►
A
Recording,
okay,
hi,
everyone
welcome
to
the
ipfs
core
implementations
weekly
sync:
it's
November,
the
4th.
If
you're
here,
please
add
your
name
to
the
attendees
list.
I've
just
put
the
link
to
the
hackpad
in
the
chat
and
if
you
are
here
at
some
point,
please
add
your
weekly
update
for
async
review
to
the
bottom
of
the
hackpad
and
we
will
not
go
over
the
in
a
meeting.
A
But
people
can
look
at
it
and
see
what
you've
been
doing
for
the
past
week
and
comment
and
ask
you
questions
when
this
the
meeting
notes
get
PR
to
the
team
management
repo.
So
thank
you
please,
and
your
update.
If
you
are
here,
we
have
a
note
taker
Jacob.
Thank
you
very
much
and
now
we
shall
begin
by
going
through
the
initiatives.
A
And
we
will
start
with
the
upgraded
release
process,
so
I
have
to
say
that
I
think
last
week
I
said
0:39
actually
got
released
and
we
we
started
the
0:40
release
issue,
but
it
we
have
not
yet
branched
off
and
I
was
meant
to
do
that
today.
But
I
did
not
do
that,
but
there's
a
couple
of
things
in
there
in
this
next
release,
which
is
a
repo
migration
tool
which
I'm
currently
waiting
kind
of
blocked,
I
guess
on
a
pull
request
for
that
Adam
is,
is
working
on.
A
This
Adam
is
an
awesome
contributor
who
has
been
doing
a
whole
bunch
of
stuff,
but
he
has
a
job
and
he
does
not
always
have
time
to
continue
things
so
I
might
have
to
pull
that
over
the
line
so
yeah,
that's
that's
the
status
of
that.
Maybe
that
gets
dropped
from
this
release
anyway.
So
that's
the
only
thing
I
have
to
say
about
the
release
process
for
the
chase
of
your
first
Steven.
Do
you
have
anything
on
go
ipfs,
releasing.
A
C
Yeah
I
mean,
let's
say,
let's
change
since
last
week,
so
we
just
got
like
I
think
it
was
I
kept
anyhow,
it
does
miss
right
now.
At
the
moment
she
was
trying
to
get
as
many
notes
as
he
can
running
the
DHT
test.
I
need
you
with
us
today
today
were
trying
to
get
that
done,
basically
trying
to
spinning
up
and
I'm
basically
working
automatically
and
then
also
get
like
some
basic
network
controls
like
latency
is
like
that
done.
A
C
A
E
F
E
Perfect
yep
and
I
also
while
I
was
at
that
I
realized,
we
sort
of,
could
add.
Support
for
I
can
like
see
IDs
in
I
pianist
paths
without
sort
of
like
being
blocked
by
this
entire
I
think
refractor
to
bubble
up
this
PID
change.
And
if
you
look
at
the
second
PR,
it's
basically
like
one
line
change
that
does
not
require
any
any
third-party
dependency
bumps.
E
E
It's
separate
topic,
separate
topic
is
switching
to
a
single
version
of
JSP,
ID,
I
sort
of
have
those
changes
like
on
my
machine
and
I
feel
I
can
like
submit
them
as
a
separate
PR
to
not
block
this
change,
so
I
probably
will
submit
them
as
a
a
PR
draft.
So
if
someone
is,
it
gets
eventually
to
the
peer
ID,
because
I
don't
know
how
to
bubble
that
up
on
the
b2b
side,
but
at
least
on
the
J
side
we
will
see.
What's
the
scope
of
changes
require
yep
I,
guess
that's
it.
D
A
A
D
So
the
big
discovery,
pub/sub
PR,
landed
and
there's
been
a
release,
so
huzzah
next
up
is
integrating
that
into
the
pub
sub
router
and
then
bubbling
that
up
into
go
idea.
Fest
so
should
be,
should
be
fairly
quick,
because
these
changes
have
already
been
made
and
reverted.
They
just
need
to
be
redone
again
and
yeah.
Hopefully
that
should
be
available,
and
if
s
Masterson.
D
Yeah,
alright,
so
so
we
did
some
more
like
looking
around
and
it
looks
like
in
order
to
make
I
cannot
like
in
order
to
make
ID
performance
faster.
We
basically
just
need
to
have
a
database
with
asynchronous
writes.
Instead
of
waiting
for
every
time
we
write
for
the
thing
to
get
flushed
to
disk.
You
can
either
do
this
a
statement
gun.
Well,
we
don't
absolutely
need.
D
So
we
I
I,
guess
it's
a
matter
of
phrasing.
We
we
definitely
need
this.
You
can
either
do
this
by
just
having
a
database
that
supports
asynchronous,
writes
or
you
can
do
this
by
writing
a
layer
on
top
of
a
synchronous
database.
That
effectively
makes
it
asynchronous
right
by
like
offering
stuff
in
between.
One
of
these
is
like
flip,
a
switch
and
badger
and
the
other
one
is
make
the
buffer
dag
part
of
our
API
perform
better.
D
You
could
also,
you
could
also
just
try
and
like
get
around
this
a
little
bit
by
putting
in
a
synchronous
layer
on
top
of
the
synchronous
one,
just
like
caching,
all
the
transactions
and
like
periodically
flushing
them,
but
that
feels
even
happier
than
than
just
making
the
buffer
dag
better.
So
we're
gonna,
try
and
push
and
the
links
in
the
in
the
notes
there's
a
link
to
the
making
the
data
allowing
the
data
stores
to
be
asynchronous.
I'm
gonna,
try
and
push
on
some
people
this
week
to
see.
D
C
A
Super
thank
you
all
right:
migration
to
Martina
she's
in
block
store.
This
is
still
pending.
The
migrator
tool
will
actually
being
merged
into
ipfs
repo.
The
pull
request
has
been
reviewed
and
that
just
needs
to
nice.
The
changes
to
be
done
and
then
to
be
washed
and
released,
and
then
then
we're
good
to
start
that
will
create
the
migration
that
does
that.
Does
that
work
and
I
think
that
actually
has
already
been
done
in
our
requests
for
it,
but
it
just
needs
to
be
then
reviewed
as
well
so
yeah.
A
A
B
C
C
A
G
Yes,
so
the
plaintext
to
internal
crypto
refactor
for
the
p2p
and
private
networking
has
all
been
merged.
So
it's
good
to
go.
Bosco's
currently
working
on
the
peer
store
refactor.
So
it's
the
pure
book
and
then
the
Registrar,
which
is
going
to
be
part
of
connection
manager
and
that's
gonna.
Let
us
register
like
topologies
and
custom
topologies
and
stuff,
which
is
going
to
be
quite
nice,
and
then
the
pub/sub
PRS
are
also
in
progress.
Getting
close,
they
just
need
some
stuff
from
the
registrar
in
the
peer
store
and
then
identify
reef
actors
under
way.
A
G
A
Okay
cool,
so
then
we
have
design
review
proposals
we
I
so
last
week,
I
said
that
I
would
start
a
proposal
thing
uniting
the
files,
API
and
I've
been
working
on
that
I've
been
writing
a
lot
of
words
down
and
I'm
thinking
about
it.
A
lot.
It's
not
quite
finished.
I
haven't
submitted
the
pull
request,
but
it's
kind
of
nearing
it.
But
it's
it's
a
proposal
and
kind
of
I.
Guess
I
put
it
down
here
today,
just
to
sort
of
see
how
gauged
they're
kind
of
what
people
are
feeling
like
is
it?
A
Is
it
a
non-starter
or
is
it
is
it?
Is
it
something
that
we
feel
could
could
take
off
like
if
you've
got
big
concerns
with
it?
All
it
didn't.
Please
like
just
get
in
touch
with
me,
I
see,
or
you
know,
slack
or
whatever,
but
I'll
I'll
probably
get
it
well.
I
need
to
read
through
it
a
number
of
times
and
see
and
there's
a
but
there's
a
few
things.
I
need
to
kind
of
finish
off
and
add,
but
then
I
will
submit
a
PR
and
open
it
up
to
people
to
have
a
look
through.
A
We'd
have
to
make
as
small
as
I
can,
but
there's
a
couple
of
small
things
that
we
feel
we
can
do
to
really
improve
things.
But
there
are
some
bigger
things
in
there
as
well
so
like.
If
you're
interested
in
that
then
have
a
look.
I
will
hopefully
submit
the
proposal
and
we
can
maybe
schedule
like
a
and
maybe
the
next
week.
I'll
have
a
like
scheduled
meeting
and
we
can
kind
of
people
who
are
interested
in
it.
Can
we
can
go
talk
over
some
of
the
changes
and.
A
Get
everyone
involved
who
needs
to
have
a
say
in
this
because
I,
obviously
it's
it's,
it's
the
files
API
it's
quarter,
ipfs!
It's
the
the
thing
that
people
are
going
to
be
interacting
with
most
when
they
use
IP
FS,
so
it
needs
to
it
needs
to
be
good.
It
needs
to
be
low,
streamlined
and
easy
to
use
and
I'm,
hoping
that
we
can
maybe
make
it
better.
So
yeah,
that's
me
and
then
alex
has
a
parole
for
improving
the
agents
being
API.
Would
you
like
to
talk
to
that.
H
H
So
if
you
have
a
streaming
interface,
then
where,
with
all
this
async/await
stuff
we're
moving
towards
having
streaming
stuff
everywhere,
if
you're
in
the
browser,
you
know
browser
support,
HTTP
trailers.
So
if
you
you're
processing
a
stream
and
the
stream
errors,
then
the
stream
just
ends,
and
you
don't
get
an
error
message
or
anything.
This
has
worked
in
node
previously
because
no
does
support
HTTP
trailers,
but
since
we're
moving
everything
to
use
fetch,
it
started
to
bring
browser
semantics
into
node
and
words
along
with
those
browser
semantics.
That
means
no
more
HTTP
trailers.
H
So
it
means
that
our
error
handling
in
node
doesn't
work
either,
and
it
means
that
the
browser
has
never
worked
if
you
ever
used
fetch
or
any
kind
of
streaming
API.
So
I
took
a
look
at
the
HTTP
API
and
say
a
long
time
ago,
when
I
first
start
working
up
here,
I
looked
at
it.
I
was
like.
Why
is
it
inside
messing
around
with
a
wrapper
for
it
that
would
kind
of
make
it
rusty
so
that
you
can
use
tools
to
do
things
like?
H
Have
you
know,
documentation
generated
from
it
and
have
tools
that
you
can
then
use
to
ping?
The
API
and
all
this
kind
of
stuff
generate
clients
for
accepting
such
a,
never
really
went
anywhere,
but
I
did
do
a
lot
of
thinking
about
how
it
could
all
fit
together,
and
so
that
has
kind
of
informed
the
design
I'm
proposing
in
this
PR.
Also
the
up
that
stuff.
H
There
Alan's
been
working
on
and
and
his
unified
files
API,
which
is
kind
of
we've,
been
talking
about
for
a
while
and
kind
of
taken
some
of
those
ideas
forward
and
putting
them
into
the
HTTP
API,
which
means
we'll
have
a
smaller
surface
area.
There'll,
be
like
more
like
the
interruptions
that
you
have
where
there
will
be
simple,
which
will
shift
some
the
complexity
onto
the
client,
but
that's
kind
of
necessary.
H
If
you
don't
want
to
have
to
do
lots
of
weird,
have
lots
of
extra
arguments
like
the
DHCP
API
is
kind
of
modeled
on
the
go
on
and
they
go
on
is
modeled
on
the
CLI,
and
so
you
end
up
having
lots
of
kind
of
behavior
that
you
might
not
necessary
want
to
expose
over
HTTP
things
like
like.
If
you
take
a
path,
there's
an
argument
to
import
it
recursively,
which
makes
no
sense
over
an
HTTP
API.
H
Just
keep
sending
me
files
like
do
the
recursive
pausing
on
the
client
and
just
keep
sending
request
stuff
like
that
they're
trying
to
remove
some
of
that
stuff.
It
should
be
simpler
to
to
implement
simpler,
to
understand
and
simple
to
maintain.
So
yeah
it's
in
a
pre
pre,
full
review,
pull
request.
State
comments.
Welcome,
please
do
get
involved.
H
A
H
A
Okay,
so
yeah,
please
do
get
involved
with
the
pull
requests
for
the
HTTP
API
and
give
it
a
once-over.
And
if
you
were
interested
in
in
that
sort
of
thing,
that
would
be
super
cool
because
it
would
be
nice
to
have
have
that
become
a
little
bit
more
saying
a
little
bit
more.
What
developers
expect
and
and
yeah
a
bit
better
at
dealing
with
errors
for
starters
and
yeah
that
would
be
very
cool
and
I
will
hopefully
submit
a
pull
request
to
the
for
what
I'm
working
on
the
Uniting
the
files
API
just.
D
A
You
a
kind
of
idea
this
has
been
asked
for
since
I
guess,
I
think
the
issue
the
first
issue
I
found
for
it
was
from
2016,
so
it's
been
in
the
pipeline
for
a
while,
so
yeah
we're
gonna,
make
it
better
and
hopefully
soon
but
yeah
cool,
and
so
next
up,
where
are
we?
We
I've
got
not
lot,
not
a
lot
of
questions.
Efforts,
blockers,
asses,
no
blockers
and
ask
there.
Is
anyone
got
anything
else?
A
D
C
C
C
Going
to
be
some
long
way
to
go,
a
team
that
has
a
handle
or
a
copy
of
the
of
this
like
in
memory
I
gave
us
were
a
reference
to
it
didn't
see.
Definitely
that
sounds
like
it
might
be.
What's
happening.
D
D
C
C
A
F
A
Nice
serving
the
PR
now
all
right,
cool
thanks
everyone
for
coming
along.
It's
been
really
fun
this
week
as
always,
and
have
a
lovely
week
this
week
and
I'll
see
you
next
time
from
some
more
core
implementations,
action
and
and
it'll
be
even
more
exciting,
and
it
has
been
this
today
so
bye.
Everyone
see
you
next
week.