►
Description
Meeting Notes: https://github.com/ipfs/team-mgmt/pull/1119
About IPFS GUI and Browsers Weekly: https://github.com/ipfs/team-mgmt/issues/790
IPFS Mirror: https://ipfs.io/ipfs/bafybeigi4jfvlvylw3fdplq5gd4xtykwwyve62op2h5njxvj62efp7h3zm
For more information on IPFS
- visit the project website: https://ipfs.io
- or follow IPFS on Twitter: https://twitter.com/IPFS
Sign up to get IPFS news, including releases, ecosystem updates, and community announcements in your inbox, each Tuesday: http://eepurl.com/gL2Pi5
A
And
welcome
it's
26
of
February
2020.
This
is
three
and
in
web
browsers
and
connectivity.
Specialty
group
interest
group
weekly
call
with
assortment
of
people
and
topics
last
week.
This
call
did
not
happen
because
we've
been
on
location
as
a
team.
I
profess
team
week
happened
and
we
had
some
problem
solving
sessions
and
overall
small
hug
time.
I
want
to
quickly
go
over
two
topics
relevant
to
web
browsers
that
we've
talked
about
during
the
week.
A
A
Decisions
now
like
very
briefly
and
after
the
call
I
try
to
either
copy
notes
from
those
documents
here
or
just
link
to
them.
So
if
anyone
who
is
interested
in
back
who
was
participated
in
the
session
or
what
topics
were
discussed,
we
had
pretty
good
notes
from
each.
The
session
on
the
future
of
browser
integrations
was
more
or
less
revisiting
the
history,
how
our
browser
extension
evolved
over
time.
Who
was
the
the
majority
of
end-users?
And
what
was
the
purpose
and
we've
trying
to
answer
the
question?
Is
it
still
the
same
user
base?
A
It
expects
you
to
run,
go
up
events
locally
and
there's
a
welcome
screen
which
asks
you
to
install
go
activist
or
aquifers
desktop,
and
we
believe
going
forward
will
probably
remove,
remove
features
from
companion
or
maybe
like
not
remove,
features,
remove
them
from
companion
and
either
add
them
to
desktop
or
web
UI.
Mainly
due
to
the
fact
how
browser
extension
ecosystem
is
changing.
A
A
Believe
that's
more
or
less
the
plan
remove
like
so
I.
When
we
add
new
functionality
expose
new
functionality.
It
should
be
only
something
that's
possible
in
the
browser
extension,
namely
redirect
to
local
ironic
node
or
installing,
like
the
protocol
handler.
If
we
have
such
API
at
some
point
and
all
the
user
interfaces
should
probably
be
dedicated
to
IP,
why
that
was
it
for
this
session.
Another
session
was
it
shorter.
It
was
mostly
an
effort
to
agree.
A
When
do
we
switch
to
CID
v1
as
a
default
and
what
we
mean
by
switching
to
CID
you
want,
and
the
decision
we've
made
is
that
we
will
probably
switch
to
switch
the
output,
for
example,
when
you
add
something
to
idea
first,
using
a
give
us
add
command,
and
then
you
get
a
CID,
that's
the
ID
right
now
we
see
a
DV
0,
which
is
multi
hashing
base
58.
We
will
switch
that
default
output
to
see
a
DV
1
in
verse,
42,
and
we
are
planning
to
do
that
around
go
FS
0.6.
A
A
You
can
do
I
give
us
our
CID
version
1
and
you
will
get
ready,
V
1
in
the
output.
The
difference
is
that
it's
not
only
changing
the
version
in
the
output.
It's
also
changing
the
CID
version
used
on
every
node
in
the
dog
that's
created,
and
that
means
the
final
multi
hash
in
the
root
will
change
also
due
to
the
fact
that
those
intermediary
intermediary
nodes,
UCI
dv1.
It
means
we
now
have
multi
codec
as
a
part
of
CID
and
thanks
to
that,
the
leaves
of
that
tree
can
be
roll.
A
If
so,
the
multi
codec
of
those
nodes
will
be
zero
and
that
also
changes
all
the
hashes.
That's
why
we
won't
be
changing
this
behaviour.
This
flag
will
produce
the
same
CID.
We
will
just
update
like
help
to
be
very
clear
on
what
CID
version
does
I
believe.
That's
was
the
decision
all
in
this
session,
I,
probably
PR
and
Link
those
changes
below
another
session
checkup
who
do
want
to
take
this
one.
B
C
Where
are
things
going
to
be
because
realistically,
on
a
webpage
running,
the
DHT
isn't
terribly
realistic.
Just
because
of
the
sheer
number
of
connections
that
we
have
to
do
even
in
client
mode
to
query.
But
there
are
other
contexts
ingest
that
it
wouldn't
make
sense
to
run
a
DHT.
So
we
want
to
make
that
we
are
accounting
for
all
of
us.
A
C
C
We
need
to
lock
down
how
we
want
to
handle
the
security
of
like
client
to
server,
because
when
we
get
the
direct
connect,
the
final
connection
that
secured
to
Sakai
oh-
but
we
have
had
some
discussions
around
start
us
running
because
we
are
running
when
you
connect
to
the
server
that's
currently
going
to
be
over
WebSockets
are
which
is
already
secure.
So
if
that
also
additionally
runs
over
Sakai.
C
C
On
top
of
that,
so
we
can
either
leave
it
open
and
then
like
take
the
hit
with
the
double
encryption
and
say:
okay,
if
you
want
to
dial
over
start
a
server
over
TCP,
whoever
TC,
you
can
totally
do
that,
but
then
we
would
need
to
lock
that
down
more
than
likely.
Otherwise
we
risk
getting
really
muddy
with
configuration
options
or
just
what
it
supports.
The
other
option
is
to
just
lock
it
down
and
say:
no,
we
only
support
WebSockets
tar
that
way.
C
Well,
we
don't
or
WebSockets
secure
WebSockets
so
that
we
don't
have
to
worry
about
it,
and
then
we
can
leave
it
plain
text.
So
we
don't
take
that
hit
on
double
encryption,
so
we
just
need
to
decide
which
path
forward.
More
than
likely.
Our
use
case
is
we're
really
trying
to
cover
the
use
case
of
browser
to
browser
and
browser
knows
discovering
browser
notes,
and
so
that's
really
the
thing
that
we
should
be
optimizing
for
so
we'll
probably
end
up
just
locking
it
down
and
optimizing
for
those
connections.
B
C
Running
over
that
relay
as
a
web
web
sock
is
here
so
that
wouldn't
be
double
encryption
from
the
browser
to
the
server,
so
that
would
run
over
WebSockets
tar
and
then,
when
we
made
the
underlying
connection
that
still
gets
encrypted.
So
the
underlying
stream
stays
encrypted,
but
we're
just
saving
on,
like
the
connection
to
the
server
to
negotiate
that
contract
and
dial
and
then
like
as
we're
streaming
peers
from
that
server.
That
would
be
only
over
the
WebSocket
secure
connection
right.
B
C
B
C
C
It
was
a
similar
scenario
that
the
big
difference
is
like
we.
The
reason
like
we're
having
this
problem
with
start
us
is
that
we're
basically
running
a
the
refactor
is
running
a
lip
p2p
server,
rather
than
it
being
like
a
socket
IO
server.
It's
just
no.
This
is
a
lip
PDP
server
and
it
runs
socket
connections
so
really
using
our
own
tech
to
be
the
server.
So
then
we
just
have
to
handle
the
take
accepting
the
WebSocket
connection
as
the
server
a
secure
web
socket
connection
so
having
the
actual,
like
HTTP
connection
in
front.
B
There's
there's
no
there's.
No
reason
why
why
we
have
to
have
double
encryption
is
just
doesn't
work
for
those
transports
to
be
aware
of
the
context
that
they're
running
in
no
they
don't.
We
don't
need
to
the
underlying
transport
it's
already
encrypted,
so
we
don't
need
to
set
encrypt
as
well.
Are
we
moving
away
from
Sakai,
oh
okay,
yeah.
C
A
C
B
Think
like
right
right
now,
like
especially
given
the
stage
a
lot
of
this
stuff
is
I.
Think
the
trait
it's
a
much
better
trade-off
to
just
broaden
our
set
of
connectivity
options
and
worry
about
performance
optimization
later,
because
if
that
tree,
if
we're
double
encrypted,
what
we're
we're
gonna
be
slower
but
safer
already,
but
I'm
having
more
routes.
Think
from
a
from
a
higher
level
is
way
more
important.
We
don't
solve
double
encryption.
Ever
computers
will
just
get
faster
anyway,
and
so
will
network
connections
so
yeah.
C
A
Like
I
I
become
very
skeptical
when
we
say
something
is
a
stopgap
because
we've
been,
we
have
some
stuff
got
like
some
stop
gaps
still
around
and
probably
will
be
and
like.
Even
if
we
have
this
like
the
centralized,
signaling
thingy
at
some
point
in
the
future,
some
people
will
still
want
to
run
Stardust
just
to
make
it
faster
and
I
feel
there's
another
problem
of
like
threat
modeling.
A
A
C
So,
basically
like
what
we
can
do
is
when
we
go
when
we
upgrade
connections
like
we
can
give
transports
upgraders
and
those
upgraders
will
determine
what
the
upgrade
path
is
or
whether
it
needs
which
encryption
it
needs,
whether
it's
plain
text
or
Sekai
or
noise.
So
we
can
change
that
potentially
based
on
context,
but
yeah
the
way
I
look
at
it
is
like.
Ideally,
what
we
would
be
doing
in
the
future
is
replacing
start
us
with
a
relay
server
and
that
relay
server
is
going
to
be
running
Sakaya
or
noise
encryption.
C
We
probably
aren't
going
to
necessarily
have
no.
We
will
so
like
a
being
able
to
have
a
WebSocket
secure,
endpoint
on
that
relay
so
that
WebSocket
browsers
can
hit
it.
So
basically
doing
the
same
thing
like
we'll
have
to
deal
with
there
too,
but
we're
running
like
a
full
circuit
relay,
and
so
this
is
like
a
stopgap
of.
B
Just
for
remote
prioritization
feedback
I
have
met
with
two
different
groups
in
the
last
week
who
both
were
saying
that
this
in
browser
connectivity
is
the
number
one
blocker
for
them
in
terms
of
their
there's.
They
want
to
be
able
to
ship
ipfs
web
apps
over
HTTP
and
that
that's
gonna
be
their
model
for
four
years
to
come.
Right
like
you,
even
once
we
get
native
native
IP
fence,
which
is
still
talking
about
a
couple
of
different
couple
years.
B
Just
for
you
know
the
best
case,
you're
one
year
for
proof
of
concept
and
then
a
couple
of
years
for
a
standardization
and
browser
adoption,
probably
yeah
three
five
seven
years
possibly-
and
so
this
ipfs
over
HTTP
web
case,
especially
browser
to
browser,
is
becoming
more
more
part
of
what
people
are.
Gonna
depend
on
one
of
the
pieces
of
feedback
that
I
heard
yesterday
from
a
group
that
I
met
with
that
have
their
own
proof
of
state
blockchains,
the
developer,
the
ecosystem
they're
building
out,
but
an
incubator
accelerator.
B
So
they
have
a
bunch
of
companies
that
are
building
on
top
of
the
stack.
Is
that
I
profess
this
one
of
the
few
projects
in
the
space
that
actually
kind
of
speaks
directly
to
in
browser
as
use
case,
so
some
of
the
other
projects
don't
really
treat
browser
as
a
first-class
customer
and
they
really
thought
that
we
did.
But
the
problem
was
that
it
just
does
from
the
foot
I
think
the
way
they
described
it
was.
B
We
treat
it
seriously,
but
it
just
doesn't
work
yet
so
they're
like
we
love
that
you
treat
it
seriously
and
then
the
Brahe
you
think
of
browser
running
in
browser
should
be
first-class
citizen
and
we
just
need
it
the
basics
to
work
and
what's
the
basic
connectivity
works,
we
can
build
so
much
more
stuff
on
top
of
it.
So
this
is
from
a
you
know
the
unblocking,
the
people
that
want
to
build
on
our
stack
and
really
making
our
stack
work.
First-Class
for
how
people
build
and
develop
apps.
A
I
made
a
draft
of
grant
for
coming
up
with
native
protocol
handler
for
browser
extensions,
and
it
will
probably
land
in
the
plant
sweeper
soon.
But
basically,
the
idea
is
to
create
a
new
API
that
enables
browser
extension
to
register
a
handler
that
is
capable
of
returning
arbitrary
bytes
to
the
browser
to
the
rendering
process
and
that's
unlocks
a
lot
of
interesting
things
of
the
value
is
for
the
centralization
with
no
longer
give
us,
but
also
protocols,
locked
out
and
secure.
A
Skuttlebutt
could
use
this
new
API
for
returning
data
fetched
from
distributed
network
without
relying
on
some
sorta
like
third-party
service,
local
and
offline
use.
Cases
could
be
built.
On
top
of
that.
A
lot
of
extensions
rely
on
some
servers,
which
act
as
a
half
of
the
function
like
are
responsible
for
half
of
functionality
and
for
idea
fest.
We
of
course,
would
not
be
forced
to
run
local
HTTP
gateway
for
users
who
want
to
very
light
experience
just
to
upload
file
or
data
file.
A
A
The
problem
is:
when
you
fetch
it
bites,
you
are,
you
are
not
able
to
provide
them
to
to
the
renderer
process
to
display
HTML,
so
we
been
talking
with
open-source
console
consultancy
called
Galia
I
believe
Dietrich
mentioned
them
before
right
now,
I
wrote
a
lot
of
like
prior
art,
so
we
have
those
wet
IP
is
where
regular
website
can
see
with
register
this
like
handler,
which
is
just
a
redirect
to
some
web
service
somewhere.
So
those
are
like
web-based
protocol
handlers
which
require
web
Plus
prefixes
for
most
of
protocols.
A
Another
prior
art
we
have
is
in
Firefox
browser
extensions,
are
able
to
put
protocol
handlers
in
manifest
of
their
extension
and
that
automates
this
redirect
based
registration
is
still
a
fake
protocol
because
it's
redirects
to
some
other
URL,
so
you
need
to
have
a
server
somewhere
who
will
handle
the
URI?
And
finally,
we
had
native
handlers
before
we
had
one
in
new
on
base
brave
before
they
switch
the
chromium,
and
we
had
this
famous
famous
proof
of
concept
when
we
had
Firefox
nightly
with
the
keyword
experiments,
which
was
basically
what
we
want
from
this
grant.
A
But
the
thing
that
we
want
in
this
grant
is
that
sort
of
a
functionality
but
for
chromium
based
vendors.
That's
mostly
because
the
browser
vendors
who
are
most
vocal
about
most
interests,
it's
about
enabling
artifice
support
our
chromium
based.
So
that's
the
plan
for
this
grant
is
to
unlock
that
type
of
integrations.
I
should
be
posting
this
at
some
point
this
week,
maybe
next
week,
but
basically
we
want
to
create
a
bi
specification
which
could
be
adopted
by
other
browser
vendors.
So
it
should
be
like
generic,
not
specific,
IP
affairs.
A
A
So
this
grant
will
be
not
only
like
interested
from
the
grant
like
grant
creating
perspective,
but
also
if
anyone
is
interested
in
history
or
what's
actually
possible
on
the
web
platform,
it
will
have
like
a
lot
of
resources
and
like
list
of
chronological,
chronological
private
art,
so
I
feel
it
would
be
interesting.
Resource
I
will
link
it
later
when
it's
published
something
definitely
power.
A
B
That's
it
there's
a
fantastic
sparking
into
the
history
of
how
many
people
have
tried
to
do
something
like
this,
and
one
of
the
things
that
we're
doing
here
is
kind
of
an
end
run
around
browser
vendors.
That
are,
you
know.
The
big
browser
vendors
are
not
really
interested
in
adopting
this
type
of
really
powerful
functionality.
You
know,
even
in
even
in
Mozilla
the
the
pushback
from
an
engineering
and
engineering
groups,
also
the
security
groups
around
some
of
these
we're,
pretty
strong
and
and
and
not
without
good
reason.
B
Right,
like
really
really
well
founded
concerns,
and
that's
something
there
where
we're
using
this
to
really
push
on
the
boundaries
of
that
when
you
have
partners
in
browsers,
smaller
browsers,
but
still
big
enough.
They're
are
willing
to
do
that.
Experimentation
take
those
risks,
so
they're
kind
of
we're
short-circuiting
not
a
bit
where
if
we
can
get
something
landed
in
chromium-
and
this
will
be
our
first.
B
You
create
this
environmental
in
the
pressure
on
on
these
organizations
around
how
they
make
decisions,
how
they
evaluate
functionality
and
how
they
make
those
trade-offs
around
a
developer
adoption
and
a
number
a
number
of
other
things.
So
it's
a
pretty
nuanced
to
play
in
in
in
figuring
out
where
our
levers
are
for
having
influence
in
these
organizations
and
be
able
to
have
somebody
like
the
Collier
who's
really
interested
in
both
that
model,
but
also
have
the
keys
when
it
comes
to
being
embedded
in
those
communities
and
landing
code
is
really
really
powerful.
B
B
They
share
with
them
and
get
their
feedback,
because
it's
gonna
be
an
iterative
process
to
be
able
to
figure
out
what
the
right
format
is,
that
both
meats
are
or
I
could
post,
have
grants
kind
of
format,
requirements
and
also
something
that
has
that
level
of
detail
and
structure
that
they
need
to
go
and
do
the
work
so
I'd
say
first
off
get
this
in
front
of
them
and
get
feedback
there.
This
is
fantastic
thanks
for
writing.
This
up.
B
That
the
video
yes
yeah
I
met
with
a
unstoppable
at
Eve
Denver
and
had
an
interesting
conversation.
We've
talked
with
them
before
and
but
the
the
new
wrinkle
was
you
know.
I
did
a
talk,
kind
of
about
our
ecosystem
group
and
also
the
browser
work
that
we're
doing
at
the
decentralized
network
summit
the
day
before
with
Denver,
and
they
were
there
and
there
they
were
basically
like
this.
You
just
talked
about
you.
Just
did
our
talk
for
us.
B
These
are
the
kind
of
things
that
we
want
to
see
happen
with,
with
IP
fussing
and
and
and
distributed
web.
Your
browsers
so
they're
very
interested
in
doing
this
kind
of
work.
They
shipped
that
demo
browser
which
has
a
supple
domain
stuff
built
in,
but
also
has
a
built-in
ipfs
node.
It
also
has
built
in
several
built-in
gateway
options.
So
it's
worth
yeah.
Actually
just
going
like
you,
you
load
the
browser,
and
the
first
thing
you
see
is
this
banner
across
the
bottom:
that's
like
you,
don't
have
an
IP
FS
no
running.
B
You
should
run
one
to
be
able
to
support
the
distributed
web,
but
then,
if
you
actually
go
turn
on
your
I
think
that's
desktop
you're
like
make
sure
you're
like
I
gotta.
Stop
writing
what
the
hell
like
I
have
a
node.
You
know
detect
a
local
node.
They
actually
run
a
node
in
browser
fat,
just
so
fascinating.
So
it's
pretty
and
I
don't
really
dug
into
the
source
code
at
all
like
did
they
just
bundle
Jesus
IP,
IP
fast
and
run
that
node?
Are
they
running
a
CO?
B
Node
I
don't
know
so
it
would
be
worth
spelunking
a
little
bit.
I
filed
an
issue
in
their
repo
around
implementing
our
addressing
spec
and
there
they
replied
immediately
to
me
an
email
saying:
hey
yeah,
we'll
totally
look
at
that,
but
they
really
they
really.
They
said
that
they
want,
for
this
browser
to
be
a
place
for
people
to
experiment.
So
this
is
something
where
we
can.
B
Maybe
you
know,
will
work
and
and
and
talk
about,
yeah
and
reallyy
demo
and
hack
on
some
of
what
chromium
underlying
chromium
changes
would
be
like
the
the
weird
bit
is
like
you
know,
this
is
I,
think
electron
based
it's
not
like
they
forth
brave
or
something
like
that.
So
you
don't
get
the
full
browser,
but
you
know
one
of
the
basic
functionality.
Is
there
so
really
I?
Think
socializing
and
highlighting
these
type
of
experiments
really
helps
us
push
them.
B
You
know
socialize
with
the
boundaries
of
what
about
other
web
browser
could
do
what
what
could
look
like
help
network
if
they
implement
the
addressing
spec.
We
we
then
now
now
have
a
desktop
browser
that
you
can
download.
That
has
a
protocol
handler,
which
would
be
pretty
amazing
and
also
might
help
prove
out
some
of
those
protocol
handler
paths
right.
Multiple
invitations
is
always
gonna,
be
nice,
so
even
if
we
say
have
the
native
Cryme
one,
that's
going,
you
know
maybe
be
a
longer
term
project
of
Gallia.
B
But
then
we
have
something
like
this,
where
we
have
an
experimental
vehicle
that
we
could
play
with
and
fork
like,
we
could
even
fork
this
and
just
ipfs
put
an
IP
pencil,
go
on
it
and
play
with
it
and
push
stuff
out
there
and
have
their
services
built
in
we
could
we
could
I'm
really
interested
in
having
lots
of
these
lots
of
these
experiments
in
name
resolution
be
easily
accessible
to
developers
right
now.
These
these
feel,
like
both
ENS
and
unstoppable,
are
different
ways
of
solving
this
naming
problem
and
I.
B
Think
we're
gonna
see
a
lot
of
competition
and
experimentation
in
that
place
over
time.
There's
not
one
winner.
There
might
be
a
pattern
that
emerges
and
standard
standard
around
this
and
talking
with
them
anything
very
they're,
also
interested
in
kind
of
standards,
track
participation
as
well,
so
worth
worth,
checking
out,
I
told
them
I
putted
them
at
the
at
the
list
for
I,
confess,
weekly
and
I,
said
we'd
love
to
have
them
demo
and
introduce
the
the
browser
to
the
community
talk
about
how
it
was
implemented
in
the
future,
so,
hopefully
they'll
they'll.
B
B
Jim
is
starting
to
move
on
to
the
mobile
mobile
IP
fest
design,
research
that
grant
just
got
approved,
but
that's
that's
more
like
mobile
as
a
big
circle,
with
browsers
as
a
sub
mobile
browsers
as
a
subset
of
that
work,
so
broadening
the
scope
there,
but
also
some
some
stuff
related
to
browsers,
but
for
the
the
first
one
though
guidelines
I
still
need
to
do.
The
final
pass
on
the
content
fix
up
the
iconography
section,
because
I
think
we
made
some
changes
there.
Fixture
your
feedback
and
comments.
B
A
I
believe
I
did
like
a
pass
before
team.
Okay,
just
around
visuals
I
was
just
very
skittish
about
like
sort
of
like
having
fake
addresses,
which
are
not
actually
following
our
a
dressings.
Okay,
though
I
added
like
comment
for
for
every
occurrence,
I
found,
so
that
may
be
helpful.
Another
thing
identified
is
sort
of
like
maybe
not
a
blood,
so
like
a
blank
spot
around
iconography,
its
distinction
between
mutability
and
immutability.
A
That's,
maybe
even
like
a
wider
like
a
separate
research,
how
people
understand
that
or
how
to
communicate
that
to
people,
because
a
lot
of
people
don't
even
know
that
this
distinction
exists
and
like
in
ipfs.
We
got
like
replacing
the
green
padlock
means
not
only
indicating
like
integrity
guarantees
not
only
like
the
transport
encryption
guarantees,
but
I
believe
also,
which
sort
of
should
have
a
way
of
telling
this
is
immutable
address,
or
this
is
an
address
that
may
like
the
content,
may
change
yeah.
B
Like
like,
like
there's
this,
it's
basically
a
new
trust
model.
It's
not
even
like
a
wrinkle
or
change
the
existing
trust
model,
it's
a
type
of
guarantee
of
content
validity
that
is
not
rooted
in
external
trust
systems,
but
is
more
like
internal
integrity
and
the
trust
model
is
out-of-band
and
that's
something
that
we
really
wrestled
with
and
went
back
and
forth
on
and
ultimately
I
think
like
for
this.
B
For
the
purposes
of
this
first
version
narrowing
the
field
of
view
to
be
able
to
focus
on
like
what's
the
minimum
set
of
thing
that
browser
should
this
should
design
for
in
a
way
that
also
respects
and
protecting
the
user
and
like
as
as
much
as
possible,
it's
really
a
hard
it's
hard
to
draw
boundaries
there
without
really
boiling
the
ocean.
So
I
think
it's
narrowing
the
field
of
view
for
the
for
each
one
of
these
and
making
them
as
bite-sized
as
possible.
It's
gonna
be
yeah.
A
Especially
given
the
fact
like
for
this
demo,
browser
we
can
see
like
for
us
aside,
give
us
project.
We
know
that
ipfs
namespace
is
immutable
and
IP
NS
namespace
is
mutable,
and
for
us
it's
just
like
like
water
for
fish,
we
don't
even
know
it's
there.
However,
when
people
start
implementing
it
that
they
put
like
mutable
names
in
immutable
protocol
name
and
also
like
regular
people,
don't
really
know
what
either
of
those
things
is
so
probably,
instead
of
like
relying
on
protocol
scheme
or
I.
A
B
A
B
Are
the
questions
that
we
really
have
to
answer
like
in
order
to
be
able
to
achieve
mass
adoption?
We
need
to
be
able
to
figure
like
we
need
to
get
that
part
of
our
house
in
order
and
communicated
really
well
and
really
well
understood
in
a
way
and
have
the
resources
for
people
who
are
thinking
about
adopting
to
be
able
to
have
them
make
these
evaluations,
especially
if
they're
good
they're
designing
for
millions
and
millions
of
users.
It's
super
important.
It's
not
gonna
be
easy.
B
B
So
they
we
have
a
latest
bill
from
them.
I
think
that
they
were
still
mostly
on
on
track
for
releasing
pretty
soon
but
we're
you
know
playing
this
dance
with
with
the
release
of
0.5
of
call
us
and
getting
the
CID
and
said
you
mean
sport
and
soda
main
gateways.
Support
really
well
settled.
Having
that
those
changes
propagate
amongst
to
all
the
the
big
gateway
partners
to
it
was
really
important.
B
B
Folks,
the
the
lead
p2p
folks,
especially
even
if
not
participating,
these
discussions
coming
in
to
be
parting,
the
hallway
conversation
and
and
lurking,
and
getting
to
know
the
people
that
are
making
these
decisions.
There
are
whole
tracks
around
distributed
networking
and
a
lot
of
these
like
technologies
that
we're
kind
of
hacking
around
or
building
on
top
of,
because
the
underlying
transport
doesn't
meet
our
needs
decisions
about
how
those
underlying
transports
are
being
made
in
these
meetings.
B
We're
gonna
do
that
in
a
couple
of
big
ways.
This
year,
I
think,
but
really
these
are
parallel
tracks
from
from
an
influence
and
and
and
change
making
change
that
we
want
to
see
in
the
work
standpoint.
So
being
being
part
of
that
conversation,
while
we're
also
short-circuiting
or
hacking
around,
really
helps
them
understand
our
vision
so
and
and
and
our
needs
and
where,
where
the
current
transport
level
transports
aren't
meeting
our
needs.
So
I'd
encourage
you
to
go.
Look
at
the
t's,
IETF
107
schedule.
B
Gonna
recommend
for
this
particular
group.
If
we're
doing,
if
you're
doing
work
on
J
sled
p2p,
if
you're
doing
for
the
whole,
if
you
to
be
generally,
should
probably
be
there
but
are
specifically
our
work
around
our
browsers
and
and
and
network
in
networking
and
connectivity
in
the
web
content
context.
This
is
something
where
I'd
almost
say
like.
B
Let's
just
take
take
the
entire
week,
not
take
it
off,
but
take
off
our
ipfs
hat
and
just
really
sit
these
conversations
in
Madrid
in
July
and
soak
it
all
in
and
learn
who
the
people
are,
how
they
communicate
with
each
other.
How
the
IETF
works
at
the
beginning
of
every
ITF
work.
They
actually
have
like
a
newbies
Meetup
the
day
before
the
conference
starts
where
they
have
like
an
introduction
to
have
the
ITF
words.
B
So
things
like
that,
a
really
nice
from
an
operating
perspective,
but
I'd
love
to
see
as
many
people
as
we
can
from
this
group
be
part
of
that.
And
then
what
we'll
probably
do
is.
Maybe
you
know
tack
on
three
days
at
the
beginning
of
that
to
be
able
to
do
browser-specific
hacking
work,
but
that's
that's
tentative
for
now,
but
kind
of
block
that
out
on
your
calendar
for
I
think
it's
mid
to
late
July
in
Madrid.
The.
B
B
Hugo
demoed
all
this
amazing
in
browser
and
extension
level
testing
framework
last
week.
How
does
that
affect
us?
Lionel?
Do
you
have
any
idea?
We
have
this
testing
matrix
of
all
these
things
that
we
wanted.
Eventually
we
haven't
got
there
yet
are
any
of
these
testing
scenarios
for
desktop
or
extension
or
web
unlocked
for
us
from
the
work
that
Hugo
did?
Is
there
any
transition
that
we
need
to
be
able
to
make
there
yeah.
A
So
most
of
our
modules
are
like
most
of
jaeseok
is
using
edger
and,
as
you
was
using
karma,
which
hugo
aims
to
replace-
and
you
know-
and
he
effectively
like
wants
to
the
moment-
he's
able
to
do
the
same
things
in
this
new
runner.
He
wants
to
switch
to
that
because
it's
like
not
only
much
smaller
like
it.
There's
like
the
the
difference
is
not
that
visible.
A
But
if
you
like,
look
how
complex
the
old
solution
was
and
how
small
and
relatively
elegant
the
new
one
is
it's
like
no
brainer
for
us
to
switch
so
I
think
a
lot
of
switching
will
happen
behind
the
scenes,
just
as
you're
switching
to
this
new
runner
without
the
need
for
doing
anything
for
activist
companion,
specifically
I'm
excited,
because
it's
for
the
first
time
enables
us
to
run
tests
against
actual
web
extension
runtime.
Instead
of
like
seen
on
mocks
and
stops.
The
problem
is
right.
A
A
A
B
That's
great
and
ready
companion
actually
be
able
to
test
companion
in
a
real
browser
environment.
This.
It's
really
really.
Nice
I'm
also
interested
in
his
work
to
get
this
this
infrastructure
running
with
test
grounds.
So
then
we
can
say
you
know.
Spinner
spend
a
whole
lot
of
lot
of
browser
events,
yeah
and.
A
Cool
I
added
it
to
the
highlight
so
I've
added
two
things
to
the
highlight
section:
one
is
that
I've
released
the
new
version
of
IPS
companion
to
the
beta
Channel
and
it's
a
version
that
is
addressing
our
problems
with
chrome
webstore.
So
there's
a
an
issue
with
never
I
made
one
issue
because
I
would
be
creating
issues
over
and
over
again
and
now.
I
just
have
this
one
and
I
reuse
it
for
all
the
problems
we
have
for
Chrome
Web
Store.
A
So
it's
basically,
what
happened
is
that
Chrome
Web
Store
now
ask
for
permission
justification
for
every
permission
that
your
extension
ask
for
it's
pretty
good
idea
in
theory,
but
in
practice
it's
super
automated
and
when
you
get
rejected
you
don't
get
like
human,
readable
feedback.
It's
just
like
robot,
saying
hey.
You
did
not
comply
to
our
privacy
policy,
please
fix
it,
but
they
don't
say
what.
Luckily
we
had
some
help
from
Simon
from
Google
and
we
were
able
to
fix
those
issues.
A
I
made
the
new,
better
release
which
removed
unused
permissions
and
hopefully
will
pass
the
review.
It
takes
time
about
a
week
to
get
approved.
So
the
moment
the
beta
is
released,
we
will
be
able
to
reuse
the
same
descriptions
for
stable
Channel.
Why
it's
important,
because
the
Chrome
Web
Store
so
far
is
used
as
a
distribution
platform
for
all
the
chromium
based
browsers,
including
Microsoft,
edge
brave
and
if
you
take
alone,
crazy,
yeah.
And
if
you
take
a
look
at
our
user
base.
A
A
Well,
I,
don't
think
it
will
get
better.
Google
is
investing
into
manifest
to
be
free.
I
probably
mentioned
that
previous
calls,
but
in
case
someone
wants
to
read
more
about
what
manifest
v3
is.
We
got
this
properly
numbered
issue
in
ipfs,
companion,
repo
and
I.
Try
to
like
follow
up
developments
related
bugs
and
things
like
that.
A
We'll
see
what
happens
when
Google
decided
like
when
chrome
webstore
decides,
we
only
accept
manifestly
free
extensions.
That
would
probably
mean
brave
opera
edge
and
all
the
other
chromium
based
browsers
that
still
supports
the
old
API
will
either
provide
own
stores
or
will
a
community
based
store
will
be
created?
D
A
Green,
all
checks
are
passing
if,
yes,
a
background
is
like
chocolate
is
like
a
package
manager
for
Windows
right
and
we
got
a
package
for
ID
because
they
stop
there.
So
if
you
are
using
that
you
can
install
latest
version
desktop
with
these
commands
and
then
you
will
get
updates
over
time,
I
believe
that's
it.
From
my
end,.
A
Unless
you
are
using
Catalina,
then
you
have
nasty
pop
I
believe
we
have.
We
just
need
to
invest
time
into
that
notarization
step
because
we
are
already
signing
I
like
I,
try
to
understand
prior
out
in
our
continuous
integration
setup,
and
we
are
already
like
we
already
have
keys.
We
are
already
signing
and
signing
happens
on
our
CI,
so
every
time
we
make
a
release,
it
gets
signs.
Then
there's
this
additional
step.
A
B
A
B
Really
I
think
that
cats
act
that
catallena
it
necessarily
no
one's
burning
gonna
burn
us
that
we
need
to
be
that
one
sent
what
said
whatever
you
info
you
have
after
after
the
call
and
then
I'll
just
follow
the
rest
of
the
citing
steps,
because
I've
done
that
birth
I
did
it
for
mobile
stuff.
I've,
never
done
for
desktop
one
but
I'm
familiar
with
just
how
crappy
and
terrible
that
workflow
and
those
you
eyes
are.