging.
Various
people
in
this
call
about
just
to
clear
up
things
that
I
don't
understand
like
what
is
bit
swap
and
how
do
you
actually
spell
it
things
like
that,
so
yeah
just
keep
looking
there's
something
you
should
look
at.
B
Here
is
an
initial
update
to
the
changelog.
I
will
put
it
in
the
notes
under
the
release
section.
B
It'll
stimp
in
the
chat,
if
you're
interested
in
seeing
what's
coming
in
the
release,
it
doesn't
have
everything,
but
it
has
the
errata
all
the
stuff
that
people
need
to
know
or
most
of
the
stuff
that
people
need
to
know.
So,
like
it's
things,
it
might
be
interesting.
It
could
be
helped
by
previous
content
on
ips
wps
like
this
will
tell
you
like
the
important
things
that
like
may
be
changing
or
may
affect
you,
okay,
okay,
so
that's
box
asks
actually
another
ask
I
have
there
is,
let's
see.
B
Yeah,
please
contribute
release
notes
if
you
know,
if
anything
has
changed
that
I
don't
please
update
it
and
I'm
running
over.
Anyone
have
any
questions.