►
A
B
B
C
Spec
itself
and
kicked
off
supporting
js
http
client,
which
will
be
using
new
pin
remote
commands
from
work
in
progress,
pull
request
in
go
ipfs.
I
think
we
need
to
like
wire
up
config
and
figure
it
out.
The
way
a
config
will
be
up
like
updated,
like
how
we
add
and
remove
pinning
services
if
it
will
be
possible
to
like
modify
like
specific
one
in
the
way.
That's
just
like
an
array
of
objects
in
the
config.
C
C
B
D
B
Yeah,
well,
it
just
means
like
if
I
rename
you
know
your
pinning
service
from
alex,
you
know
alex's
pinning
service
and
to
alex's
pinning
service
back
up,
and
then
I
make
a
new
one
and
call
it
alex's
pinning
service
right
like
I
can
like.
I
can
like
change
the
names
around
so
that
when
you're
sending
to
me
you're,
actually
sending
me
like
a
different
key
than
you
meant
to
so
we'll
just
have
to
make
that
not
a
thing
that
you
can
do.
That's
it's.
A
B
D
B
B
E
Yeah
I
so
the
after
relay
is
still
in
progress.
The
initial
implementation
got
merged
last
week
in
the
0.30
range,
which
is
a
branch
for
the
jsonptp
feature
release
the
milestones.
23
are
ready
for
reviewing
we'll
review
them
this
week.
They
include
the
network
update
to
inform
the
other
peers
using
an
identified
push
that
we
have
now
a
new
listening
address
that
people
could
use
to
reach
to
us,
and
I
also
needed
to
update
the
self-signed
peer
records
to
being
basically
recreated
with
the
new
address
and
the
thermal
sound.
E
Basically,
when
we
don't
have
in
enough
relays
to
listen
on
according
to
how
user
configured
the
liberty
we
can
try
to
find
them
on
the
on
the
network
using
by
now
only
b2b
content
routing.
So
in
ipfs
people
would
basically
use
the
delegates,
as
we
still
don't,
have
the
dhc
enabled
by
default.
But
I
plan
to
also
offer
the
next
week's
follow
up
this
with
rendezvous
integration.
E
It's
to
don't
do
don't
announce
the
private
addresses,
which
is
a
problem
that
we
already
have,
and
usually
we
have
people
creating
issues
saying
that
we
have
warnings,
for
example,
in
the
browser
saying
that
we
are
trying
to
dial
to
local
addresses
in
the
browser
which
it's
basically
because
other
peers
are
advertising
them,
as
listening
addresses
that
people
should
be
able
to
dial,
which
is
not
the
case.
So
basically,
we
will
need
to
know
that
we
have
the
auto
relay
to
get
rid
of
this
issue
yeah
and
finally,
regarding
the
connection
manager.
A
E
I'm
we
are
still
aligning
like
how
involved
she
has.
The
ptp
would
be
in
the
car
side
if
it
would
basically
be
intelligent
enough
to
just
protect
everything
or
if
we
should
guide
people
on
who
build
protocols,
and
now
they
should
protect
the
peers.
That's
some
of
the
things
that
we
need
to
polish.
While
we
implement
it.
I
think.
A
B
D
I
invited
several
people
here
to
it
on
friday,
but
if
you're
not
on
that,
let
me
know
and
I'll
add
you
to
like
a
beyond
bit
swap
research
discussion,
because
the
resident
lab
team
is
working
on
beyond
bit
swap
so
what
the
future
of
the
bit
swap
or
block
swapping
looks
like.
So
that's
on
friday
to
just
what's
the
state
of
that
research
where's
the
project
going
solicit
feedback,
so
this
could
be.
D
F
I
crossed
it
out
I'll
take
that
off
the
list,
because
I'm
not
spending
much
time
on
it
so
once
I'll
actually
start,
then
I'll
put
it
back
on
the
list.
The
next
one
is
also
mine.
Improving
web
ui
file,
let's
sing
so
a
lot
of
stuff
has
already
landed
last
remaining
piece
that
I'm
working
on
is
actually
doing
the
ui
piece.
So
it
will
let
you
know
how
it's
progressing.
F
F
F
F
It
just
generates
types
separate
from
the
tree
and
tells
typescript
where
to
look
for
them
and
then
there's
a
older
pull
request
from
hugo.
That
does
a
doc
generation
from
the
types
and
I
forgot
exactly
what
few
other
things
so
I'll
be
splitting
up
those
into
separate
tasks
and
lend
them
one
by
one.
F
Once
we
have
a
ipfs
type
generation
working
I'll,
probably
go
on
and
try
to
add
the
similar
mechanism
to
other
recalls
as
well
and
take
the
learnings
that
we
get
from
there
so
except
expect
pull
requests
there.
Part
of
the
reason
is
also
like
for
ipfs
to
generate
meaningful
types.
It
should
be
able
to
understand
the
dependencies
pulls
in
so
if
flippy
t2p
is
not
typed
and
none
of
the
ipld
is
typed
and
it
can't
really
make
much
sense
of
it.
Yeah
so
through
that.
F
A
A
The
idea
being
that
you
can
just
require
the
core.
You
don't
have
to
like
go
through
the
installation
of
all
the
dependencies
for
the
cli
and
the
http
api,
which
you
usually
don't
need,
because
that's
kind
of
something
that
people
complain
about
quite
a
lot
and
then
so.
That's
kind
of
why
I
haven't
got
to
the
touchscript
stuff
yet
because
I've
been
moving
on
so
far
as
well.
A
A
Another
test
against
guy
progress,
9.7,
which
entailed
bringing
the
icfs
npn
registry,
mirror
up
to
date
because
it
shares
a
whole
bunch
of
the
code
which
has
been
interesting
and
I
was
exposed
to
the
misty
v8
still
welcome
because
it
cools
through
into
the
noses
tp
module,
which
requires
either
buffers
or
strings
to
be
sent
you're,
not
using
aos,
which
is
mega
tedious
clearance.
Unfortunately,
that's
different.
A
B
B
I
don't
know
if
it's
going
to
be
this
this
week
or
next
week
would
be
better
but
discussing
how
we
want
the
pinning
to
look
in
light
of
some
of
the
changes
to
in
light
of
the
work
vital
has
done
on
the
pinning
services
api
in
particular
around,
like
you
know,
named
pins
or
having
multiple
pins
that
target
the
same
cid
and
like
what
might
be
an
appropriate
way
to
manage.
All
of
that,
I
don't
think
this
week
is
going
to
be
the
time
for
that,
but
maybe
next
week.
F
D
I
think
there's
like
there's
like
the
the
short-term
data
storage
thing
and
then
there's
like
a
long-term
just
moving
pinning
completely
over
to
threads,
but
I
think
that
is
going
to
require
a
much
much
larger
conversation.
But
is
there
something
that
we
can
do
in
that
interim?
That
gets
like
that
pinning
performance
improvement
for
the
the
pinning
services
so
that
they
can,
you
know,
just
improve
performance
on
their
systems
until
we
get
to
the
the
thread
thing
or
whatever
that
ends
up
being.
B
I
suspect
that
the
big
question
will
just
be
like
for
now
I'll
just
be
like
what
is
the
thing
to
be
the
primary
key?
What
is
the
correct
thing?
Is
it
like
a
cid
or
a
name
or
some
uuid,
and
then
once
you
have
that
figuring
out?
Okay,
what
do
you
want
to
do
in
the
cases
where
you
want
to
you
wish
the
primary
you
know
key
was
something
else.
B
F
G
Yeah
right,
so
I
have
some
code
sitting
around
that
looks
through
every
repo
in
a
a
user
or
org
and
looks
for
a
file
and
then
sees
if
it
matches
the
file
that
I
have
locally
and
then
replaces
it.
If
it
doesn't
using
my
github
token,
I
use
it
to
update
all
of
the
github
action
workflows
whenever
there's
a
change,
because
I
have
like
50
repos
that
all
have
the
same
workflow.