►
Description
Meeting description here:
https://github.com/ipfs/team-mgmt/issues/992#issuecomment-581457823
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
You
don't
want
recorded
alright.
Here
we
are
it's.
The
recording
has
started
now,
hi
everyone.
Oh
hello,
welcome
to
the
ipfs
core
implementations
weekly
sync,
it
is
the
3rd
of
February.
It
is
2020
already
and
we're
in
the
second
run
for
ready.
Can
you
believe
it
cool?
If
you
are
here,
please
add
your
name
to
your
attendees
list.
A
I
will
post
the
link
in
the
chat
just
in
case
anyone
missed
it
and
if
you
have
an
async
update,
if
you
have
been
working
on
some
things
and
would
like
to
update
some
people
on
what
you've
been
up
to,
we
won't
go
through
a
sync
updates
in
the
meeting,
but
please
do
add
them
to
the
bottom,
and
so
everyone
can
check
out
what
you've
been
up
to
and
ask
questions
he's
synchronously.
A
But
what
we
will
do
now
is.
We
will
go
through
our
high
priority
initiatives.
We
have
a
note-taker
already
Thank
You
Jacob,
so
high
priority
initiatives.
So
let's
get
started
upcoming
and
all
shipped
releases
I
have
an
update
here
for
well,
it's
kind
of
an
update,
Jer,
cyberfest,
Syria,
41
DRC
is
landing
imminently.
A
Essentially,
I
want
to
publish
it.
After
this
meeting,
so
that's
cool,
there
was
a
blog
post
and
explain
about
blog
post
to
explain
a
lot
of
the
API
changes
that
are
happening
in
this
release,
the
reef
factors
that
have
been
done
and
the
reasons
behind
them
and
to
give
everyone
a
chance
to
digest
those
changes,
figure
out
what
what
it
means
for
them
and
their
codebase
and
and
and
hopefully
get
a
head
start
on
those
changes
before
they
actually
land.
So
that's
been
released.
The
link
is
there
in
the
notes.
A
Jacob
is
here,
but
since
last
time
the
PTP
has
shipped
not
so
much
news,
I
guess
other
than
it's
there
there's
a
blog
post
in
the
in
the
pipeline.
Well
done
to
everyone
who
has
worked
on
that,
and
thank
you.
That's
really
cool
news
and
the
other
final
thing
is
that
go
ipfs
zero
4.23
got
released
a
bug,
fix
release,
fixing
some
critical
issues,
there's
also
a
blog
post,
to
explain
what
is
going
on
in
that
release
on
the
website.
B
A
C
So
from
testament
perspective,
many
things
happened
last
week,
so
we
merged
compositions,
which
is
based
to
the
feature
that
allows
you
to
run
various
groups
of
different
images
on
test
ground.
We
merged
a
feature
which
allows
you
to
basically
store
anything
that
you
want
from
your
test
line
on
an
s3
bucket.
So
now,
if
you
run,
let's
say
a
thousand,
no
two
amount
with,
let's
say
the
DHT
test
plan,
you're
free
to
store
whatever
assets
or
artifacts.
You
want
on
the
s3
bucket
and
then
easily
browse
through
them.
C
Apart
from
that,
many
other
small
improvements
and
tomorrow
will
probably
be
ready
with
the
version
0.1
release.
We
are
just
up
on
the
locks
and
the
outputs
part
of
it.
So
that's
pretty
much.
What's
going
on
I
guess
during
tomorrow's
go
to
be
more
stuff
communicated.
A
C
C
A
E
F
I
put
that
in
there
he's
gonna
be
working.
He
got
feedback
last
week
from
Raul,
so
he's
gonna
work
on
getting
that
feedback
done
and
just
work
trying
to
get
it
over
the
hill.
While
it's
not
a
the
biggest
priority
and
the
grand
scheme
of
content
routing,
it's
almost
done
so
we
want
to
just
finish
it
now.
B
Okay,
my
update
here
is
also
we
have
some
open
desks
and
know
what's
currently
working
on
these.
So
if
you
are
finding
yourself
at
loose
ends,
we
have
a
son,
Athena's,
bow
bug,
and
let
me
be
this
mu
there
for
ages,
but
like
we're
not
like.
Basically
it's
something
we
can
fix
easily,
but
something
we
can
sort
of
work.
Obviously
the
problem
is
two
pairs
connect.
They
both
see
the
in
Banff
connection,
so
they
cancel
around
connection
because
they
haven't
seen
that
finished
yet
in
the
gold
walk
away.
B
What
we'd
like
to
do
is
just
say
file
they
connect
and
they
just
keep
two
connections
and
whatever
we
deal
with
it,
the
issue
is
on
an
improvements
they're
fun
to
bugs
in
the
ordinary,
build
plenty
of
relying
on
this.
So
if
you're,
just
this
talk
to
me
and
all
like
when
you
write
different
things,
I
need
to
get
fixed
here.
Finally,
tests,
if
you're
less
familiar
with
go,
we
could
really
use
help
like
building
tests,
especially
I
should
think
they're
a
bunch
of
them
leftover
bits.
B
F
A
Nice,
thank
you
all
I
am
I
would
like
to
help
you
with
that
stuff,
I'm,
hoping
that
by
next
time,
this
next
time
we
meet
this
time
that
quick,
I
will
have
been
able
to
put
down
the
or
hand
over
the
reins
to
on
the
JSI
PSS
train
to
the
aching
brain
Alex,
Co,
CDs
and
infant.
He
could
drive
it
for
a
while
and
I
can
help
out
on
the
on
the
content.
Routing
stuff
I
am
excited
to
get
involved.
A
Let's
do
it,
but
hopefully,
hopefully
by
the
Sun
that
fake
the
Jessica
press
release
will
be
out
the
door
and
I'll
be
freed
up,
but
yeah.
If
other
people,
obviously
you
know
watching
the
cool
want
to
help
out
then
also
come
help
us.
We
need
help
alright,
rad
content,
routing
done
set
the
main
gateway
Lloyd.
Who
do
you
have
an
update
for
us
yep.
G
G
What
remains
to
be
done
in
the
PR,
our
harness
tests
for
all
the
edge
cases,
I
got
distracted
by
aquifers,
desktop
release,
and
that's
like
zero
for
me
and
this
he
won
now.
So
when
I
after
I
give
us
the
stop,
release,
I
hope
to
finalize
this
this
week.
What
we
also
agreed
with
Stephen
is
kind
of
increasing
a
scope
of
this
PR
a
little
bit
and
adding
this
hardening
of
pathways
gateways.
G
B
I
believe
I'm.
Sorry,
everybody
I
realize
now
that
I'm,
a
terrible
person,
increasing
scope
of
PR
is
probably
not
but
I
should
have
told
you
to
do
it's
not
that
big
increase.
It's
just
like
another
flag
when
you
thread
and
if
it's
simpler,
like
it's,
whatever
simpler
I
give
it
some
purchased
emerge,
but
we
currently
have,
and
then
you
this
is
a
second
patch.
Let's
do
that.
B
H
A
I
I
can
verify
the
Steven
is
not
a
terrible
person.
It's
you
reviewed
my
9000
line.
Pr
the
bits
are
proof
of
concept
and
we
finally
got
it
merged.
So
that
was
pretty
exciting
moment
in
my
life,
probably
the
best
day
of
my
life.
So
far,
so
so,
basically,
last
week,
I
spent
quite
a
bit
of
time
doing
performance,
profiling
and
I
made
some
grass
and
stuff
there's
a
link
to
an
issue
there.
That
has
all
the
graphs
in
it
a
little
bit
of
a
little
bit
of
sieve
analysis.
I
Guys
see
that
yes,
so
so
yeah
his
a
few
of
the
performance
analyses
we
were
done
this
week,
so
this
is
comparing
master
or
what
was
master
at
the
time.
So
the
old
bit
swap
this
is
the
the
Nuba
top
that
we've
just
merged.
So
we're
mostly
concerned
with
how
long
it
takes
to
get
stuff
and
a
number
of
duplicate
blocks
that
get
sent
around,
in
other
words,
extra
data
that
we
don't
really
need
to
be
sending
if
we're
100%
efficient.
I
So
this
first
graph
or
this
first
set
of
graphs,
is
simulating
a
network
with
five
millisecond
latency
and
with
varying
bandwidth.
So
this
first
one
is
one
gigabit.
This
is
hundred
sixty
Meg's
forty
megabytes,
20
megabytes
and
ten
megabits
megabytes
of
bandwidth,
and
you
can
see
the
blue
lines
are
the
old
bit
swap
and
the
yellow,
orange
and
red
lines
are
the
new
bit
swap
so,
for
example,
at
one
gigabit
bandwidth,
you
can
see
that
these
three
red
lines
are
a
lot
faster
than
the
blue
lines.
I
So
the
darkest
line
here
is
when
there's
four
seeds
and
one
leech
the
next
one
is
to
season
one
leech,
and
this
is
when
there's
once
even
one
leech,
so
the
new
bits,
what
outperforms
orbits
are
pretty
significantly
when
there's
high
bandwidth,
when
we,
when
we
get
down
to
pretty
low
bandwidth,
it's
quite
similar,
but
still
this
still
look
quite
a
big
improvement
there,
and
then
this
set
graphs
is
just
about
comparing
performance
when
there's
very
variable
numbers
of
leeches
in
seats.
So
in
this
case
there's
three
leeches
and
three
seeds.
I
So
I
guess
the
question
we're
asking
is:
are
we
taking
advantage
of
downloading
from
all
the
seeds
or
do
we
sort
of
send
all
our
traffic
to
one
seed
and
see
that
the
other
line,
the
new
bid
swap
performs
significantly
better
than
the
old
bit
so
I,
particularly
for
high
bandwidth
scenarios?
This
one
is
about
to
placate
blocks,
so
basically
the
new
new
bid
swamp
doesn't
really
send
duplicate
blocks.
Instead,
it
sends
out
a
request
to
all
the
peers.
I
B
I
go
on
about
that.
No
dude,
we
don't
just
say
we
don't
have
to
franchise
every
block
and
said
the
guy's
one
of
the
peer,
this
block
and
all
the
other
peers.
Do
you
have
the
block?
We
also
have
a
feature
really
that
people
come
back
until
snow
I.
Don't
have
a
lot,
so
we
can
me
to
leave
like
look
at
the
other
peers.
I've
said
yes,
I
have
the
watch
angel
and
ask
them
like
one
that
we
know
will
actually
work.
So
it's
not
like.
We
have
like
two
accurate
requests.
B
Now,
it's
more
than
like.
We
have
a
best-case
one
round
trip
worst
case
two
round
trips.
Instead
of
like
send
a
request.
Wait
for
time
to
have
another
request
related
to
requests,
wait
for
timeout
said
for
requests,
wait
for
timeout
and
they
keep
us
any
more
of
a
request.
If
I
don't
get
back
ton
of
watts
office.
I
Thanks
yeah,
that's
that's
a
good
point
yeah.
So,
basically
we're
able
to
just
take
advantage
of
these
new
message,
types
to
make
things
faster
and
more
more
efficient.
So
it's
not
100%
ready
yet,
as
we
still
have
to
make
a
couple,
little
changes
to
make
sure
that
the
two
versions
are
interoperable.
That
is
that
the
new
version
can
still
speak
to
nodes,
running
the
old
version
and
I'm
still
writing
a
few
kind
of
like
fuzz
testing
plans
and
things
like
that.
We're
pretty
close
any
questions.
A
That
sounds
awesome.
I'm
really
excited
can
I
request
that
you,
because
you
did
a
presentation
on
the
changes
that
you
were
gonna
make
a
while
ago.
Can
you
put
a
link
to
that
in
the
in
the
notes
that
if
anyone
wants
to
check
it
out
and
they
can
get
a
good
overview
of
what's
actually
changing,
yeah
sweet
all
right?
Let's
move
on
I
know
we're
running
a
bit
behind,
but
that
was
that
was
worth
it.
So
we've
got
stream
based
content
based
no
stream
content
based
junking
research,
/
improvements,
catchy
Peter.
H
So
last
ones
the
end
I
actually
got
most
of
the
work
done
on
the
tool.
Member
I
think
to
be
able
to
evaluate
different
options
for
a
change
in
the
algorithms.
Then
I
had
a
meeting
with
Eric,
who
told
me
a
lot
of
stuff
about
so
as
a
result
of
that
I'm
a
little
bit
in
writing
the
entire
thing
like
there
is
not
a
single
medical
care
anymore,
but
I
am
close
enough
for
us
to
be
able
to
miss
this
tomorrow
when.
H
H
H
A
It
thank
you
all
right
cool.
Thank
you,
everyone
for
your
updates
on
initiatives,
the
we
have
a
section
called
other
initiatives.
These
are
initiatives
that
are
like
in
backlog
or
on
pause
or
but
are
still
in
in
they
are
in
cold
storage
waiting
for
us
to
bring
them
back
to
life.
Again.
At
some
point,
we
have
a
question
on
universe,
1.5.
B
I
B
Just
want
to
ask
okay
shape,
we
get
enough
for
monkeys
with
enough
keyboards
and
then
eventually,
one
of
them
time.
Okay,
the
next
one
is
a
team.
You
had
a
lot
of
numbers
from
the
XS,
add
improvements,
but
no
numbers
that
you
can
publish
and
we
have
new
like
a
cell
on
it,
takes
to
add
good
to,
but
we
also
old,
the
cell
long
takes
a
de
motu
but
they're
into
the
machine.
Is
there
any
chance
you
have
like
all
the
cell
on
takes
a
dip
into
on
the
same
machine?
J
One
that
I'm
using
right
now
is
that
it
previously
took
24
minutes
to
add
arch
linux,
and
now
it
took
11
minutes
to
add
arch,
linux
and
I.
Think
those
are
our.
You
know
apples
to
apples
but
I'm,
not
a
hundred
percent
sure.
So
if
we
just
like
update
the
issue
where
we
liked
it
all
of
that
add
performance
work
with,
like
you
know,
here's
what
it
used
to
be
like
the
table
of
the
times.
J
That
would
be
really
useful
for
you
don't
link
to
something
that
people
don't
have
to
like
scroll
to
in-flight
comments,
that,
like
tell
you
one
side
or
the
other
and
like
read
through
it
in
order
to
validate
like
yes,
Molly
isn't
lying
to
me
in
a
blog
post,
so
that
would
that
would
make
that
be
snazzy.
If
we
have
those
things
available.
D
D
F
For
distributed
signaling,
yes,
so
for
things
that
have
web
RTC
transports.
Yes,
so
basically
the
idea
as
part
of
this
we'd
also
need
direct
connection
upgrades.
So
we
would
use
a
limited
circuit
relay
connect
over
that
distribute
over
the
relay
and
then
then
merge,
but
in
theory
any
node
that
we're
mutually
connected
to
could
serve
as
the
signal
err.
Okay,.
A
So
I
was
watching
the
magic
gift.
Take
me
longer
to
get
back
to
here
than
I
expected
Thank
You
Hector,
that's
cool
all
right
design,
review
proposals.
There
is
a
proposal
for
ignoring
the
tea
size
property
in
dak.
Pb
links
is
happening
on
to
morrow
at
3:00
p.m.
GMT.
If
you'd
like
to
come
and
talk
about
that,
then
please
add
your
name
to
the
list
and
we
will
invite
you
to
the
meeting
and
he
take
us
put
your
name
on
the
list
in
the
notes.
D
B
A
J
K
Exactly
why
everyone
I'm
the
newest
member
of
protocol,
apps
from
Moxie
and
I'm,
really
excited
to
with
this
opportunity
to
work
with
an
international
team.
It
will
be
a
big
challenge
for
me,
because
I
never
worked
remotely,
but
I.
Think
with
my
good
communication
skills
I
will
be.
I
will
be
successful,
with
your
help
of
course.
So
let's
do
it.