►
Description
Meeting description here:
https://github.com/ipfs/team-mgmt/issues/992#issuecomment-578806684
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
B
C
C
There
yeah
yeah
in
the
right-hand
corner
your
you
might
have
some
people
like
covering
it
right
now,
but
if
you
move
those
people
out
the
way,
then
you
will
see
that
version
42
the
actual
bubble
size
is
way
way
smaller
than
it
used
to
be
so
hooray
for
that
and
that
so
that's
just
the
pre-release.
We
will
get
to
the
final
release
at
some
point
soon
and
then
the
second
thing
I
wanted
to
talk
about
was
that
Jase
ipfs,
Syria
41
is
coming
soon.
C
The
the
reason
that
there's
a
pre-release
of
httpclient
is
that
the
async/await
pull
requests
all
got
merged,
so
we
are
just
finishing
up
bits
and
pieces
that
need
to
be
done
for
the
JSI
PFS
release
and
the
highlights
are
the
async/await
stuff,
but
also
the
UNIX
FS
version
1.5
stuff.
That
Alex
is
all
of
the
API
changes
for
which
there
are
a
lot
and
they
are
breaking
are
in
the
release.
Issues.
So
have
a
look
at
that
and
there
is
a
migration
guide
for
everyone.
C
Again
there
we
go
cool,
so
these
are
the
results
of
a
test
run
in
the
interrupt
tester
suite
on
the
left
is
the
new
IP
FS,
with
async
await
stuff
merged
and
on
the
right
is
the
current
version
that's
already
out,
and
so
what
this
is
is
this
is.
This
is
doing
interrupt
test
between
two
different
Jas
I
saw
two
different
IPS
implementations,
and
you
can
see
here.
We've
got
like
a
J's
to
go
and
chairs
to
j/s
nodes
here
and
what
the
tests
do.
C
Is
they
spin
up
two
nodes
transfer
a
single
file
between
those
two
nodes?
There's
no
one
else
involved
so
they're,
pretty
simple,
really
but
I'm.
Looking
at
these
results
here-
and
you
can
see
that
you
know
on
the
left-hand
side,
every
everything
here
is
taking
a
whole
lot
less
time
they're
used
to.
If
we
look
at,
for
instance,
this
whoops,
so
this
67
Meg
file-
this
takes
about
eight
seconds
now
it
used
to
take
27
seconds
so
whoa,
okay
way
faster
in
some
cases.
C
D
C
E
B
Okay
in
terms
of
God
profess,
we
think
that
is
your
program
for
23
will
come
out
today.
Unless
we
have
a
lot
of
discussions
around,
she
bought
ghosts,
but
the
blog
post
looks
great.
So
please
take
a
look
at
it
quickly,
so
we
can
ship
to
release
yeah
everything
seems
to
be
stable
and
fixed
a
bunch
of
bugs
okay.
D
F
Well,
I'm
tight
from
my
site
is
that
infrastructure
is
getting
more
stable
right
now.
You
can
easily
start
like
500
or
1000
test
plan
instances
on
kubernetes.
It's
documented
and
it's
very
straightforward,
just
like
a
few
steps
that
you
need
to
run
and
wait
for
your
cluster
to
come
up
and
be
running
where
I
go
above
500
test
plan
instances
on
the
DHT
test,
I'm
starting
to
see
some
timeout
issues,
but
it's
still
not
clear
whether
this
is
infrared
light
or
whether
it
is
test
related.
F
For
example,
I
see
that
we
have
some
retry
mechanism
there
and
the
retry
doesn't
seem
to
be
working
because,
even
though
we
retry
the
dialer
has
some
back
of
time,
so
actually
we're
not
really
trying
any
dials
so
yeah.
If
someone
knows
more
about
the
DHT
test
plan,
somehow
there
would
be
appreciated
or
I'll
be
looking
at
that
in
the
coming
days
as
well.
Yeah,
that's
pretty
mature
from
my
site.
Yes
Steven!
You
want
to
take
cover
there
more
updates
on
the
intro.
Okay.
B
B
E
Jacob
give
update
yeah
yeah,
so
we've
gone
through
and
created
all
of
the
high-level
like
epic
level
or
feature
set
issues
in
there
all
in
github,
with
notes
on
all
of
the
design
notes
that
we've
worked
through
and
welles
testing
scenarios
and
what
success
means.
So
this
week
we're
going
to
be
working
on
breaking
down.
The
remainder
of
the
work
in
a
smaller
task
is
smaller
tasks
and
estimating
those
out
so
that
we
have
a
better
idea
of
when
we
believe
that
everything
will
land.
B
G
G
What's
going
on,
we
put
together,
I
was
working
Stephen
last
week
a
lot
we
could
see
either
some
like
graphs
to
kind
of
see
like
what
the
state
of
the
connections
of
the
network
look
like,
because
they
were
seeing
lots
of
like
disjointed
networks
which
is
pretty
concerning,
and
maybe
these
are
just
occurring
it
like
very
small
network
sizes,
it's
easier
to
occur,
but
the
test
is
gonna.
Have
some
issues
like
multiple
bootstrappers
can
call
it?
The
way
they're
set
up
right
now
will
cause
destructed
networks
or
the
connection
manner.
B
If
you
both
have
time
today,
they
do
get
together
and
try
to
like,
and
how
could
you
start
adding
on
using
kubernetes?
It
hasn't.
B
G
F
H
Good
news
is
that
it's
mostly
on
the
basic
functionality
is
implemented.
The
remaining
part
is
writing
tests
and
dog
finalizing
stuff.
The
only
thing,
maybe
that's
more
for
like
a
separate
review
but
I
think
it's
small
enough
to
mention
here
is
that,
like
under
gateway
section,
we
got,
we
historically
had
this
thing
called
path
prefixes.
It
was
not
documented
for
a
long
time
and
all
it
like
provided
dogs
that
it's
about
like
this
header
and
the
problem
sort
of
like
a
problem
is
that
we
are
reusing.
This
disliked
name.
H
In
those
like
hostname
specific
sections,
but
we
use
it
for
a
kind
of
totally
different
feature
which
is
not
related
to
this
header.
It's
not
wild
up,
so
something
that
remains
to
be
decided,
maybe
like
in
I,
think
I'll
write,
write
those
concerns
up
and
ask
you
for
a
review.
Do
we
want
to
like
basically
to
split
it
into
two
things
or
clarified
the
docs?
B
B
H
I
No,
it's
not
me
it's
my
my
doppelganger
over
there
in
Germany,
so
big
swap
updates.
Basically
I've
been
running
some
tests
against
test
rounds,
so
I
wanna
give
a
shout-out
to
Anton
and
Stephen
and
round
for
getting
that
into
a
state
where
you
can
actually
run
some
tests
and
try
out
different
variables.
It's
pretty
pretty
exciting,
so
I've
managed
to
produce
a
few
graphs
that
I'm
gonna
I'm
gonna,
wait
or
hold
off
on
showing
until
next
week.
I
So
there's
a
little
teaser
for
you,
but
I
just
want
to
make
sure
that
all
my
tests
are
running
consistently
and
then
I
can
show
some
some
decent
comparisons.
The
good
news
is
that
it
seems
like
the
changes
we've
been
making
are
effective
and
the
got
some
pretty
big
improvements
over
the
old
bits,
but
can
I
give
a
spoiler.
B
B
J
C
Really
I
have
put
at
the
end
of
their.
Let's
maybe
remove
this
initiative
now,
because
I
think
the
updates
will
be
like
there
won't
really
be
anything
else,
we're
looking
to
maybe
release
JSI
BFS
release
candidate
this
week.
So
there's
not
really
going
to
be
much
here
from
now
on
anyway,
so,
let's
just
remove
it
unless
there's
any
location.
C
A
G
A
G
C
C
B
J
Streaming
yes
connected
news
for
last,
so
I
really
really
was
hoping
that
I'll
be
able
to
show
the
actual
thing
today
during
lightning
talks.
But
after
a
marathon
session
turned
out
that
yeah,
it's
not
going
to
happen.
So
I
am
having
a
lot
of
trouble
with
not
going
fast
enough
for
test
purposes
and
I
have
one
more
idea
to
try.
J
So
you,
so
you
basically
can
experiment
with
with
different
chemicals
and
stuff
like
that,
without
him,
turret
whole
bunch
of
code,
and
unfortunately
it
is
a
little
slower
than
going
theaters,
which
was
definitely
is
definitely
going
to
be
showstopper.
But
I
will
help
you
with
this
further
down
the
road.
So
next
step
is
to
put
everything
back
together
can
do
it
was
today
for
this
day
and
have
it
actually
available
tomorrow
for
people
to
play
with.
B
J
A
little
bit
I
think
I.
It's
like
it's
it's
one
of
those
things
that
I,
it's
not
that
I'm
doing
things
inefficiently,
like
my
allocation
profile,
is
essentially
non-existent.
I
pretty
much
okay,
the
huge
block
in
the
beginning,
I
just
do
a
lot
of
stuff
that
doesn't
have
to
do
because
of
the
scan
for
the
repeat,
bytes
and
stuff.
Like
that,
you
know
the
one
thing
that
I'm
going
to
try
is
to
basically
add
more
hints
of.
When
can
we
skip
stuff?
B
J
J
D
J
D
J
B
J
J
C
J
J
B
C
C
D
B
E
D
B
A
C
B
D
Yeah,
well,
we
were
talking
about
tea
size
field
of
the
dyed
links.
Basically
like
it's
an
unreliable
field,
it's
a
vector
for
attack.
It
can
be
expensive
to
generate,
depending
on
the
api's
that
you're
using
but
yeah.
So
I
proposed
deprecating
it
because
we
don't
have
to
remove
it
because
you
to
help
red
spots,
have
a
zero
value.
It'll
come
back
a
zero
for
them,
specifying
it's
an
optional
field,
but
yeah
it's
given
points
out.
You
know
a
15-2,
lien
might
break
and
that
kind
of
thing
I
just.
B
Feel
like
if
we
had
a
call
to
just
talk
about
this,
we
could
quickly
sync
up
on
all
the
problems.
Like
my
perspective
on
this,
is
it's
been
ever
ever
and
one
who
rely
in
it
anyways
so
I
don't
want
to
say.
Well,
you
know
it's
not
you,
but
I
feel
like
you'd
be
happy.
You
might
want,
have
a
special
others
unless
you
don't
produce
necessary.
B
D
I
mean
can
refactor
how
Joe's
impulse
as
it
likes
that
the
problem
that
we
have
is
that
the
IPL
dapi
is
don't
tell
you
the
serialize
size
of
a
thing
after
you
see
realized,
it
said
the
only
way
to
find
out.
How
is
what
you
need
for
the
for
the
end
to
know
the
size
of
the
thing
that
you're
linking
to
the
a
way
to
do
now
is
to
serialize
it
and
then
pass
it
to
IP
OD,
see
the
CID,
but
from
IP
LD,
and
then
you
have
the
serial
Oh
sighs
from
your
own
serializing.
D
So
you
end
up
serializing
it
twice,
which
is
17
years,
but
we've
kind
of,
like
we've
talked
in
the
past
about
refactoring
the
important
to
not
use
IP
LD
to
just
use
like
the
I
feel
D
tag.
Pb
module
to
you
know
know
how
to
read
and
write
the
protobufs,
but
to
not
actually
use
IP
o
D
itself
to
do
the
serialization.
B
I
know
you're
trying
to
integrate
new
system
and
it's
tricky
okay,
I,
but
ideally
we
should
probably
actually
schedule
call
for
us.
Do
you
thinking
to
try
to
find
time
that
one
can
meet?
Yeah
sure?
Is
it
if
anyone
is
interested
in
discussing
link
sizes
and
act
bébé,
please
put
your
name
down
here
and
let's
go
or
make
a
call.
D
B
B
Yeah,
just
going
back
async
over
these
discussions
will
take
a
while,
so
we
have
any
because
if
it's
discussion
of
I
do
actually
out
get
everyone
on
the
same
page,
because
I
know
I
missing
a
lot
of
state
on
how
like
the
motivation
for
why
we're
doing
this?
Probably
looking
to
this
okay,
that
covers
sign
of
you
proposals
as
far
as
any
no
one
listens
has
another
one.
G
I'm
not
sure
whether
here
are
a
different
venue
is
appropriate,
but
we're
gonna
have
to
have
some
discussions
about
the
connection
manager,
trimming
stuff
and
like
what
approach
we
want
to
take
to
dealing
with
that.
There's
a
few
options
available.
I
don't
put
some
design
proposal.
If
that's
just
the
content
routing
group
having
a
call
tomorrow,
yeah.