►
A
A
Yeah
so
release
process,
yeah,
there's
a
ball
draft
assigned
to
the
blog
post.
Those
are
all
up:
hey
yeah,
the
about
post
should
hopefully
well
it'll,
be
out
in
2nd.
The
signs
were
out
and
the
dines
were
looking
pretty
good,
but
well
haven't
you
one
final
view
of
those.
So
today,
maybe
tomorrow
will
say:
yeah
JSF
s
is
your
point.
Three
seven
isn't
released
and
I
want
is
saying
that
he
will
be
the
release
process
for
@jseifer
fest
to
manifest
your
role
in
all
the
list
should
okay,
there
are
no
updates
to
the
testing
info.
A
A
Here
is
the
design
doc.
Is
this
just
his
proposal
for
like
how
happy
should
fit
together?
Ideally,
we
can
reuse
parts
of
existing
systems
just
make
this
work,
but
he
strike
trying
to
come
up
with
a
very
generic
system
that
can
kind
of
like
supersede
us
where
you
just
like
plugging
in
custom,
doctor
I'm,
just
plugging
a
custom
test,
parcels
kind
of
stuff,
I.
A
A
C
A
A
C
So
jisub
PDP
consolidation
is
almost
done.
There
are
a
couple
modules
that
I
probably
won't
include
in
this
round
is
like
the
NAP
manager
and
keychain
one,
because
nap
manager
isn't
actually
done
yet
and
then
keychain
is
actually
not
currently
in
the
p2p,
it's
kind
of
standalone
by
itself,
and
it's
only
being
used
right
now
in
JSI
BFS.
So
I'll
probably
wait
to
do
that
as
a
separate
one,
so
that
because
that
will
cause
some
API
chains
yeah,
but
I
should
be
done
with
that
this
week
and.
C
Yep
so
going
through
an
example
of
distributed
signaling
over
a
relay
as
I've
been
going
through,
that
I've
been
finding
miscellaneous
issues
like
J
s,
multi
adder
wasn't
handling
circuit
relay
peered
when
you're
announcing
you're
really
on
there.
So
just
fixing
stuff,
as
going
through
with
that
one
thing
that
I'm
not
tackling
right
now
is
distributed
ice,
but
we're
probably
going
to
need
to
take
care
of
that
at
some
point,
because
otherwise,
if
we
just
use
like
the
default
Google
I
servers
that
becomes
very
easy
to
block
in
places
like
China
ice.
A
Is
discovering
your
own
external
dress
right?
What's
that
I
use
discovering
your
external
address?
Yes,
yes,
so
we
have
that
with
AlterNet
right,
which
I
think
may
be
sufficient
in
many
cases
for
like
you,
just
as
you
can
to
be
for
servers,
various
places
like
you
can
just
ask
them
hey.
What
is
my
external
address
and
it'll
start
dialing
back
to
you,
so
I
think
we
may
be
covered
there.
It's
not
really
it's
distributed.
It's
like
we
still
like
you
have
to
find.
These
servers
can
be
tricky
but
yeah.
A
Another
thing
that
she
go
into
his
header
is:
these
Oh
has
had
this
PR
up
or
spec
TR
about
yeah,
please
this
is
effectively
I
know
what
does
it
call
I
guess
stun
gave
the
right
term
over
relay
we're
lucky
yeah,
it's
trying
to
figure
out
her.
Basically,
you
can
X
one
of
a
relight
and
then
they
tell
you
like,
or
than
you
coordinating
miss
connect
to
try
to
use
like
I.
Don't
they
said
a
bit
or
I
get
these
out
so
he's
open
to
bypass
routers.
So.
A
You
should
read
that
anyone
was
talking,
could
read
that
and
like
try
to
it's
all
they
they
concern
there
is
that
it
does
require
establishing
a
phone
connection.
First,
it
shouldn't
be
that
expensive.
If
we're
already
using
relays,
maybe
phone
with
our
current
relay
system,
is
it
doesn't
scale?
Well,
then
we
yeah.
C
A
I
think
to
make
this
fully
useful
I
think
they
may
have
to
have
a
packet
switching
system
somewhere,
because
relays
have
this
problem
with,
like
he'd
want
to
kick
to
one
node
on
every
single
relay.
You
have
to
collection,
every
single
relay,
which
means
everyone
needs
attention
to
every
way
which
just
doesn't
work,
but
we
could
do
the
section
about
back
as
much.
You
might
just
like
make
it.
So
you
connect
one
relay
another,
be
let's
all
connect
to
each
other.
So
then
you
have
to
hop
forwards,
but
we
don't
currently
do
this.
A
We
slightly
easier
for
just
had
like
backups
from
you.
Basically,
you
think
your
outbound
relight,
you
couldn't
do
it.
You
said
now
all
your
days
that
rely
and
then
that
relay
routes
it
to
the
correct
on
the
relay
and
then
it's
basically,
everyone
makes
it
like
gateway,
node
and
I
think
it
may
be
hard
as
you
work
like
this
would
require
fair
amount
of
like
thinking
and
work
to
make
it
work,
yeah,
okay,
so
moving
on
all
right,
eNOS
didn't.
D
Yeah,
so
one
of
the
pair's
got
merged
huzzah,
which
means
there.
We
are
one
out
of
two
PRS,
because
I've
separated
out
some
stuff
to
make
it
so
that
we
can
actually
like
use
this.
You
know
use
this
and
have
it
you
know
be
functional
that
one
is
following
some
designer
boots
from
last
week.
Thank
everybody
for
helping
with
that.
Almost
it
is
it
the
end
of
peers
insight.
A
couple
of
design
issues
came
up
related
to
the
lib
p2p
record,
validator
interface,
that
you
should
take
a
look
at
if
you're
interested.
D
Basically,
we
can't
figure
out
if
things
are
equal
and
we
only
seem
to
care
about
this
scenario
of
picking
the
best
record
when
there
are
many
ways
where
you
could
combine
records
and
there's
this
issue
on
like
having
a
separate
name,
sis
that
I
know
some
folks
were
interested
in
I'm,
not
sure
if
they're
on
the
call
or
not.
But
if
you
care
please
comment
so
that
I
can
do
what
you
want
and
that's
all
for
me.
A
A
A
A
This
is
beginning,
well
attraction,
and
this
point
where
Hayes
can
start
work.
So
if
you
have
comments
or
ideas,
but
it's
well,
that's
their
other
update
is
I
had
to
go
talk
with
Alex
this
morning
today
about
a
performance,
so
I
guess
well.
This
is
more
package
matters
update
the
really
effective
measure
stuff
date,
but
everything's
good.
Actually,
okay,.
B
Thing
I
someone's
added
this
here
and
rightfully
so
it's
a
ongoing
long-running
refactor
to
switch
to
using
a
promise
based
API
is
for
other
throughout
the
codebase
instead
of
callbacks
it
is.
It
involves
ipfs
the
p2p
I
kod
multi
formats.
There
are
67
repos
that
need
to
be
refactored
like
link
to
the
issue
there.
There
are
lots
of
good
reasons
for
doing
this.
There
are
justifications
in
that
issue
like
things
like,
like
smaller
bundle,
sides,
better
staff,
traces,
yeah
stuff,
like
that,
you
know
just
how
working
with
promises.
B
It
is
kind
of
a
more
modern
way
of
working
in
JavaScript,
and
hopefully
we
might
be
able
to
gain
more
people
who
are
interested
in
working
with
ipfs.
If
we
don't
present
a
like
a
an
API,
that's
seen
as
kind
of
out
of
date
in
any
way
so
like
it's,
there
are,
there
are
reasons,
go
and
read
them,
and
it's
been
going
on
for
a
long
time.
There
are
a
lot
of
repos
and
we
we
have
not
resourced
people
to
solely
work
on
this
during
their
quarters.
B
We've
currently
done
around
64
percent
of
what
we
need
to
do,
and
we
are
the
kind
of
notable
exception.
Is
lip
e2p,
stuff,
I
think
that
it's
mean
there
are
many
many
live
feeds
of
the
repos,
and
only
two
people
working
full-time
on
it
and
not
working
full
time
on
this
particular
endeavor
and
the
process
of
refactoring
them
to
be
async/await
based
instead
of
callbacks
is
more
tricky
because
we're
not
just
refactoring
async/await,
it's
just
it's
just
it's
using
async
iterators
as
well,
instead
of
a
regular
full
stream.
B
So
we
need
to
get
that
motion
and
done
next
as
a
priority.
But
we
did
it's
worth
mentioning
that
we
did
do
an
exercise
of
prioritizing
these
repos.
In
that
issue,
we
have
weave
p0
the
ones
that
basically
have
no
dependencies
on
other
modules,
and
then
we
gave
the
others
p1
p1
p2
p3,
and
when
we
should,
when,
when
abouts,
we
should
be
be
refactoring
them
it's
tricky,
because
every
every
refactor
is
a
breaking
change.
C
E
Sorry
can
I
make
a
follow-up
question
on
the
previous
topic.
Yes,
I'm
curious.
If
putting
people
like
something
that
was
discussed
over
a
thread
last
week,
was
you
know
what
would
it
look
like
if
we
did
a
sprint
or
a
spike
on
this,
or
had
a
couple
of
people
like
dedicate
full
time
to
this
in
q4?
Would
that,
like
significantly
improve
and
help
us
be
able
to
knock
this
thing
out
instead
of
having
it
continue
on?
E
For
you
know,
at
pace
like
we've
been
working
on
it
for
like
what
you
know
six
six
months
and
we're
64
percent,
so
presumably
we'd
have
like
another
six
months
to
be
100%
like
what
is
our
pathway
to
that
is
bringing
on
someone
or
helping
people
externally
in
the
who
are
not
already
contributing
those
to
pitch
in
on
this
sort
of
stuff
feasible.
What's
our
path
forward.
B
Yeah,
so
it
was
suggested
that
we
could
do
a
sprint
where
the
Jas
core
team
worked
solely
on
this,
like
I
find
it
because
it
was
started
so
long
ago.
Now
it's
quite
hard
to
kind
of
justify
that
the
package
managers
goal
other
than
this
will
make
Jess
I,
give
s
significantly
easier
to
debug,
significantly
easier
to
maintain
and
work
with
in
the
long
run.
But
it's
just
something
that
needs
to
be
used
to
be
done
so
yeah.
B
We
could
get
an
outside
agency
to
do
it
too,
but
again
it
would
be
only
beyond
a
certain
amount,
and
we
I
think
that
you
know
the
ones
that
were
very
easy
have
been
largely
done.
I
need
to
double
check
on
that,
but
yeah
I'm,
just
a
little
bit,
hesitant
I,
guess
I'm,
not
I'm,
not
completely
averse
to
it.
B
Yeah
like
I
would
love
to
just
pull
everyone
in
a
J's
core
team
and
just
say
work
on
this
till
it's
done,
but
I
know
that
the
there
are
the
things
we
want
to
achieve
this
cause
and
that's
largely
why
it
hasn't
been.
You
know,
done
really
really
quickly,
because
it
has
always
been
a
when
you
get
time
to
do
this.
B
C
We
can
start
actually
outsourcing
that
of
like,
if
that
would
be
easier
for
people
in
the
community
to
be
able
to
say
hey.
This
is
how
things
used
to
be
done
with
pull
streams.
This
is
how
you
do
that
same
thing
with
the
async
iterators
and
now
that
that
makes
it
easier
to
do
even
in
spare
time
or
whatever
it's
just
having
those
those
examples,
but
getting
the
dedicated
time
to
do
those.
Those
four
would
be
good.
C
B
I
think
so
we
have
an
implementation
I've
spent
time
this
week.
Sorry
last
week,
working
on
multi
stream
select
and
I
I.
In
my
opinion,
it's
about
90%
done.
We
have
two
PRS
for
WebSockets
and
TCP
transports
open,
so
they
are
just
in
the
process
of
being
reviewed.
So
like
we're
close
with
that
goal,
yes,.
D
A
D
A
A
Unfortunately,
we
have
an
HTTP
link,
so
we
won't
get
up
to
point
to
need
some
HTM
points
at
a
Houston
point
and
in
many
cases
we
actually
means
like
include
the
domain,
the
reason
of
at
endpoint.
So
there's
a
good
in
this.
We
do
say
like
sausage
to
keep
SAS
domain
name,
slash
path,
because
almost
all
HTM
points
actually
like
they
expect
you
to
pass
in
a
domain
name,
it's
required
in
aged
1.1
and
most
when
a
wouldn't
out,
because
it's
been
out
for.
A
Unfortunately,
this
HP's
currently
defined
as
taking
no
arguments,
and
we
already
have
use
cases
that
you
live
without
taking
any
arguments.
So,
for
example,
if
you
look
at
this
stuff,
we
have
such
HTTP
slash,
p2p
dec
web
CRT
best
direct,
so
you
need
to
figure
out
how
to
deal
with
this
I
am
going
to
try
to
phone
with
a
puzzled
by
Wednesday
or
whatever
you
know
madness,
meeting
but
yeah.
This
is
something
we
see
this
off.
A
We
have
other
random
questions
around
the
HTTP
HTTP
and
TLS,
where
ideally,
should
be
/c,
less
sh
t
as
you
compose
these
things.
Instead
of
having
a
custom,
each
yes
same
thing
for
WebSocket.
What's
outfit
secure,
we
need
something
to
upgrade
path,
a
lot
of
fun
stuff.
Basically,
we
need
to
have
a
conversation
about
this,
so
I
would
like
to
have
a
meeting
to
discuss
this
again,
probably
Wednesday.
A
B
I
have
a
PR
open
on
the
JSA
ipfs,
the
I/o
website
and
the
so
I
think
Moxie
are
not
working
specifically
on
that
website
anymore.
So
it's
up
to
us
to
review
and
and
merge
that
sort
of
thing.
It
was
more
involved
because
so
what
I
was
doing
was
updating
examples
on
the
website
to
use
the
new
constructor
for
Jess
IVFs
in
the
latest
version.
B
D
D
A
We
should
do
that
in
some
cases
like
it's
just
this
problem
half
ago.
Actually
like
there
really
two
things
we
use
interfaces
for,
we
use
them
for
contracts
when
it
published
we're
saying
like
we
expect
you
to
satisfy
this
interface,
it
actually
yes
like.
If
I
could
we
use
for
documentation,
anyways
and
that's
why
we
come
there
I,
we
probably
should
do
more
of
like
define
the
interface
where
we
consume
them,
but
in
some
cases
that
look
requires
like
redefine
the
interfaces
over
the
place.
A
That's
what
we
don't
it
I
think
it
really
just
depends
the
module
even
even
a
case.
We
also
just
use
interfaces
less
often.
Sometimes
we
do
that
because
we
think
you
know
what
eat
like
like.
Okay,
so
they
can
kill
you're,
not
supposed
to
return
interfaces.
You
must
return
the
concrete
type,
but
sometimes
we
just
use
interfaces
anyways,
because
we're
saying
look
like
we
may
want
to
change
the
concrete
type
in
future.
A
You
may
want
to
add
additional
concrete
types,
so
we're
just
going
to
turn
the
coffee
type
ever
just
confronted
interface
like
we
do
this
all
over
the
place
and
in
mostly
utter
and
stuff
like
that,
he
does
have
a
cost
like
if
I'm
type
cost
of
allocations,
so
need
to
be
more
careful
about
this.
But
that's
going
reasons
like
it
depends
on
the
case.
I
guess
so.
A
lot
of
salsa,
just
like
people,
came
into
work
about
this
experience,
but
the
languages
they
brought.
A
Those
ideas
instead
of
two
minutes,
go
away
well
like
if
you're
looking
lib
p2p
on
the
counter,
like
you,
have
the
host
interface,
all
kind
of
stuff
or
the
swarm.
The
connection
look
guys
stuff
like
that.
That's
that
that's
just
because,
like
we
want
to
abstract
of
this
we're
saying,
like
hey
user
use,
these
things
here
like
here
are
the
types
you
should
be
using
because,
like
you
can't
rely
on
the
concrete
information.
Basically,
that's.
D
What
we
do
that
they
just
like
as
an
example
of
some
of
the
issues
that
I
posted
about,
like
you,
know,
Golu
p2p,
record
types.
I,
don't
have
to
have
a
long
discussion
with
anyone
about
changing
the
record
type.
I
could
just
make
a
new
interface
and
the
library
it's
used
in
the
library.
The
thing
is
being
consumed
by
and
like
no
no
global
changes
to
look.
You
two
people
are
required
in.
A
Some
cases,
yes,
the
concern
here
is
like
it's
like.
If
you
basically
have
something
smokes,
you
can
form
some
interface,
then
you
have
to
make
sure
the
key
supports
the
interface.
The
other
issues
like
I
said
like
if
you
like,
if
you
change
one
thing
and
add
a
feature
to
it,
and
then
you
change
something
else
to
sailing
I
need
something.
Has
this
new
feature,
then
anything
else
that
might
know
that
feature
doesn't
doesn't
work.
This
is
a
big
one
of
the
prime
reasons.
A
We
use
interface,
a
force
like
basically
support
department,
Go
Go's
written
by
Google
for
Google
for
like,
like
a
single,
monitor
ethos,
and
it
works
really
well
in
our
situation,
you
have
maybe
those
many
different
people
working
on
these
things,
so
that
we
need
some
point
where
we
could
say,
like
you
know,
look
really
things
like.
That's
like
you
know
basis
for
it's
like
this
is
the
contract
we
are
going
to
satisfy.
A
So
like,
if
you
redefine
the
same
database,
it
up
a
bunch
of
different
actions
to
consume
something.
Is
it
a
finding
it
one
such
application
is
then
it's
totally
unclear
like
when
upwards
happened,
and
you
can
have
situations
like
things
just
don't
work
for
magical
reasons
that
make
sense.
So
talking
about
your.
D
A
So
if
you're
talking
with
a
valley
area,
basically
you
want
to
add
a
feature.
Validator
interface,
like
the
reason
we
haven't
faced
find
here,
is
it
then,
like
I
know
that
anywhere
I
can
find
a
validator
and
then
I
could
pass
it
off
dive.
You
guys
and
then
it
will
always
like
be
a
valid
validator
and
I
keep
looking
at
be
pheasant
peanut
base
itself
like
we
have.