►
A
Hello:
everyone
welcome
to
the
ipfs
core
implementations
weekly
sync
for
monday,
then
the
16th
of
november
2020.,
I
am
making
brain.
I
will
be
your
host
we're
going
to
play
the
game
of
high
priority
initiatives,
other
initiatives,
q,
a
parking,
lot
questions
all
that
kind
of
stuff.
It's
going
to
be
great.
A
So
without
further
ado,
let's
talk
about
a
high
priority
initiative,
so
the
first
thing
to
talk
about
is
upcoming
ship
releases
they
ship
some
jsopfs
stuff.
So
we
shipped
a
new
http
client
with
a
long-awaited
patch
from
iraqi
to
report.
Progress
uploads,
which
is
great.
A
It's
gonna,
make
the
web
ui
well
good,
so
that
is
out
now
hooray,
also
some
bugs
with
the
message
for
client
around
dependencies
and
hopefully
that
shouldn't
happen
again
in
the
future,
because
we've
shored
up
our
ci
around
that
that's
from
js
life,
I
see
there's
a
go,
live
p2p.
B
C
Yes,
the
streams
they've
been
refactored
yeah
people
have
been
running
into
foot
guns
every
so
often
around
the
fact
that
when
you
close
a
stream,
it
was
only
closing
it
for
writing,
but
not
for
reading
close
now
closed
for
writing
and
reading,
which
is
a
breaking
change.
We've
also
removed
the
helper
functions
that
sort
of
did
a
background
task
that
waited
for
the
stream
to
have
an
eof
sent
to
you,
because
we
don't
need
that
anymore,
because
close
will
close
the
read
and
write.
C
D
Yes,
so
pinning
services
the
last
week
we're
finishing
up
some
migration
work
and
throughout
the
weekend
there's
been
some
migration
pains
that
need
to
be
resolved.
I
think
we're
very
close,
but
I'm
going
to
see
if
the
dean
can
work
with
me
to
resolve
those
last
couple
issues
and
it
has
to
do
or
this
point
with
using
a
new
versus
old
version
of
go
ipfs,
that's
linked
into
a
migration
code,
and
we
might
need
to
change
how
we're
we're
doing
a
couple
things
to
get
that
that
working
properly.
D
Make
a
few
changes
to
how
we
do
migrations,
but
we
should
be
able
to
get
that
done
today,
other
than
that,
the
pinning
services,
nothing
there's
nothing
new,
nothing
that,
since
last
week,
since
we
talked,
I
know,
there's
been
some
activity
asking
for
these
things
on
some
of
the
issues
and
they're.
Looking
at
waiting
for
this
in
the
next
release,
I
haven't
responded
to
any
of
the
issues
comments,
but
hopefully
we
can
get
this
in.
That's
all
we
have
left
is
to
make
sure
the
migrations
are
good.
C
Yes
and
related
to
that
a
once,
the
release
is
out.
The
the
next
high
priority
item
will
likely
be
making
fs
repo
migrations
like
function.
We've
basically
been
skating
by
since
we
switched
away
from
gx
to
go
mod.
We
haven't
had
to
do
anything
that
seriously
required
any
dependencies,
because
it's
just
been
editing
a
config
file,
and
now
we
finally
did
a
thing
that
required
actually
having
some
understanding
of
the
code
and
bringing
in
dependencies
and
everything
kind
of
blew
up.
C
So
this
is
we'll
call
it
paying
off
technical
debt
from
the
gx
gomod
switch.
D
In
retrospect,
it
may
have
actually
been
easier
to
do
the
work
moving
forward
this
weekend,
as
opposed
to
trying
to
get
it
to
work
the
old
way,
but
I'll
need
a
second
set
of
is.
We
can
evaluate
where
we
are
right
now,
so
we'll
have
an
answer
by
the
end
of
the
day,
either
it's
going
to
say
it's
where
a
thumbs
up
or
we
know
exactly
what
we
need
to
do
to
fix
it.
C
And
then
for
the
the
rest
of
the
pinning
services
stuff
lidl
has
has
done
a
bunch
of
work
with
with
petar
to
kind
of
move
this
closer
to
the
to
the
finish
line
and
actually
get
the
the
cli
and
http
api
that
we
are
going
to
be
stuck
with
for
a
while
ironed
out.
E
I
think,
like
the
main
question
remaining
question,
is:
should
we
block
by
default
or
not
over
the
weekend,
I
leaned
toward
blocking
by
default
like
andrew
convinced
me
why
that's
probably
a
better
default,
so
probably
we'll
experiment
with
that
this
week.
We,
if
we
do
that,
we
will
need
to
add
a
flag
which,
for
for
reporting
progress
similar
to
the
old
pin,
add
command.
When
you
pin
you
can
pass
progress
which
prints
you,
how
many
like
notes
you've
been
locally.
That
makes
sense
locally.
C
Just
for
for
added
context
for
those
not
following
super
closely
when
you
currently,
you
know,
when
you
do
ipves
pin,
add
a
thing
it
blocks
until
you've
downloaded
the
whole
dag.
That's
not
gonna,
that's
not!
How
sort
of
the
remote
pinning
is
gonna
work
you're,
not
sort
of
waiting
on
this
http
connection
until
the
thing
is
finished
being
downloaded
by
the
service.
You
just
tell
them,
please
get
it
for
me
and
then
they
go
get
it.
C
So
the
question
is,
you
might
want
to
pull
until
they're
done
so
that
you
know
when
it's
done,
but
should
that
be
the
default
behavior
or
not,
and
then
how
we
map
that
into
the
fact
that
we're
going
to
be
pulling
the
same
api
to
be
sort
of
the
new
local,
pinning
api
and
we'd
like
those
things
to
look
the
same,
because
it
would
be
weird
if
you
were,
if,
if
ibms
pin,
you
know,
pin
remote
add
was
not
blocking,
but
if
this
pin
local
ad
was
blocking,
that
would
be
pretty
confusing
for
people.
C
I
guess
I'll
just
do
the
next
one,
which
is
secio
removal,
yeah.
I
I
pinged,
I
ping
michael.
I
think
today's
the
day
for
turning
off
secio
for
a
little
bit
and
seeing
if
everyone
freaks
out
or
if
nothing
changes,
hopefully
the.
C
F
Yes,
so
the
custom
announced
multi-other
function
that
I
had
previously
done
is
now
merged
after
the
review.
I
also
got
reviews
for
the
auto
relay
example
in
the
dial.
Multi-Other
custom
start
function
last
week
that
I
will
already
also
address,
and
I
think
they
should
be
ready
to
go
now
as
jacob
only
point
out
now,
some
minor
things
left
then
I
I
worked
on
the
discover
and
connect
to
the
closest
peers
feature.
F
I
think
it's
also
10
and
it
will
neither
the
review
and
yeah
so
far.
This
week
my
main
focus
will
be
on
the
rendezvous
work.
F
A
Noise
next
up
is
bi-directional
streaming
and
streaming
errors
in
the
browser.
So
I
added
tests
that
prove
that
the
streaming
is
bi-directional
and
the
errors
work,
which
is,
which
is
great.
It's
a
massive
improvement,
I'm
gonna,
try
and
like
so.
A
The
only
thing,
that's
the
only
thing
that's
using
this
at
the
moment
is
add,
and
I
would
like
to
get
a
like
an
experimental
release
out
by
the
end
of
next
week,
so
people
can
start
playing
with
it
and
we
can
gather
some
feedback
and
that
kind
of
thing,
so
I'm
gonna
just
tidy
up
around
the
documentation
and
like
making
sure
we're
using
the
right
library
and
and
that
kind
of
thing,
because
it's
a
bit
weird
at
the
moment
using
different
grpc
libraries
on
the
server
and
the
client,
which
is
great,
interrupt
right.
A
But
it's
just
yeah
like
different
paradigms
either,
and
so
that
needs
to
be
looked
at.
Yeah
good,
good.
A
G
Yeah
thanks
alex
for
shipping
the
progressing.
It's
finally
landed
new
hooray,
so
now
we
can
integrate
into
the
web
ui.
It
should
be
quick,
I'm
hoping
by
next
week.
We
should
have
it
in.
A
Sweet
you're
up
again
next
for
typescript
integration
for
geoscience.
Those
updates
were
not
from
me.
So,
oh
maybe
somebody
else
did
it.
A
new,
a
new
typescript
challenger
has
entered
the
ring.
H
Okay,
so
yeah
a
pull
request
is
up
on
the
ag,
repo,
basically
to
add,
support
for
type
checking,
not
so
much
as
the
full
typescript
environment.
We
have
another
like
global
flag
just
to
enable
typescript
support
in
other
places.
This
is
more
for
or
more
specific
for
type
checking,
generating
documentation,
documentation
from
the
the
type
declarations
that
that
are
generated
and
all
those
use
cases
around
it.
So
yeah
please
review
and
give
some
feedback,
I'm
testing
it
on
the
js
ipfs
repo
everything
seems
to
work.
H
H
A
Amazing
very
exciting
next
up
is
use
of
shared
node
from
service
worker.
G
A
Oh,
I
will
get
to
that.
I
think
it
was.
It
was
like
pretty
much
all
there
so
that
will
land
imminently
next
stops
budget
two
support.
D
Nothing
to
report
it's
just
waiting
to
see
if
they
are
going
to
do
a
release
so
just
waiting
on
badger.
That's
all.
C
Don't
call
us
we'll
call
you
when
there's
they
having
a
decision.
I
guess.
B
Yeah
so
this
week
and
for
the
last
week,
arches
is
working
on
the
easily
stun
hardening.
So
there's
some
incongruities
that
we've
noticed
between
autonat
and
kind
of
observed,
addresses
being
shared.
So
right
now
he's
working
on
improving
that
so
that
we're
advertising
the
correct
public
addresses,
because
we
will
need
that
to
leverage
autonet
for
for
sun
coordination.
A
C
No,
I
think
the
contributor
is
still
still
doing
some
stuff.
I
I
need
to
give
some
reviews.
D
All
right
after
testing
at
scale
last
week,
I
needed
to
make
some
change
to
the
design
those
were
supposed
to
be
out
on
friday,
but
those
got
pushed
back
due
to
the
work
I'm
doing
with
the
migrations
for
the
pinning.
A
A
C
I
guess
I'm
just
I'm
just
like
looking
over
some
stuff
that
andrew
sent
me
over
the
weekend.
This
this
refund
migration
stuff
could
be,
could
be
pretty
gross.
I
suspect,
we're
gonna.
We
can
make
some
some
solid
progress,
but
I
there
may
be
some
hairy
things
that
pop
up
here
due
to
us.
I
think
a
number
of
changes
in
the
migration
process
that
we
haven't
done
before
we
support
both
like
level
db
and
badger
by
by
default.
C
Now
right,
there's
all
sorts
of
plug-ins
for
the
data
store,
we're
using
go
mod
like
there's,
there's
lots
of
new
things.
I
think
that
are
coming
up
in
this
migration
that
we
just
haven't
had
to
deal
with
before.
But
now
is
the
time,
so
I
guess
just
putting
that
out
there
as
a
heads
up
this
may
take
longer
than
than
expected.
You.
C
D
A
The
js
replay
migration
tool
is
similarly
hairy.
Not
wanting
to
do
things
like
create
a
whole
copy
of
a
repo,
because
it
could
be
enormous,
and
what
do
you
do
if
you
fail
halfway
through
like
moving
backwards
and
forwards
between
the
things
and
just
keep
running
it
until
you've
got
nothing
else
to
migrate?
You're
like
oh,
I'm
done
what
fun.
H
Anyway,
I
got
a
question
so
regarding
the
pinning
services
or
the
http
client
support
for
pinning
services
who's
gonna
take
that
I
think
gonzalez
started
working
on
it,
but
then
me
jacob
and
alex
talked
about
me
taking
on
that.
So
what
should
be.
G
Yes,
so
I
don't
know
I
I
started
looking
at
it
on
friday,
if
you
want
to
take
it,
I'm
happy
to
let
you,
but
I
can
also
work
on
it.
G
I
think
there's
also
a
few
things
that
could
be
paralyzed
if
you
want
to
go
that
way,
given
that
it
will
like
so
maybe
I'll
just
lead
to
the
next
item
that
I
had
to
the
parking
lot
because
they're
pretty
much
the
same,
so
we
will
have
to
figure
out
the
way
how
we
can
land
all
the
pieces
and
given
that
we
seem
to
want
to
land
it
in
one
lump
so
problem
is
that
to
test
it
we
need
some
sort
of.
G
We
need
to
go
ipfs
to
test
against,
and
then
we
also
need
a
some
pinning
service
to
run
it
again
to
like
pair
it
with
and
test
it
with.
There's
a
ruby
implementation,
but
it
seems
like
will
be
very
complicated
to
try
to
get
the
running
all
on
ci.
G
So
I
think
a
better
approach
would
be
to
write
a
simple
ser
like
a
echo
server
that
we
already
have
or
something
along
those
lines
to
do
it
yeah
and
then
trying
to
land
go
ipfs
and
also
lend
the
client
and
also
run
blend
the
tests
and
all
do
it
together
seems
like
I.
I
don't
really
know
how
we
can
pull
it
off,
including
web
ui.
G
I
made
this
piece
so
I
think
maybe
we
should
have
a
conversation,
how
we
can
do
gradually-
and
maybe
I
don't
know
the
things
me
and
lyle
talked
about.
Maybe
we
can
have
a
flag
in
http
clients
that
you
can
opt
into
and
we
can
land
certain
patches
that
are
not
tested
yet,
so
we
can
ship
the
go
ipfs
with
the
web.
Ui
that
uses
it
and
then
once
it's
in
then
we
can
add
the
tests
to
test
again
that
I
know.
Maybe
there
are
some
other
options
that
we
can.
G
Maybe
we
can
have
a
like
next
release
of
http
client
and
then
only
ship
it
once
the
gold
ships
but
yeah.
I
don't
know,
let's
figure
out
how
to
do
it
and
in
terms
of
parallelization
we
can
split
it
up
into
like.
As
I
mentioned,
we
need
to
do
the
server
into
the
other
pieces.
Maybe
we
can
do
it
together,
so
we
don't
have
to
wait
on
each
other
or
like
not
one
person
has
to
do
it
seriously.
G
H
Cool,
I
reckon
one
question
little
can't
we
generate
a
mock
server
from
the
spec.
E
E
We
we
generated
both
server
and
client
for
go,
I
believe
as
a
test,
and
we
actually
use
the
client
in
go
ipfs.
So
generation
works
pretty
well
with
open
api
yeah
so
and
and
if
we
have
that
lighter
version,
then
because
for
the
context
right
now,
we
got
this
ruby
app
created
by
andrew
nesbit.
E
But
that's
like
postgres
red
is
ruby
on
rails.
It's
not
something
you
want
to
run
on
the
ci.
Instead,
we,
if
we
have
this
lighter
echo
for
echo
like
service,
we
could
even
switch
go
ipfs
to
that,
and
that
would
solve
a
bunch
of
problems.
We
got
there
as
well.
A
C
H
C
H
C
H
G
Yeah,
so
do
we
need
another
session,
or
do
we
need
to?
Can
we
figure
out?
How
are
we
going
to
ship
all
this.
G
A
A
Okay,
great
thanks
for
coming
everyone
that
has
been
the
ipfs
core
implementations
weekly
sync
from
monday,
the
16th
of
november.
Please
remember
to
fill
in
your
async
updates.
People
do
read
them
and
it's
very
exciting
to
know
what's
happening
and
what
everyone's
working
on.
Thank
you
very
much.
Everyone
we'll
see
you
next
week,
bye.