►
A
Hi,
everyone
welcome
to
the
ipfs
score
orientations
weekly
sync
for
monday,
the
9th
of
november
2020..
I
am
a
king
brain.
I
will
be
your
host
we're
going
to
play
the
game
of
high
priority
initiatives.
Other
initiatives
parking
lot
q,
a
questions,
it's
going
gonna,
be
amazing.
We're
gonna
have
a
lovely
time.
So,
let's
start
with
the
high
priority
initiatives.
The
first
thing
is
upcoming
and
ship
releases
hot
off
the
press.
I
see
there's
a
js
p2p
who
can.
B
C
There
there
will
also
be
a
go.
The
p2p
release
upcoming
that
takes
into
account
the
golo
p2p
core
changes
to
the
stream
interface,
so
for
all
of
you,
who've
been
wondering
why
go
get
causes,
compile
errors
because
goget
just
grabs
the
latest
version
of
everything
with
no
carrying
at
all
for
versioning.
A
Amazing
also
shipped
about
20
minutes
ago,
js
ipfs
52
went
out
the
door,
so
that
would
have
taken
the
p2p
dna
resolution
stuff
with
it,
which
is
great,
so
that's
in
the
hands
of
everyone
right
now.
The
notable
things
in
that
is,
we
stripped
out
the
like
anything
that
isn't
any
ipod
codex
to
on
taxi
board,
doug,
pb
or
or
raw,
because
they're
not
very
commonly
used.
A
So
that's
things
like
bitcoin
and
zed
cache,
and
that
kind
of
thing
we
bundled
all
those
with
like
it
was
always
always
bundled
with
the
node
installed
like
it
wasn't
in
the
browser
bundle.
So
it's
not
like
the
browser.
Bundles
go
any
smaller,
but
things
like
ipvs
core
now
are
smaller
and
quicker
to
install.
So
you
know
everyone's
ci
will
thank
us.
A
B
C
I
hope
so
everything
is
in
its
like
final
stages
of
of
review.
I
think
the
pinning
services
stuff
is
like
basically
there
I
might
want
to
go
over
to
go
over
a
little
bit.
More
may
tag
lydle
to
check
a
couple
things,
but
I
think
we're
basically
done
last.
I
checked
it
was
just
like
adding
some
tests
and
andrew's
stuff
is
also
good.
D
E
C
D
F
F
F
This
was
basically
because
lib2
websockets
was
allowing
tcp
and
dns
addresses
both
suite
websockets
are
websockets
secure
to
be
dialed.
This
was
causing
the
famous
browser
ever
attempted
to
connect
to
the
insecure
websocket
endpoint
that
people
add
complaint
for
a
while.
So
basically,
we
when
we
are
changing
the
default
to
only
allow
to
dial
dns
websockets
secure,
addresses
in
the
browser
and
dns
websocket,
secure
our
simpler
socket
in
node,
but
also
supporting
a
custom
function
to
override
this,
because
we
need,
for
example,
for
local
testing
purposes
or
other
scenarios
that
people
need.
F
Basically,
we
want
to
find
and
connect
to
the
closest
peers
that
we
know
over
time
in
order
to
exchange
our
signed
peer
records
and
make
our
visibility
in
the
network
better,
and
with
this
I
will
also
work
on
the
production
guides
for
the
relay
in
the
rendezvous,
as
we
want
to
have
good
documentation
on
how
people
should
set
up
their
info
before
launching
this
on
the
next
leap,
to
release
and
yeah.
That's
it
for
this.
A
A
A
The
only
thing
to
do
right
now
is
bi-directional:
grpc
methods,
so
there's
no
like
unary
or
anything
like
that,
because
it's
just
going
to
fall
back
to
the
http
client
for
the
time
being,
for
everything
that
is
in
the
streaming
response,
which
would
be
bad.
So
work
is
ongoing
on
that
I've
heard
draft
pr.
G
G
G
G
It
removes
some
of
the
type
magics
that
we
had
in
place
to
try
to
work
around
the
api
manager,
swapping
apis
by
removing
api
manager
and
adding
a
bunch
of
other
types.
I
also
kind
of
spent
some
time
looking
into
the
issues
that
turns
out.
Typescript
has
so
it
does
omit
type
parameters
in
generated
types
which
is
not
great
because
and
it
turns
them
into
any
or
it
would
complain
when
people
are
when
people
have
a
structure
of
a
setting
set
in
their
projects,
because
emitted
types
are
not
technically
valid.
Typescript.
G
So
I
don't
know,
a
team
at
microsoft
seems
to
recognize
that
it
is
an
issue,
but
it's
gonna
be
a
while
before
it's
fixed,
and
I
was
trying
to
see
if
I
can
fix
it.
So
we
don't
have
to
wait
it
too
much,
but
it's
not
non-trivial.
So
there's
another
issue
that
turns
out
kind
of
intertwined
with
this.
So
typescript
has
a
notion
of
interfaces
and
types
and
you
can
import
interfaces
in
typescript
syntax,
but
you
can
only
import
types
in
jazz,
doc,
syntax,
but
then
you
can't
implement
type.
G
You
can
implement
interface
anyway,
yeah
sorry
for
bubbling,
but
there's
a
bunch
of
things
that
still
need
to
be
worked
out
worked
out.
I
also
put
another
item
which
was
last
couple
of
weeks.
People
had
been
expressing
interest
in
shared
worker
stuff
in
within
the
service
worker,
so
that
prompted
me
to
do
some
work
there.
I
wrote
an
example
that
is
a
pr
now
to
a
js
ipfs
that
tries
to
exercise
all
of
those
things
and
show
how
they
can
be
wired.
G
There
are
some
interesting
findings
there,
two
that
I
hope
to
kind
of
write
about
and
including
their
blog
posts,
namely
there
are
a
bunch
of
limitations
with
how
shared
workers
work.
Sorry,
service
workers
work,
and
I
think
we
might
have
to
think
more
about
how
we
can
work
within
those
limitations
rather
than
trying
to
fight
them,
but
yeah.
I
think
I'll
try
to
capture
it
somewhere
else
and
share.
A
C
By
the
way,
I
guess
I
guess,
there's
a
semi
announcement
thing.
We
used
to
have
a
a
frequently
skipped
over
unix
fsd
1.5
other
issue.
Here
someone
has
picked
it
up
again
and
so
I'll
do
some
some
reviews
and
and
see
how
that
works
and
if
we
can
can
finally
push
it
over
the
finish
line
and
go
that
was
picked
up
as
a
bounty
right.
A
I've
been
commenting
on
that
thread
just
to
make
sure
that
the
argument
types
are
the
same.
You
know
because
we
had
to.
It
was
tedious.
We
had
to
change
some
stuff
in
js
because
of
some
limitations
of
browser
apis
boring
anyway.
That's
cool
yeah.
I
stuck
a
note
into
the
document,
so
we
can
get
updates
on
every
week
now
until
it's
done
next
up
is
dns
adder
resolving
ngs,
which
we
kind
of
talked
about
a
little
bit.
F
B
B
We
can
have
that
conversation
more
this
week
to
really
really
button
that
down
yeah
so
for
anybody
who's
not
clear
what
that
is
it's
the
idea
that
we
just
do
hole
punching
strategy
through
relays,
and
if
that
fails,
you
continue
to
just
exchange
data
over
the
relay
versus
having
a
relay
coordinate
the
hole
punch
for
you,
so
that
you
don't
actually
run
any
data.
That's
not
related
to
the
hole
punch.
B
This
is
like
a
bandwidth
saving
thing
for
the
relay
and
having
hole
punching
relays
versus
like
data
exchange
relays
was
mostly
cost
effective
scenarios
there
pros
and
cons
to
both
strategies.
The
dean
has
mentioned
that
in
the
issue.
We
will
try
to
figure
out
that
soon,
which
one
we
want
to
do,
but
yeah.
If
you
have
opinions
scream
in
my
direction,.
D
Yes,
I
am
currently
working
on
on
that.
I
have
a
proposal
which
I'm
just
finishing
up
today
as
it
covers
what
we
want
to
do
in
terms
of
general
approach
and
some
of
the
algorithms
that
I'm
going
to
be
suggesting
that
we
we
look
into
employing
for
this,
and
what
I'm
doing
right
now
is
trying
to
gather
some
actual
statistics
so
that
we
can
have
some
real
numbers
to
look
at,
and
so
we
can
incorporate
that
into
the
proposal.
D
D
A
D
C
I
suspect
that
this
is
going
to
sort
of
keep
popping
up.
Routing
hints
is
the
idea
that
if
you
build
multiple
routing
systems
for
finding
content
that
it
would
be
nice
to
have
a
heads
up
as
to
which
system
to
look
for
it
doesn't
have
to
be
the
only
one
you
can
check
all
the
other
ones,
but
it's
a
hint
as
to
where
it
might
be.
C
So
you
can
choose
to
optimize
for
checking
there
first
or
checking
there
later,
as
a
fallback
sort
of
how
that
works
is
information
that
we
sort
of
know
out
of
band
for
some
data
that
we
might
want
to
have
some
some
way
of
of
conveying
it
yeah.
So
I
I
don't
think
this
is
gonna,
be
that
would
be
this
week.
But
perhaps
if
people
are
interested
in
brainstorming,
then
we
can
do
some
github
stuff
this
week
and
design
review
next
week.
B
Yeah
just
along
those
lines,
because
I
know
like
we're
in
november
and
december's
around
the
corner-
and
one
thing
I
want
to
be
cautious
of
is:
we
have
a
lot
of
things
in
progress
going
on
in
parallel
here.
I
know
we're
getting
a
lot
of
people
back
from
file
coin
efforts
and
wanting
to
ramp
up
on
things,
but
we
should
also
take
the
time
to
slow
things
down
on
some
of
those.
B
Let's
ship
go
ipfs.ed
and
whatever
things
that
we
have
in
the
pipeline
and
then
spend
a
little
bit
more
time
talking
through
some
of
these
efforts
and
prep
for
for
next
year,
so
that
when
we
get
even
more
people
back,
we
can
have
these
smaller,
concerted
efforts
to
push
that
stuff.
So
I
would
prefer
to
favor
everyone
favor
design
discussions
over
trying
to
push
too
many
things
in
parallel.
G
C
Yeah
yeah,
I
remember
the
sort
of
how
how
threads
how
threads
does
this?
This
is
yeah.
That's
like
sort
of
one
of
the
options
part
of
it
is
like
how
we,
how
we
embed,
how
we
make
this
accessible
to
users
right
so
threads
does
this
by
they
introduce
like
a
thread,
structure
or
protocol,
so
we
have
ipfs,
they
have
slash
thread
and
that
embeds
the
information
inside
of
it.
Maybe
that's
what
we
want.