►
From YouTube: Updates on Companion, Desktop and Web UI - IPFS GUI and In Web Browsers Weekly, 2019-08-14
Description
IPFS Mirror: https://ipfs.io/ipfs/bafybeihroplzc5z52nq4cxxrz5ztjjtarmyd5sfihtagboslbsitaawuuq/
A
A
B
So
a
quick
update
on
under
the
jcpoa
facing
brave
he
landed
in
the
stable
channel
and
the
improvements.
Are
it
if
you
switch
to
ember
that
jess
ipfs
in
brave?
It
will
automatically
pick
check
if
the
party's
available
and
if
not
it,
will
pick
the
next
one.
So
there
are
no
port
collisions,
it
will
automatically
persist
external
node
config.
B
What
the
problem
with
member
the
J's
idea
first
in
brave,
is
that
we
don't
have
tcp
transport
yet
so
you
are
not
able
to,
and
we
don't
have
the
HD.
So
we
are,
unless
we
are
already
connected
to
a
node
that
has
data
you
want.
You
want
to
be
able
to
get
the
data
because
you
were
not.
You
won't
have
any
means
of
finding
peer
ID
of
people
who
have
that
data
and
what
delegated
peers
and
content
routing
does.
It
basically
gives
you
DHT
lookup
capabilities
without
running
TCP
or
DHT
ingest.
B
B
Yeah,
so
this
is
brave,
knightly
I,
think
in
settings
we'll
go
to
the
very
bottom
and
their
sacrifice
component.
It
landed
in
the
stable
channel,
so
we
just
enabled
here
and
it
should
just
work
yep.
So
I
have
no
idea
us
running
locally,
so
it's
already
connected.
But
if
I
go
to
settings
and
if
I
switch
to
embedded
Plus
Chrome
sockets,
you.
B
B
Or
maybe
it's
too
slow
for
now
yep
so
anyway,
why
delegated
to
think
is
important
is
because,
if
I
was
not
connected
to
anyone
with
webui
before
I
was
not
able
to
fetch
it
and
now,
if
it
there
was
a
person
in
the
swarm
that
has
this
content.
I
want
I
can
get
its
PID
and
their
addresses
using
delegated
routing,
and
also
when
I
add
stuff,
oh
gosh,
when
is
it
him
so
if
I
add
stuff
to
ipfs
now.
B
It
was
added
to
my
local
node
and
it
it
loaded
from
remember
that
node
and
what's
interesting
now,
is
that
if
I
add
stuff
to
my
local
node
I
want
other
people
to
be
able
to
hedge
this
data
from
me,
however,
I
don't
have
public
IP
in
web
browser,
yet
so
this
is
where
delegated
routing
helps
when
I
add
stuff
to
my
local
node.
He
delegated
routing
automatically
asks
delegate
nodes
to
prefetch
the
content,
I've
just
added,
and
then
that
node
will
make
the
content
available
on
the
network.
B
B
So
that's
more
or
less
than
update
what
I
want
to
talk.
Don't
mention
around
this
I
mother,
Jessica,
fussing
brave
is
that
the
delegated
peers
and
content
routing
does
not
fully
solve
the
discovery
problem
if
I
want
to
open
if
I
had
staff
it
gets.
Added
to
the
network
and
network
is
notified
about
that.
However,
if
I
want
to
open
things
that
are
new
and
opened
them
from
the
Gateway,
that's
when
it
gets
tricky,
because
my
my
oh
gosh,
so
my
gate,
my
gateway,
ranking
in
brave.
B
It's
not
able
to
fetch
this
content.
It
has
no
peers
that
are
capable
of
talking
to
offer
what
WebSockets
and
we
are
discussing,
adding
preload
for
all
NFS
operations,
I
believe
it
did
I
linked
that
yeah.
So
that's
like
the
second
link
under
delegated
content
and
purity
in
mythic
notes.
The
idea
is
that
if
we
we
request
content
from
embedded
gateway,
we
would
promote
new
speed
of
nodes
to
preload
that
content
to
pull
out
node
and
that
way
the
embedded
Jacek
give
us
will
be
able
to
open
it.
B
That's
just
a
stopgap
until
we
have
a
peer-to-peer
transport
and
the
TCP
transport
will
probably
be
the
next
thing.
I
will
be
working
on
as
soon
as
I
finish
up
some
other
stuff.
I
realized.
It
was
bit
chaotic
today,
sorry
about
that,
but
maybe
there
are
some
questions
that
will
clarify
clarified
only
and
then
Jim
the
delegate
knows
who's
looking
after
those
I
believe
right
now
those
have
the
same
status
as
like
delegate
nodes
and
preload
nodes
have
the
same
status
as
WebSocket
star.
B
B
And
for
the
record
by
default
delegate
nodes
are
not
enabled
in
Jessica's
only
preload
nodes.
So
it's
like
a
messy
situation.
We
probably
need
to
clean
up
this
API,
but
for
now
we
have
preload
nodes
which
actually
are
delegate
nodes
but
have
different
host
names
yeah,
so
host
names
are
different
but
actually
point
at
the
same
servers.
Okay,
that
makes
more
sense.
Yeah.
B
Like
everyone's
game,
it's
available
I'm,
pretty
sure
in
brief,
the
brief
nightly
and
probably
better
and
DEF
channels,
I'm,
not
sure
if
it
already
landed
in
I,
mean
I'm,
not
sure
if
the
toggle
in
settings
screen
landed
in
the
stable
channel.
Yet,
however,
if
you
have
that
toggle,
it
means
you
can
just
install
it,
and
the
toggle
will
always
install
the
stable
version
of
Eclipse
companion.
B
B
Honestly
I
believe
we
may
just
disable
the
legates
before
that
happens.
My
plan
is
to
land
like
kill,
two
big
transport,
and
then
that
should
also
like
solve
or
at
least
get
us
to
the
better
place
when
it
comes
to
content,
discovery
and
pure
discovery
and
remove
the
need
of
both
preload
nodes
and
delegate
nodes,
because
we
would
need
to
explicitly
disable
preload
with
it
would
like.
If
a
lot
of
people
start
running,
that
it
would
just
hammer
our
preload
nodes,
which
are
already
not
the
best
shape,
agree
that.
B
So
right
now
the
WebSocket
transport
is
the
only
one
we
enabled
in
embedded.
Note,
however,
like
the
plan
is
to
either
expert
either
I
would
not
invest
time
or
in
WebRTC
at
least
not
in
brave
cause
in
brave.
The
more
more
effective
way
is
to
spend
some
time
and
get
like
actual
TCP
transport
and
that's
the
plan,
so
we
would
still
have
both
WebSockets
and
TCP
embrace
that's
a
plan.
It's.
E
B
B
E
A
A
Would
I
would
love
to
have
a
visualization
of
all
of
our
connection,
strategies
that
we
use
and
how
we
cascade
and
fallback
to
different
strategies
in
different
environments?
I'm
not
sure
it
would
ever
reflect
reality
would
probably
be
permanently
out
of
date,
but
wow.
It
would
be
very
helpful
to
understand
that,
especially
for
people
trying
to
debug
where
failures
are
happening.
So,
if
they're
trying
to
publish
something,
they
don't
know
this
entire
layout
that
the
topology
of
this
is
pretty
complicated.
Then
it's
done
very
few.
People's
heads,
yeah.
B
Like
right
now,
we
in
ingest
ipfs
with
me,
we
have
a
section
about
setting
up
web
sockets
are
or
section
about
WebRTC,
and
that
honestly
should
be
like
replaced
with.
The
thing
you
describe
like
here
here
are
available
options.
Here.
Is
the
state
known
problems
for
people
to
maybe
pick
up
issue
or
dig
into
yep.
A
F
Okay,
so
remove
this
here
we
have
the
new
ipod,
SS
top,
which
comes
oh,
this
might
not
be
better
yeah,
but
it's
from
us,
whatever
we
have
like
now.
The
these
two
menu
options
here,
which
are
can
be
used
for
these
actions
that
we
already
had
before,
but
we
could
only
use
them
by
pressing
the
Rockets
Monken
with.
Indeed
here
you
can
also
move
the
repository
location.
Some
people
don't
want
to
put
it
on
their
own
folder
and
you
can
also
have
like
a
co-hosted
websites
list
which
gets
updated
every
12
hours.
F
F
F
Okay,
strange
you
already
had
like
a
demo
last
week,
but
remembering
the
main
feature
is
that
you
can
now
never
can
normally
get
to
all
of
the
ipfs
or
IPS
universe.
While
you
already
found
a
bug,
you
can
also
check
the
pins,
and
basically
that's
mainly
it
I
would
like
to
know
if
you
could
like
install
it
in
your
system
check
if
there
are
any
bugs
before
we
can
release
the
final
version.
D
F
F
Hopefully
it
isn't
slowly
now
it
will
take
a
while
I,
don't
know
if
I
should
like
talked
about
the
last
byte
energy
in
the
agenda,
which
is
related
type
in
essence
of
what
do
you
think
we
don't
know?
What
that
point
is
you
should
just
talk
about
it,
yeah
the
cycle
of
the
tree.
I.
Couldn't
we've
talked
about
it
before?
How
can
I
hide
this
like
zoom
bar
it's
a
night,
we
don't
say,
though
yeah
I
know,
but
I
see,
and
it's
all
right.
It's
in
the
middle
of
the
screen.
Okay
Erica
said
like
a.
F
It
doesn't
have
much
time
to
work
on
this
I
really
enjoyed
like
the
styles
of
the
icons,
my
quest,
which
I
think
it's
better
than
having
like
colored
I
can
which
doesn't
match
any
of
the
others.
The
problem
is
on
Windows
and
other
platforms
or
where
we
can't
detect
if
the
user
as
a
white
or
dark
mode,
so
the
eye
constant
gets
very.
F
F
D
It
feels
like
there
desire
to
blend
in
with
the
existing
icons
to
go
monochromatic,
but
if
that
causes
problems
on
windows,
then
I
would
just
pause.
Work
on
this
endeavor.
Until
we
have
a
better
one
there
like
it
doesn't
I,
don't
think
it's
causing
us
any
problems
and
existed
the
existing
colorful
one
works
in
dark
mode
and
light
mode
on
all
operating
systems.
I.
B
Ideally,
we
would
ship
this
in
both
ipfs
desktop
on
all
platforms,
but
also
like
ipfs
companion,
at
least
in
Firefox
I,
believe
the
browser
icons
there's
like
a
recent
capability
of
detecting
night-night
mode
by
the
browser
extensions.
So
at
least
we
could
detect
that
and
swap
the
icon
when
the
user
has
the
night
mode
or
maybe
swap
of
light
and
dark
icons
in
Firefox.
B
B
Maybe
maybe
let
me
clarify
I
think
that
in
the
in
the
issue
you
are
showing
I
commented
that
the
key
base
and
Dropbox
icons
are
a
good
examples,
because
they
are
not.
They
are
actually
the
logo
of
the
project
simplify
to
the
extreme.
So
it's
immediately
obvious
what
it
is
when
you
look
at
it
and
I
would
like
the
design
proposed
by
Eric
with
those
like
connecting
dots
and
other
stuff,
but
that
I
like
that
in
a
bigger
size
in
the
small
size,
I
just
have
no
idea
what
that
it's
just
a
small.
B
Maybe
it's
the
problem
with
my
laptop,
because
I
ran
in
even
higher
dpi
than
other
people.
However,
like
it's
just
like
it's
just
like
strange
square
for
me,
there's
all
those
details
that
it's
just
too
complicated
I
feel
I
feel
like.
We
should
have
super
simple,
obvious:
ipfs
cube
in
black
and
white.
B
B
B
B
So
that's
the
tricky
part.
We
don't
have
many
options.
You
can
like
feel,
feel
their
walls
a
little
bit
with
some
shader
or
not.
So
that's
like
an
open,
open
problem.
It's
a
tricky
thing
and
I
like
I.
Actually
I've
been
thinking
about
it
and
I'm
like
what
what
he
eats
up
this
film
to
me
that,
like
both
Dropbox
and
others,
that
they
spent
some
time
on
those
icons,
there's
like.
B
F
D
B
A
A
question
about
the
language:
it
feels
like
like
there's
that
so
first,
this
is
awesome.
This
is
pretty
amazing,
because
I
think
it
gets
pretty
close
to
what
people
think
about
when
they
don't
know
about
ipfs.
Well,
they
think
ipfs
might
mean,
but
I'm
curious
about,
like
co-hosting
I
feel
like
there's.
We've
talked
a
little
bit
about
people
who
want
to
altruistic
Allah,
keep
sites
alive
for
other
people
and
then
there's
the
I
want
to
read
sites
offline
and
those
are
really
different
ways
of
thinking
about
why
you
would
co-host
different
use
cases
for
it?
A
D
D
D
D
D
B
F
B
Yeah
so
I'd
either
added
like
sub
sub
menu
or
other
experiments.
If
you
want
to
keep
it
as
is
or
just
like
hide
it
until
it
lands
and
UI
in
settings,
you
know
what
you
I,
maybe
that's
like
the
list
of
house
names,
it's
pretty
it's
not
something
we
could
make
a
better
UI
and
then
like
text
editor,
probably
and
people
would
probably
appreciate
that
and
have
like
a
feature
itself-
would
have
a
bigger
engagement
from
users.
F
H
Will
be
keen
to
suggest
this
goes
along
with
inspects,
experiments
are
already
there
because
then
we
can
have
some
analytics
to
see
when
people
have
enabled
or
disabled
things,
then
we
might
need
additional
UI
or
Chrome
just
to
actually
list
out
the
inputs,
but
yeah
I
feel
like
it's
like
it
lives
there
or
we
just
stop
promoting
that
section.
A
bit
more
now.
A
D
G
G
B
G
F
Yep
like
because
here
you
can't
like
at
these
folders
from
business
or
any
folders,
but
on
some
it.
Oh,
it
only
works
on
in
the
file
section
it's
inside
the
directory.
If
you
open
like
a
file,
it
wouldn't
work.
So
you
have
the
other
options
related
to
files,
and
the
same
applies
to
anything
outside
of
the
file
section
of
your
own
trust.
F
D
So
maybe
we
should
pause
on
the
design
review,
but
I
think
I
think
what
this
is
pointing
to
is
a
lot
of
the
feedback
that
has
come
up
in
the
last
few
minutes.
There's
is
feedback.
Those
come
up
on
pull
requests
previously,
that
these
exact
issues
about
the
color
of
the
button,
the
pins
link,
not
looking
like
a
button,
I
think
there's,
there's
probably
a
missing
step
in
our
process
whereby
there
is
a
design
review
before
we
cut
a
release
of
web
UI
and
there's
fewer
people
focused
in
on
lucky
right
now.
A
B
B
First
term
we
use
a
sweet,
direct
and
redirects
IP.
Fest
companion
started
as
browser
extension
that
the
text
ipfs,
that
the
text
request
for
ipfs
resources
and
redirects
them
to
your
local
node,
and
it
was
like
very
technical
extension
and
we
just
kept
that
language
and
it's
just
retire
right
through
directs
and
we've
added
other
options
and
there's
like
redirect
this
redirect
dots
and
right
now
the
UI
looks
like
this.
You
have
a
redirect
to
Gateway
and
then
for
active
tab.
B
You
can
toggle
the
redirect
for
this
specific
domain,
but
it's
like
unclear,
which
has
a
precedence
and
stuff
like
that.
So
there's
like
more
details
on
this
and
some
proposed
changes,
basically
like
just
dropping
the
name
and
going
with
different
language,
like
prefer
local
note
or
try
local
note
first,
mainly
because
at
some
point
we
will
have
an
native
support
in
browser
vendors
or
a
native
protocol
handler
and
browser
extension
and
effectively.
We
will
no
longer
like
redirect,
but
we
will
like
provide
a
native
handler
which
just
handle
handle
is
the
request.
B
So
will
the
redirect
will
no
longer
make
sense
as
a
language
provided
to
the
user,
and
another
thing
is
sorting
outs
than
language
around
toggle,
more
granule
control
when
we
redirect
power
website.
So
it's
this
toggle
was
added
as
a
means
of
just
disabling.
This
redirect
on
a
specific
website,
but
actually
we
do
more
than
just
redirect
and
what
people
effectively
want
is
to
just
like
don't
run
ipfs
integrations
on
this
website.
B
So
that's
like
change
of
language
around
that
and
some
proposed
terms
are
here
and
basically
I
would
appreciate
some
feedback
on
that,
probably
on
the
issue,
so
that
others
can
like
participate
and
the
second
one
is
spinning.
The
problem
is
sort
of
similar
pinning
is
a
low-level
term.
We
have
in
IPA
score,
however,
when
you
pin
stuff
you
just
pin
the
content
identifier,
the
cid,
you
are
not
able
to
my
other
lab
label
for
that,
and
you
basically
have
like
I
have
800
of
pins
and
I
have
no
idea
what
what's
what
so
the
idea.
B
The
plan
is
to
basically
stop
using
this
low
level
pinning
mechanism
and
instead
just
copy
files
to
em
FS.
So
when
people
are
able
to
rename
them
and
manage
them
using
web
UI
in
the
file
screen-
and
the
second
thing
is
that
people
will
have
better
control,
they
can,
if
something
is
added
to
MF
s
is
implicitly
pinned,
so
it
won't
be
garbage
collected.
And
if
someone
removes
things
from
em
fast,
then
it
can
be
garbage
collected
again.
B
So
that's
a
underlying
reason
why
we
want
to
go
away
from
using
pinning
API
and
if
we
stop
using
pinning
API,
it
makes
sense
to
stop
using
pin
as
a
term
in
our
user
interface,
to
avoid
confusing
people
and
like
mixing
different
meaning
at
different
abstraction
levels.
And
another
thing
is
that
at
some
point
we
will
have
like
a
DNS
linked
website.
We'll
have
something
similar
to
what
hogs
just
demoed
the
co-hosting
a
website.
B
You
will
be
able
to
like
go
to
a
DNS
linked
website
and
you
will
be
able
to
like,
like
team,
but
not
being
like
co-host
this
website
and
that
would
effectively
bookmark
the
website
and
that
would
be
a
life
bookmark
powered
by
DFS.
So
when
you,
your
IP
first
note
would
follow.
Ip
NS
updates
or
DNS
link
updates
and
fetch
prefetch,
a
new
version
of
stuff
like
that.
So
that's
like
motivation
behind
those
above
proposals
to
change
language
and
if
anyone
has
any
strong
or
even
weak
opinions.
A
A
B
I
kept
as
desktop
well
every
time
web
UI
is
released
and
my
question
is:
would
it
make
sense
to
change
this
process
and
unburden
web
UI
maintainer
because,
like
we
don't
expect
library
maintainer
to
PR
every
project?
That's
using
library,
it's
the
other
way
around.
So
that's
my
proposal
to
just
go
and
agree
with
go
and
Jas
and
other
projects
so
that
they
have
a
checklist
item
on
and
they
check
if
there's
a
newer
version
of
web
UI
that
way.
Web
UI
maintainer
does
not
need
to
worry
about
that.
D
B
A
D3I
I
would
like
to
provide
a
counterpoint,
I
feel
like
the
person
that
the
making
decisions
around
and
other
projects
resourcing
it
isn't
going
to.
It
will
always
keep
up
in
that
that
that
amount
of
resources
that
is
unaligned.
So
we
have
explicitly
said
that,
right
now,
the
desktop
GUI
are
lower
priority.
If
the
project
with
the
higher
priority
has
an
explicit
dependency
in
need
and
the
onus
should
be
on
them
to
own
that
need.
A
I
also
worry
that
they
will
chip
that,
if
that
dependency,
I
guess
may
be
taken
seriously,
isn't
the
right
way
to
describe
it,
but
that
they
actually
have
a
hard
dependency
on
web
UI
working,
and
this
is
kind
of
where
the
discussion
about
the
test
matrix
between
web
UI
and
desktop
and
go
and
JSI
PFS
and
those
all
working
in
concert.
All
the
time
was
around
elevating
that
understanding
of
that
hard
dependency
that
we
can't
say
something
like
what
UI
is
deprioritized
when
go
ipfs
has
to
ship
with
it.
A
So
I
would
like
to
actually
make
those
dependency
is
more
clear
and
formal
and
make
go
at
the
FS
than
gia.
So
if
s
have
that,
even
even
if
we
don't
change
our
web
UI
process
and
it's
still
our
process
to
a
PR,
those
projects,
I
would
still
really
really
encourage
those
projects
that
have
a
dependency
on
web
UI
to
have
that
web
UI
checkbox
in
their
release
process,
especially
now
that
they
go,
has
a
release
process
because
it
reduces
it,
puts
the
the
responsibility
on
both
sides
and
formalizes.
A
B
Something
I
realized
because
I'll
have
a
good
point.
That's
like
go
up.
Your
first
teams
are
already
overburdened
and
actually
we
may
have
more
webui
releases
than
go
ipfs
releases
and
if
we
make
a
PR
to
go
ipfs
each
time
we
make
a
web
UI
release.
We
actually
introduce
like
overhead
of
multiple
PRS,
but
that's
just
like
a
second
club.
If
they
have
a
checklist,
they
just
check
the
moment.
They
make
a
release.
A
And
we
probably
want
to
get
to
where
they're
pulling
in
official
releases
automatically
and
their
tests
break
if
they
pull
in
a
release
and
web
and
they
they
pull
it
a
release
in
run
web
UI
tests
in
it
and
something
breaks
that
should
be
a
kind
of
a
halt.
Hunting
State
for
both
sides,
but
yeah
I,
don't
think
we're
there.
Yet
I
feel
like
at
least
adding
it
to
the
checklist
gets
us
like,
like
baby
step
towards
that
world.
B
And
something
to
remember
is
that
we
are
planning
to
support
loading.
The
latest
version
of
web
UI
from
IPA
from
DNS
link
path,
so
that
could
be
something
using
tests
as
well
for
like
for
this.
Automated
testing.
You've
just
mentioned
that
person
ongoing
work
to
support
DNS
over
HTTP,
but
then
we
should
unpack
that
yeah
only
just.
A
Way,
I
think
iPod
post
was
really
astray,
so
we
have
five.
We
have
five
more
minutes,
it
sounds
like
we.
We
have,
maybe
maybe
a
decision
interesting
towards
adding
new,
at
least
having
it
on
the
list.
But
what
do
I
process
doesn't
change
to
PR?
The
three
projects
should
the
next
I
remodel.
This
is
a
PSA
yeah.
B
It's
a
also
very
quick
one
stiffen
at
the
TLS
multi
colic.
So
now
we
can
have
like
slash
TLS
in
multi
others.
The
problem
is
that
we
like
HTTP,
is
ftp
in
wrapped
in
TLS,
but
we
already
had
like
slash
HTTP
and
in
web.
We
use
it
in
our
you
ice
for
things
like
ipfs
api
address
and
we
show
it
in
multiple
places.
B
So
slash,
HTTP
is
already
used
and
there's
like
an
ongoing
discussion
on
supporting
/
TLS,
slash
HTTP
as
an
equal
one,
and
then
there's
like
an
open
question
which
should
be
the
canonical
version,
because
we
have
like
libraries
that
convert
urls
to
multi
others
and
multi
others
to
urls
and
at
some
point
we
need
to
decide.
So
if
anyone
has
any
strong
opinions
comment
on
the
issue
I
linked
and
there's
a
call
tomorrow.
If
someone
who
is
really
passionate
about
this,
that's
that's
the
announcement.
A
Thank
you.
How
that
we
talked
about
the
icons
earlier
last
item
on.
There
is
a
question
that
I
added,
if
there's
an
issue
tracking
with
the
known
differences
and
go
ipfs,
NGS
ipfs
for
running
as
a
server
are
I
was
thinking
about
doing
something
like
adding
those
differences
in
a
table
or
something
like
that.
I,
don't
I
had
talked
about
previously
somewhere,
so
that
people
understand
what
the
difference
is.
A
So
if
you
have
and
know
of
any
issues
that
track
those
specific
differences
or
if
there's
a
meta
issue
that
has
a
list
of
all
those
differences,
please
add
it
to
the
notes
and
we
can
do
more
things
with
it
and
harness.
Even
you
know,
1%
of
those
12
million
developers
to
come,
make
chess
ipfs
better.