►
From YouTube: GUI and In Web Browsers Weekly, 2019-02-13
Description
IPFS Mirror: https://ipfs.io/ipfs/bafybeif57ylbap2tuqrl7nb25y7hxlncm2ebjd74u4ck2hpmljitylmxc4/
Meeting Notes: https://github.com/ipfs/team-mgmt/pull/874
A
A
My
act
of
someone
who
cares
about
gooeys
I
have
increased
the
font
size
in
this
document,
so
we
can
actually
read
it
when
it's
face
lit
up
on
calls
tada,
we
have
an
agenda,
we
have
a
note
taker,
that's
Jim
and
the
recording
has
started
so
I
will
start
with
two
demos
and
then
we'll
go
into
miscellaneous
other
demos
right
so
who's.
First
up
as
always,
it
is
vital
we
like
their
here's
with
your
highlights
from
last
week,
title.
A
B
B
So
far,
I
got
positive
feedback
and
some
good
hints
I
already
like
change
the
namespace
to
no
longer
collide
with
mdns.
So
if
anyone
is
interested
how
local
gateway
could
look
in
web
browser
zone,
this
is
a
good
issue
to
follow.
How
it
would
work
automatically
is
that
ipfs
companion
is
able
to
set
HTTP
proxy
for
arbitrary
URLs,
so
the
users
would
just
install
activex
as
companion
and
the
proxy
would
be
automatically
set,
so
those
domains
local
host
domains
would
be
proxy
by
go
ipfs.
B
B
E
B
Like
no,
it's
like
a
long
issue
of
issues
like
a
long
list
of
issues
and
I
managed
to
resolve
our
reproducibility
issues
in
the
companion
itself,
but
the
reviewer
noticed
that
it's
not
possible
to
build
web
UI
like
to
get
exactly
the
same
unified
code
and
I
open
issue
in
activist.
What
do
I
will
have
a
discussion
on
this
later,
but
basically
it's
like
an
ongoing
thing.
That
will
be
a
priority
for
me
this
week,
so
I'm
blocked
on
that
and
first
I
want
to
resolve
that
problem,
and
then
I
will
be
probably
finishing.
B
F
F
This
is
a
file
and
each
one
of
these
have
inside
another
component.
That
was
this
one
context
random.
So
you
guys
can
imagine
the
huge
Dom
node
that
the
huge
dome
it
would
appear.
If
you
have
a
list
with
100
files,
we
would
have
a
hundred
in
new
files
with
a
hundred
context
menu.
So
I
had
to
decouple
all
of
that
project,
and
now
we
only
have
one
hundreds
menu.
It
works
with
every
file
that
was
a
bit
of
work.
I
know
it's
just
that
idea,
but
performance
wise.
F
F
It's
not
finished,
yet
he
has
a
lot
of
bugs
yeah
they're,
still
rendering
issues
I
still
need
to
fix
some
some
things,
but
I
hope
to
finish
this
this
week.
I
can
show
you
guys
an
example.
I
want
to
do
the
same
here
as
you
can
see.
This
is
using
you
see
this
little
are
playing
diseases
in
react,
virtualized,
it's
only
rendering.
What's
inside
this
window,
but
as
you
can
see,
we
have
two
Scrolls.
We
have
the
main
scroll
all
the
way
loaded
and
the
little
scroll
or
react
actualize.
F
These
things-
I,
don't
like
it,
but
while
I
was
checking
the
how
to
do
this
in
the
file,
I
noticed
that
there
is
a
component
that
we
can
use
to.
Basically,
this
is
using
react
fertilized,
but
it's
overriding
the
scroll
on
the
page,
so
we
can
scroll
the
page,
but
we're
actually
just
growing
on
the
reaction
for
life
what's
happening.
As
you
can
see,
there's
a
little
glitch
in
the
files
they
come
up
and
down.
F
I'm,
not
sure
why
that's
what
I
have
to
fix,
but
other
than
that,
although
he
remembers
all
the
glitch
that
was
happening
here,
not
happening
just
this
little
one,
okay,
so
basically
it's
it's
almost
and
I.
Think
then
I
have
to
work
with
the
other
interactions.
That's
missing
the
chef
of
the
planet
check
them
they're,
not
appearing,
but
there's
just
some
small
big
girls.
I
think
this
is
under
way.
A
F
A
G
F
F
Yeah
and
by
basically
at
least,
was
the
only
thing
that
was
left
to
use
the
expert
realized.
Is
we
already
using
the
export
and
the
peer
stage?
So
it
makes
sense
to
to
apply
it
here,
but
this
is
a
bit
more
complicated
because
it
has
a
lot
of
interactions,
drag
and
drops
checkboxes
any
other
things
so
yeah.
Hopefully
this
will
be
finished
and
we
can
check
the
new
version
of
the
IBI.
D
A
For
me,
super
cool,
thank
you
very
much.
I
think
it's
me
I
think
enrique
can't
be
with
us
this
week,
so
I'll
just
quickly.
He
is
mostly
on
holiday,
but
as
in
overworking
freshen,
he
has
done
one
important
thing,
which
is
we
got
the
IP
fest
desktop
we've
been
working
on
getting
signed,
builds
for
OS,
X
and
Windows,
and
the
OS
x
one
landed
a
couple
of
weeks
ago
and
the
windows
one
landed
last
week,
so
he's
upgraded.
A
The
build
pipeline
to
include
the
windows
certificates,
so
future
releases
of
desktop
will
be
signed,
which
means
fewer
warnings
on
install
is
good
UX
when
okay,
what
was
I
doing
no
I,
don't
think
I
need
to
share
my
screen
for
that.
Then
he
canceled
due
to
do
not
a
huge
amount
from
me.
This
week
mostly
been
doing
management
tasks,
but
the
exciting
one
is
there's
two
exciting
ones.
One
is
we're
about
to
get
count
lead.
A
My
PFS
Gayoso
will
be
able
to
get
telemetry
from
our
self
hosted
analytics
for
web
UI
and
desktop
I'm
expecting
that
to
land
this
week,
so
that
next
week
I
can
wire
it
into
it's
already
world.
At
the
web,
UI
I
just
need
to
get
that
PR
merged
and
then,
after
that,
I'm
gonna
add
additional
and
I
think
for
desktops
drop
down
menu,
but
other
than
that
we're
good
to
go
on
the
and
that's
all
it
respects
the
users
do
not
track
preference
in
their
browser
and
the
other
one
was.
A
We
got
github
actions
enabled
for
all
our
books,
so
we
can
start
experimenting
with
those.
If
that
is
our
want,
which
connects
to
the
agenda.
Point
which
is
Jenkins
has
gone
away
as
our
CI,
and
that
was
how
we
were
publishing
all
our
websites.
So
I
won't
talk
too
much
about
that.
We
can
say,
go
back
to
that
at
the
end,
and
next
week
we
are
releasing
I,
think
web
UI
2.4,
with
all
the
files
improvements
that
we
can
working
on.
I
think
we
got
to
the
end
of
the
list.
A
So
that's
very
exciting
kind
of
that's
going
to
drop
and
the
desktop
is
just
a
couple
of
tests
away
from
being
ready
for
release
as
well.
So
we
might
be
having
two
releases
this
week.
The
the
excitement
is
tangible
boom
and
then
I've
got
a
background
task
to
add
analytics
so
bunch
of
things.
But
hey,
that's
me
teri
did
you
want
to
do
an
update.
I
Sure
yeah,
mostly
what
I've
been
working
on
I'm
in
San
Francisco
this
week
with
the
community
team,
doing
a
little
hacking
and
bonding
and
whatnot
so
Michael
and
I
spent
some
time
yesterday.
Looking
at
the
file
tutorial
like
how
the
uploading
will
even
work
and
what
the
UI
would
look
like
for
that,
so
I
can
show
you
that
at
some
point,
I
don't
know
if
you
rather
do
it
now
or
later
on.
I
In
it
sure
fine,
the
oh,
so
many
windows,
okay,
sure
something
here,
look
a
slide
deck.
Okay.
There
we
go
so
for
those
who
haven't
seen
what
a
normal
lesson
looks
like
you
end
up
with
just
an
exercise
box
like
this,
and
you
update
some
code
and
you
hit,
submit
and
get
some
feedback,
and
that's
the
only
task
that
you
have
to
do.
I
I'm
working
on
kind
of
what
this
looks
like
and
I'm
very
open
to
feedback.
I
just
got
a
little
bit
of
feedback
from
the
design
team,
which
I've
implemented
a
tiny
bit
of
and
I
can
tell
you
some
other
things
they
suggested
so
making
this
kind
of
programmatically
know
whether
it's
file
s
or
not,
and
add
these
little
headers
for
the
multiple
steps
and
then
I
stole
this.
My
big
accomplishment
of
the
week
is
I
figured
out
how
to
copy
elements
from
google,
the
inspector
in
chrome,
which
I
didn't
know
how
to
do
so.
I
That
means
I,
don't
have
to
learn
where
to
find
your
HTML
and
all
your
access
to
steal
things
from
you,
it's
very
exciting,
and
then
this,
which
Michael
made
this
thing,
there's
something
broken
about
it
right
now.
So
if
you
would
like
to
click
to
upload,
you
have
to
do
it
twice.
You
can
also
drag
things
into
it.
I
But
consistency
was
one
of
the
asks
on
that
call
yeah.
So
if
you
decide
these
are
not
the
file,
so
you
want
to
be
working
with
it.
We'll
just
go
back
to
this
version,
and
then
you
would
do
whatever
a
year
and
hit
submit
right
now.
This
is
built
in
a
way
that
that
won't
work
but
pretend
that
it
will
evaluate
both
whether
you
have
indeed
uploaded
a
file
and
whether
you
put
in
whatever
code.
I
We
want
you
to
do
and
give
you
the
appropriate
messages
and
let
you
move
on
to
the
next
lesson
as
the
otherwise
would,
and
some
of
the
other
feedback
that
I
got
on
the
design
call.
Was
that
they'd
like
to
see
like
you
know,
here's
there's
first
upload
a
file.
Next
do
this.
They
would
rather
see
it
as
a
small
chunk
of
text
under
this
header
and
then
a
small
chunk
of
text
under
this
other,
and
we
talked
about
whether
we
might
want
to
have
like
little.
I
A
A
I
I
I
I
A
A
F
A
A
A
A
J
J
Secondly,
I
made
a
whole
bunch
of
progress
on
CID
v1
base
22
this
week
and
sent
out
a
request
for
there.
So
that's
good.
It's
coming
along,
we'll
be
with
you
soon
and
then
the
DHT
pull
request,
adding
DHT
to
chair
sub.
If
s
has
now
been
merged-
and
that
is
super,
good
news
will
be
available
in
the
0:35
release,
speaking
of
which
I
released
a
free
release
of
0
theta
5,
with
this
pull
request
in
it,
because
why
not?
But
there
there
is
there's
dragons.
J
If
you
are
using
this,
I
have
noticed
problems
such
as
unable
to
update
lock
within
the
state
still
threshold,
so
there's
a
bug
lair
and
it
picks
a
cpu
at
100%
if
you're
unlucky
and
but
it's
there
and
so
I
can
show
you
know
right
now,
I'm,
looking
at
okie,
you
see
my
screen
right.
Yeah
I'm
connects
I'm
using
companion.
J
Yes,
yes,
it's
actually
taking
my
CPU
right
now,
but
it's
got
mishmash
mismatch
in
number
anyway,
look
341
or
482.
Whichever
number
you
choose
to
choose
to
believe,
there
is
more
than
the
10
you
usually
get
so
things
are.
Things
are
happening,
although
there's
a
bug
that
needs
to
be
fixed
before
it
can
be
released.
So
that's,
that's!
J
He
pointed
out
that
if
we
use
the
test
profile
for
go
ipfs
and
in
our
tests
that
the
test
might
be
might
be
a
bit
faster,
so
I
thought
why
not
heck?
Why
not?
Why
not
give
it
a
go
so
I
I
gave
it
a
go
and
the
test
site
at
the
time
it
took
to
run
the
node
tests
locally
on
my
laptop
went
from
14
minutes
to
5
minutes.
J
I
was
like
wow,
that's
incredible,
and
so
that's
a
big
big
win,
I
just
applied
it
everywhere
and
that
pull
requests
have
been
merged
and
and
that's
all
good,
and
now
the
CI
has
changed
also
to
the
use
Travis.
So
we
are.
We
are
really
rocketing
with
the
CI
this
week,
so
I'm
super
happy
and
I'm
not
blocked
on
anything.
E
A
J
Seeing
is
it
only
in
the
release
notes
we
have
only
the
DHT,
it
really
depends
on
how
long
it
takes
to
resolve
the
issues
with
with
the
stability
of
lipids
being
the
DHT.
That's
in
there
right
now,
with
the
advent
of
the
DHT
I
think
we,
because
we
are
making
that
many
more
connections
and
I,
think
it's
going
to
uncover
a
whole
bunch
of
bugs
that
we
haven't
seen
before
now
that
are
going
to
need
to
be
addressed
and
at
the
moment
I.
Don't
I
can't
really
say
for
sure
how
long
that
might
be
there.
A
J
That
you,
you
I'm
sure
you
went,
there
was
a
lot
of
testing.
We
need
to
do
on
the
DHT
like
it's
not
only.
You
know
the
fact
that
it's
taking
the
cpi-u,
but
we
need
to
make
sure
that
we
know
it
works
and
it
works
well
and
isn't.
Gonna
pollute
the
pollute,
what's
out
there
already
and
in
any
way
and
yeah.
So
there's
a
way
to
go,
but
that
the
pull
request
for
that
for
adding
the
DHT
has
been
open
forever
and
the
tests
all
passed,
so
it
needed
to
get
merged.
A
H
Yes,
so
I've
been
runnin
with
Olay
and
Alex
and
Alan,
and
everyone
from
the
infrastructure
infrastructure
team,
so
basically
I
mean
all
I've
been
working
on
the
Travis
migration
and
all
I
did
was
around
topic.
A
H
H
The
one
thing
you
should
look
at
is
this
first
part:
if
in
any
way
you
are
using
or
comparing
I'll
without
the
is
your
Windows
job
will
probably
break
things
will
not
match.
So
you
should
be
careful
about
this
part.
If
you
are
doing
this
type
of
things
like
repairing
files
or
see
IDs,
we
should
add
lines
like
these
to
your
get
attributes
file
in
your
specific
repo.
H
Everything
else
is
just
straightforward,
just
copy/paste
what
everything
in
here
you
create
a
Travis
file.
You
have
your
nice
patch
to
the
readme
and
you
are
good
to
go
right
now.
We
are
in
the
process
of
moving
everyone,
because
Jenkins
died
and
didn't
like
tell
that
attire
to
everyone
else,
so
everyone
is
kind
of
finding
out
that
no
more
Jenkins
machines
are
running
so
be
careful
with
that
and
help
us
out.
H
So
for
the
next
couple
of
days,
I'll
be
looking
at
the
benchmark
repo
from
the
near
from
guys
as
we
I'm
here
with
the
info
team.
I'm
gonna
check
everything
in
there
and
make
sure
we
can
deal
with
it
and
understand
everything
that
is
in
there
and
probably
try
to
find
a
way
to
manage
and
probably
replicate
that's
infrastructure
to
other
working
groups.
So
that
will
it
be.
H
It's
also
integrating
the
benchmarks
on
the
two
on
Travis,
the
nightly
builds
or
the
night,
and
also
some
runs
on
the
PRS
which
specific
tests,
not
all
of
them,
because
all
and
take
one
time
to
finish
but
sell
them.
Two
most
important
ones
will
probably
be
running
them
on
each
full
requests
and
yeah.
As
we
finish,
the
migration
to
the
Travis
I
will
try
to
finish
the
window.
Size
be
ours,
so
they
can
be
included
on
the
next
release
of
IPA
face,
but
yeah.
We
are
having
some
troubles
with
the
PHP
and
windows.
A
I'm
gonna
jump
in
over
you,
ok,
so
key
one
midpoint
scoring.
We
have
the
in
web
browsers
working
group
and
the
IP
festivities
have
some
objectives
and
some
key
results
and
we're
halfway
through
the
first
quarter.
So
now
it's
time
to
go
through
and
review
are
we
are
we
building
what
we
said?
We
thought
was
important
to
build
a
month
and
a
half
ago
or
our
memories,
so
bad
that
we
have
drifted
off
on
to
other
things.
Hopefully
not,
but
now
is
the
point
to
to
check
in
with
them.
A
A
The
key
point
here
is
that
we
were
using
Jenkins
and
it's
gone
away.
It
was
difficult
to
maintain
it
and
it
wasn't
super
reliable.
So
it's
now
a
kind
of
choose
your
own
adventure
in
CI
world
and
the
current
recommendation
is
to
use
Travis
good
news
for
the
GUI
team.
Is
we
have
the
left,
but
we've
been
using
Travis
all
along
burr.
We
have
all
the
website
deploys
depend
on
Jenkins
who
had
the
token
to
update
the
DNS.
When
master
changed
on
github,
we
would
autumn
we
would
do
a
build.
A
Do
a
production
builder,
the
website
at
the
site,
route
to
ipfs,
get
the
CID
out
and
then
update
the
DNS
link
with
the
new
CID
to
to
cause
humans
to
be
able
to
see
the
new
website
that
doesn't
work
anymore.
So
we
need
to
fix
that
the
current
there's
a
conversation
with
Emperor
around
secrets,
secrets
management
and
the
best
way
to
do
that,
but
fYI.
If
you
push
some
changes
to
any
of
our
websites
right
now,
chances
are
even
if
the
PR
gets
merged,
you
won't
see
the
changes
go
live.
So
this
is
your.
A
C
So
we
pulled
out
Allan's
latest
version.
This
has
come
from
this
registry
rather
than
directly
from
the
NPM
standard
registry,
and
currently
the
new
pre-release
is
not
mirrored
on
to
the
ipfs
NPM
registry,
I
guess
because
no
one
has
pulled
it
yet,
and
so
it's
not
found
its
way
on
their
rows.
If
we
put
in
something
different,
then
it
should
eventually
come
up
with
a
CID,
and
you
can
click
on
this
and
go
look
on
Explorer,
but
it
never
successfully
resolves
to
anything.
C
A
A
One
thought
is
if
we
added,
if
we
connected
the
NPM
on
ipfs,
mirror
servers
to
the
Gateway
infrastructure
such
that
they
were
then
part
of
the
Gateway
nodes
swarm.
Then
you
could
just
open
the
directory
directly
on
IPS
bio,
slash,
ipfs,
slash
the
ID,
and
it
would.
When
you
look
for
blocks,
you
used
simultaneously
ask
your
connected
swarm
and
search
the
DHT.
So
if,
if
they
were
in
a
swamp
together
that
it
could
work,
you
could
link
through
to
the
directory
of
the
thing
that
you
want
to
see
cool.
E
A
You
are
on
the
money,
so
I
don't
know
it
seems
like
that.
Wouldn't
actually
be
hard
to
do.
We
just
have
to
get
the
data
there.
They
write
once
Alex
and
Andrew
have
done
all
the
work
to
make
the
thing
they
exist
and
linking
to
it
is
pretty
easy.
That's
the
special
magical
link
right.
You
get
all
this
values.
A
E
E
That's
awesome:
let's
see,
I
can
never
find
the
share
button.
Okay
here
it
is
okay,
so
I'll
hide
this
one,
so
demo
so
I
was
playing
around.
So
we
had
Kosala
eric
lee
on
the
call
in
january
and
he
demoed
Loonette
and
I
watched
it
and
I've
been
watching
what
he's
been
doing.
I've
been
reading
his
stuff
and
I
still
had
no
idea
what
it
was.
So
it
was
it,
but
there's
this
event
in
Portland
next
week
and
I
live
in
Vancouver,
so
there's
only
eight
hour
drive
and
I
tend
to
get
down
there.
E
E
So
I
really
wanted
to
figure
out
what
the
lunette
thing
was,
and
it's
got
this
the
crazy
thing
here
and
so
I
spent
a
Saturday
and
I
like
traced
for
you
and
figured
out
what
all
the
JavaScript
was
doing
and
it's
like
really
crazy.
But
so,
if
you,
if
you
don't
know
what
it
is
and
I
didn't
know
what
it
was,
it's
basically
a
way
of
taking
a
website.
E
That's
been
a
static
website,
that's
been
published
on
ipfs,
but
you
can
already
meet
you
buy
PFS
gateway,
which
is
just
giving
you
the
raw
files,
but
the
gateway
has
to
do
a
lot
of
work
like
the
gateway,
the
public
gateway
that
our
company
maintains
has
to.
You
know
we
have
to
run
all
these
ipfs
nodes
and
it's
we're
paying
the
bandwidth
cost
all
these
packets
flying
to
then
customers.
E
E
We
do
that
with
IP
FS
companion,
but
ipfs
companion.
You
have
to
install
an
extension
which
is
difficult
to
convince
people
to
do,
especially
if
they've
never
come
to.
If
they
don't
even
know
what
IP
FS
is
in
the
first
place
like
so
for
those
people,
it's
like
either
publish
on
the
public
gateway
where
protocol
Labs
has
to
pay
for
all
the
bandwidth
and
all
the
infrastructure
or
convince
them
to
install
the
companion
directly.
It's
been
sort
of
thinking
about
this
problem.
He
worked
on
extension,
extension
extensions
for
Firefox
4d
web.
E
But
again,
that's
like
installing
another
extension
he's
trying
to
think
like.
How
can
you
just
publish
a
website
and
people
can
go
to
it
and
it's
just
going
to
work
and
so
basically
would
work
by
doing
a
fallback
where
it
would
probably
probably
integrate
use
GSI,
PFS,
GCSE
IP
of
us
in
the
browser
we
go
and
ode
on
the
on
the
network,
but
we
really
want
people
to
run
their
own
demons.
E
So
so
it's
sort
of
it's
just
it's
just
another
way.
You
put
putting
this
all
together,
which
it
basically
is
going
to
people
you'll,
be
able
to
polish
a
website.
Everybody
in
the
world
be
able
access
it
on
their
phone
whatever,
but
the
power
user
so
gonna
be
able
to
render
ipfs
daemon
and
keep
everything
cached
locally
in
the
Builder
work
offline.
E
So
it's
sort
of
an
exploration
at
this
stage,
so
I,
don't
think
I
explain
this
well
anyway,
so
he
he's
put
this
thing
and
he's
got
it
installed
on
his
server
he's
got
a
dummy
here,
pier
diem
I'm,
not
going
to
run
it
again.
Cuz
he's
already
done
it,
but
what
I
just
wanted
it
install
it
myself
on.
My
machine
have
already
got
nginx
set
up
and
I
already
was
playing
around
with
getting
wildcard
DNS
certificates
from
let's
encrypt
and
so
I
went
once.
E
E
See
IDs
and
then
the
nice
thing
about
those
is
you
can
stick
them
in
a
subdomain,
so
so
how
lunette
works
is
when
you
go
to
lunette.
There's,
there's
two
file,
there's
only
two
files
and
they
have
to
be
served
up
from
on
this
domain
and
they
don't
contain
the
content
at
all,
so
they
can
be
completely
generated.
So
I
just
decided
to
create
a
little
no
GS
program.
E
E
Yeah,
so
this
is
everything
I
need
to
do
apart
from
the
nginx
setup
and
then
let's
encrypt
to
basically
have
you
know
my
own
equivalent
of
ipfs
gateway,
except
I,
don't
have
to
run
IVFs
myself,
I
might
set
my
end
like
it's,
because
lunette
is
rescuing
the
users
to
run
I,
invest
themselves
or
maybe
don't
be
the
GSI
PFS
for
that.
But.
A
A
A
E
E
E
Okay,
so
that's
just
an
HTML
file
with
the
pre
tag
and
whatever
I
type.
So,
though
that
was
instant
publishing,
which
is
nice,
but
you
know
the
thing
with
IIP,
FS
or
dad,
or
anything
like
I
built
instant
publishing
for
dad
before,
but
you
need
pinning
like
he
can't
just
like
publish
like
like
people
will
need
the
files
to
stick
around
so
this
magic
going
on
here.
So
there's
I
built
an
iframe
behind
this,
so
there's
iframe.
E
You
can't
see
it
and
says
when
I
said
pinning
stash
pending,
so
I'll
show
you
the
iframe
here
and
there's
actually
iframe
here,
but
you
can't
see
it,
but
what
I
did
was
I
put
peer
base
into
the
iframe
so
peer
basis.
The
project
I've
been
working
on
since
September
in
Pedro
script
and
that
allows
you
to
synchronize
things
peer-to-peer
from
web
browsers.
So
this
is
actually
made
different.
E
This
is
the
origin,
that's
the
iframe,
so
but
the
problem
with
before
she
could
like
run
them
both
like
if
you
ran
to
peer
bases
in
different
tabs.
That
would
just
be
really
really
bad.
You
couldn't
deal
with
that.
Firstly,
I
went
to
Tokyo
in
November
and
I
met,
Andre
septic,
and
you
might
know
him
from
post
CSS
and
he's
built.
This
thing
called
log
ox,
which
is
inspired
by
some
work,
that
Viktor
Khristenko
is
doing
who's,
also
like
something
it's
not
under
team.
Everything's
connected
this
crazy.
E
These
are
all
the
same
web
browser
and
you
can
see
this
one
has
a
role
of
leader.
This
one
has
a
follower,
so
the
one
that
is
an
iframe
here
is
also
a
follower.
So
I
can
you
can
see
the
browser
election
to
have
a
little
button
here
to
hit
resign
and
you
can't
see,
but
the
leader
is
probably
the
one
that's
hidden
in
my
frame
here
and
everything's
talking
so
I.
Can
click
can
send
messages
tea
to
each
other,
but
also
like
when
I
pinned
here.
E
It's
using
logics
to
send
this
and
then
only
in
the
only
end,
the
one
that's
the
leader,
that's
the
one.
That's
running
your
base,
and
so
in
the
leader
election
happens,
the
the
one
that
becomes
the
leader
starts
up
your
base
and
for
some
reason
it
loses
it
another
one
I'll
come
up
and
pick
it
up
so.
E
E
E
E
Pitts
there's
a
lot
of
stuff
there,
but
all
the
end-user
has
to
see
is
this
and
there
they
could
publish
way
in
this
technique.
Anybody
can
just
go
crazy
and
just
publish
stuff,
so
they'll
have
to
have
the
pitter.
You
know
configured,
you
know
in
the
iframe
with
the
robot
or
something
that's
listening
care.
That's
actually
pinning
the
content,
the
stuff
that's
in
the
iframe.
This
could
be
all
based
on
like
private
user
preferences,
so
users
could
bring
their
own
Pinner
bring
their
own
gateways.
It's
this
sort
of
technique,
but
synchronize
it
between
multiple
devices.
E
E
Yeah
so
encourage
people
to
try
it
so
starting
off
my
iMac
right
now
so,
but
the
thing
is
the
nice
thing.
Is
there
early
isn't
much
running
on
my
iMac?
The
problem
with
lunette
is
right.
Now
it
only
works
if
you
have
ipfs
daemon
running
locally,
and
you
also
have
to
have
course
headers
set
up
on
it.
So,
but
if
you
have
those
two
requirements,
it
might
work.
E
A
B
Just
like
a
brief
highlight
that
this
election
couldn't
be
used
in
multiple
context.
When
you
have
J's
ipfs,
you
know
browser.
Let's
say
you
you,
you
don't
want
to
use
local
users.
Note
you
just
want
to
have
jsap
FS
on
your
website,
but
but
your
users
have
a
habit
of
opening
multiple
tabs
of
your
website.
You
basically
could
solve
the
problem
of
having
multiple
nodes,
multiple
instances
of
J's
ideas
for
the
same
origin.
B
B
So
it's
like
useful,
not
only
in
the
context
of
the
this,
like
access
point
based,
serviceworker
experimentation
that
you
luckily
is
doing,
but
even
like
solving
the
this
separate
small
problem
for
people
who
wants
to
run
JCAP
FS
in
their
own
app
that
that
saves
the
resources
in
multiple
tasks.
So
it's
my.
A
B
A
B
B
A
Summer,
the
summary
is
IP
of
s
companion,
bundles
web.
You
are
directly
in
it
so
that
when
you
run
it
on
your
Firefox,
if
you
happen
to
be
offline,
you
can
still
see
the
web
UI,
which
is
a
mobile
goal,
but
because
archivist
web
UI
is
a
sloppy,
build
and
not
necessarily
reduce
reproducible,
regardless
of
operating
system
and
environment.
It's
now
cause
us
to
be
kicked
off.
The
firefox
add-on
store,
so
I
think
we
should
definitely
remove
web
UI
from
the
companion
yeah.
That
was
that
was
the
butt
of
that
conversation
yeah.
So.
B
Like
the
small
small
regression
would
be,
we
would
not
be
able
to
use
it
in
embedded
J's
ipfs
unless
we
restore
access
to
API
over
window
ipfs.
We
may
like
look
into
that
again.
We
could
probably
like
Whitely
specific
cid
path
and
expose
an
sandbox
version
of
api,
like
we
could
just
hardcore
specific
version
anyway.
Yeah,
so
we'll
continue
this
discussion.