►
From YouTube: GUI and In Web Browsers Weekly, 2019-01-23
Description
IPFS Mirror: https://ipfs.io/ipfs/bafybeihko5yuyt4stygselhbji6yw2y7mqppvlpwafqlhtrjgix5sltypu/
Meeting Notes: https://github.com/ipfs/team-mgmt/pull/846
B
In
that
case,
welcome
to
the
ipfs
weekly
in
GUI
and
web
browser.
Sinkholes
is
tradition.
We
all
go
and
stare
at
the
agenda
for
a
second
and
see
if
anyone
has
any
agenda
points
to
act
which
will
beseech
the
audience
for
a
most
taker.
Anyone
available
anyone
stepping
forward.
Okay,
Jim
pick
thanks,
Jim
and
also
take
a
moment
to
add
any
agenda
items
that
you
would
like
to
anything
in
these
integration
teams
purview.
B
A
So
sharing
my
screen
last
week,
a
stable
release
of
ipfs,
combined
extension
with
new
web
UI
weave
window
IDF
has
v2.
So
without
the
first
enable
mat.
It's
a
sink
method
that
lets
you
ask
for
multiple
permissions
in
bulk.
It's
ready
to
play
it
with
it
in
wild.
It's
involved
stable
and
data
Channel
and
yeah.
A
That's
ready
for
install.
We
had
the
discussion
with
brave
and
there's
like
a
post
with
plan,
how
we
can
speed
up,
bringing
ipfs
to
brave
by
avoiding
shipping
ago
ipfs
binary
and
writing
all
the
orchestration
related
to
that.
Instead,
if
we
could
empower
JSI,
TFS,
no
chronic
inactive
as
companion
to
have
similar
powers,
namely
expose
HTTP
gateway
using
some
more
powerful
IP,
is
that
our
extension
could
get
access
to
so
plumb
and
pros
and
cons
for
both
approaches.
A
It's
sort
of
long
but
I
try
to
make
it
a
good
case.
Y,
J's,
IPF
asthma
may
be
ready,
or
at
least
might
be
ready
at
some
point
in
2018
and
there's
like
a
related
issue
about
those
hidden
by
default
in
chromium
hidden
by
default
in
like
Google
Chrome.
But
if
we
get
a
blast
extension
status
in
brave,
we
could
gain
access
to
those
api's
and
have
things
like
exposing
cases
be
gateway
in
similar
fashion,
but
Co
IP,
first
us,
and
also
like
the
local
discovery
and
TCP
transport.
Similarly
to
develop
experiments.
A
We
did
in
the
past
year.
Things
like
that
on
the
cid
and
by
32
front
we
have
like
public
gateway
is
coming
along
nicely.
I.
Think
Kyle
posted
some
updates
in
this
main
issue
yeah,
without
that,
like
general
summary
from
NGO
activist
and
Jessica,
first
companion
as
well,
compile
the
text
series
in
subdomains
and
displays
page
actions
dozen
related
issue.
Why
I
mention
it
is?
How
should
we
handle
websites
loaded
like
with
CID
in
subdomain?
The
thing
is
that
if
we
load
website
that
using
CID
in
subdomain,
then
this
website
gets
security
and
isolation.
A
A
So
the
question
is:
how
should
we
handle
CID
in
subdomain
website
and
that's
sort
of
a
similar
case
within
DNS
linked
websites?
Because
if
you
have
your
own
domain,
then
you
have
origin
by
security
parameter.
But
if
we
redirect
that
website
to
local
Gateway,
then
it's
just
on
the
IP
FS
/
domain
path,
which
shares
the
origin
with
other
websites.
So
there
are
different
solutions
to
that.
Some
are
sort
of
wishful
thinking
right
now.
A
What
we
could
do
is
to
display
page
actions
just
like
we
do
for
like
right
now,
right
display
that
this
website
was
loaded
with
public
gateway
and
what
we
could
do
is
to
add
additional
option
below
to
either
open
that
website
on
local
gateway,
but
that
would
be
an
opt-in
like
conscious,
opt-in
by
the
user,
or
we
could
like
opt-in
for
automatic
redirect
so
from
for
this
website.
It's
fine
to
redirect
I'm
open
for
suggestions.
If
anyone
has
a
better
idea
or
a
way,
we
could
handle
it
more
gracefully.
A
The
issue
a
linked
will
be
great
and
that
sort
of
it's
on
Maya
and
just
like
to
finish
up
and
I,
take
questions
on
the
next.
In
the
next
week,
I
want
to
I
have
a
like
a
baked
in
PR
about
about
opting
out
from
redirects
for
specific
websites,
namely,
let's
say
on
the
top.
If
I
choose
companion,
Iranian
can
be
directing
it
may
break
the
website.
A
People
should
have
a
possibility,
just
like,
in
new
blog
to
opt
out
from
from
IOT
FS
compiled
italic
on
the
for
a
specific
website
and
that
sort
of
related
to
this.
In
this,
like
discussion
about
subdomains,
we
would
like
to
opt
in,
and
there
should
be
like
also
like
opt-out.
So
that's
like
something
I
would
do
like
to
figure
out
in
the
following
week,
and
also
maybe
if
time
allows
go
back
to
brave
and
then
stop
sharing
and
take
questions.
B
B
C
C
C
Sound
all
the
sea-ice,
all
my
prototypes
are
failing,
because
it's
timing
out,
okay,
I,
don't
think
it's
any
anything
with
my
clothes.
I
also
reply
to
the
effector,
the
drag
and
drop
to
the
whole
page,
because
we
could
only
drag
and
drop
files
and
folders
to
to
the
phallus.
When
we
only
have
one
file,
we
will
just
have
to
drop
them
in
here
now
we
can
drop
even
another
pages,
for
example
in
status,
drag
and
drop.
C
C
Because
we
have
this
overlay,
then,
when
we
drag
so
the
file
lists,
get
a
bit
of
a
penalty
and
then
highlight
the
folder,
and
if
you
don't
like
this,
I
have
to
think
about
other
way
to
do
this,
but
it's
working
I
will
make
that
in
another
four
sites
for
now
I
think
back
and
it's
not.
Let
me
have
a
high
priority
right
now.
Agree.
C
C
C
B
The
most
interesting
thing
for
GUI
world
this
week
was
actually
having
just
enough
time
and
enough
credentials
to
push
through
to
get
signed.
Installers
of
the
ipfs
desktop
apps
on
OS
X
and
oh,
my
oh,
my-
is
that
it's
surprisingly
complicated
for
a
for
a
setup
that
you'd
think
would
you
definitely
want
to
happen.
B
These
are
now
like
this
is
the
distillation
of
the
process
that
I
added
and
there's
various
like
UX
Horrors
in
there,
the
problem
being
we're
building
an
electron
app
and
we
want
to
apply
the
certificates
that
we
created
during
the
continuous
integration
process.
So
the
challenge
here
is
that
I
think
a
lot
of
the
UX
improvements
around
signing
apps
has
gone
into
Apple's
Xcode
app
and
we
are
not
using
that
we
are
using
javascript
an
electron
of
any
weight.
So
there
was
them.
B
B
So
I've
kicked
I've
incremented
the
version
I'm
just
one
thing
that
hadn't
quite
thick
for
me
with
the
gif,
its
desktop
release
president's
was
the
electron
builder
and
it's
quite
interesting
and
it
runs
in
CI,
builds
out
your
artifacts
on
every
commit
semester,
and
if
there
is
a
draft
release
on
github
releases,
that's
already
been
created,
then
it
will
publish
the
latest
version
of
the
installers
to
that
draft
release.
And
so
then
every
built-in
master
will
just
change.
It'll,
just
replace
that
artifacts
from
the
draft
release.
B
Until
you
hit,
publish
the
release
and
then
those
artifacts
frozen,
and
then
then
you
bumped
their
number,
so
you're
kind
of
always
like
this
we're
building
a
lot.
Seven.
Now
we're
not
just
a
number
in
the
package
version,
but
we
haven't
committed
to
which
particular
commit
we're
going
to
tag
as
I've
got
seven,
so
it's
the
opposite
way
around
and
how
we
do
a
lot
of
other
projects
where
we
don't
spawn
a
certain
number
until
we
are
about
ready
to
do
the
time
anyway.
B
This
is
marginally
underwhelming.
If
you
care
about
user
experience,
but
the
fact
is,
you
can
now
install
the
app
wait
a
second
now
what
used
to
happen
when
you
would
run
the
app
is
Oh
a
sexy
pop
up,
a
warning
that
says
this
application
is
unsigned
and
unknown,
so
you
can't
install
it.
So
basically
have
this
warning
message,
but
with
no
no
way
to
run
it.
B
So
this
is
in
UX
terms
of
them
it's
somewhere
between,
like
some
very
minor
improvement,
because
it
still
get
a
warning
or
a
huge
improvement,
because
the
users
can
open
and
run
so
as
to
school.
Like
don't
run
this,
it's
made
by
monsters,
you
shouldn't
touch
it
so
now,
after
many
many
days
of
talking
to
Jesse
label
and
dealing
with
bizarre
admin
rituals,
we
have
signed
binaries,
so
that
is
and
less
exciting
is
the
is
a
hundred
less
expect
means.
B
The
Windows
story
is
still
on
hold,
while
waiting
to
get
switch
slides
and
we
are
going
to
go
and
do
some
due
diligence
and
check
that
protocol
lamps,
isn't
organizations,
and
eventually
those
numbers
and
certs
no
more
go
through
the
same
dance
with
that
at
least
I've
done.
It
now
so
I
know
where
it
goes.
B
Other
exciting
games
head
with
yoga
great,
we
learned
lots
about
their
guts
of
the
IP
of
the
Explorer
code
and
we
will
do
some
more
pairing
on
those
changes.
Next
week.
Just
dealing
with
there's
been
I
kind
of
saw
that
there
was
lots
of
API
changes
coming
for
IPL
D
and
some
of
the
dependency
is
like
tag
PP,
so
I
kind
of
ignored
all
of
the
changes
for
a
few
months,
and
so
that
will
tackle
a
whole
bunch
done
in
one
go.
B
B
Otherwise,
admin
stuff
I
mean
and
interviewing,
and
the
next
steps
for
the
southwestern
analytics
I
just
need
to
do
a
call
to
action
on
the
Status
page,
because
we
want
to
make
it
opt-in
so
that
we
don't
offend
anybody
if
we
are
to
get
any
analytics
out
of
The
Selfish
Analytics
I'm
gonna
just
do
a
little
Bama
on
the
Status
page.
When
you
first
run-
and
it
says
please
help
us
make
this
better
and
enable
the
analytics.
B
B
Think
there's
probably
too
many
things
like
to
do
to
get
it
done
all
next
week,
but
we're
gonna
get
as
close
as
we
can
and
Act
es
is
not
with
us
today,
but
he
has
been
ill,
so
not
a
huge
amount
stuff,
but
he's
been
focused
on
the
context
menu
for
Windows
users,
so
integrating
agonist
is
not
with
the
operating
system
so
that
you
can
right
click
on
a
folder
and
say
that's
ID,
FS
and
I
think
this
PR
is
ready
to
review.
So
it
works,
which
is
exciting.
B
C
C
B
B
B
D
Okay,
so
a
few
things
are
at
least
somewhat
related
to
this
group,
so
Alan
spent
some
time
with
me
yesterday,
as
I
just
tried
to
start
from
scratch
at
js5
p.m.
yes
I
think
I
was
just
fine
try.
What
I
was
trying
to
do
is
find
resources
about
the
file
API.
That
already
exists
because
I'm
trying
to
build
a
proto
school
tutorial.
So
we
went
through
some
of
the
resources
that
were
there
together
and
in
the
notes
here.
D
You
can
see
a
link
to
the
issue
where
we're
talking
about
what
the
protocol
tutorial
will
look
like
so
I
laid
out
a
proposal
that
would
have
actually
two
tutorials
one
of
them
about
the
stuff.
That
I
think
is
it's
like
not
m
FS,
it's
technically
the
core
stuff,
and
you
know
cat
and
those
kinds
of
things
and
then
a
second
one
that
would
be
about
m
FS
and
when
I
mentioned
that
to
Michael
he
had
been
envisioning
only
covering
the
NFS
stuff
and
completely
skipping
the
other
stuff.
D
So
this
is
a
point
of
perhaps
some
debate
that
needs
to
happen.
So
I'd
be
very
open
to
any
feedback
on
that
PR.
Before
I
get
started
building,
so
please
go
check
that
out
and
well.
We
were
on
that
main
page
there's
like
a
file
exchange
demo,
that's
built
in
where
you
see
the
code
on
the
left,
and
then
you
see
that
app
that
it
builds
on
the
right
and
when
I
tried
to
drag
and
drop
into
there.
D
I
got
a
type
error
which
Alan
identified
as
being
because
of
a
change
that
was
made
recently
to
the
API,
where
you're
not
supposed
to
include
the
word
files
anymore
when
you
type
stuff
but
I
I'm
like
completely
lost
and
github
I,
don't
know
where
the
repo
is
for
that
website.
So
I
don't
know
where
to
blog
this
issue.
D
My
impression
is
that
you
can
do
that
and
go
now
and
not
quite
enjoy
us,
and
maybe
the
next
release
of
JSI
PFS
will
support
it
or
something
along
these
lines.
So
I
put
a
few
notes
after
kind
of
pinging,
Ollie
and
Lydell
I
put
a
few
notes
in
that
issue
to
kind
of
respond
to
the
guy
who's
there.
But
this
is
one
where,
when
I
talked
to
Michael
about
it,
he
would
like
to
prioritize
this
as
one
of
the
most
most
urgent
tutorials
to
build
after
we
get
the
file
one
out
the
door.
D
So
if
you
guys
have
any
thoughts
on
what
that
would
look
like
feel
free
to
throw
them
in
that
issue
which
is
linked
there
and
then
the
other
thing
that
might
be
of
interest
to
some
of
you
on
this
group.
I,
don't
know
how
much
you
speak
at
conferences,
meetups
whatever,
but
I
started
a
this
is
not
actually
what
I
started.
Sorry,
let
me
put
the
right
word
in
here:
I
started
speaking
Channel
and
slack,
and
this
is
one
of
my
okay
ours
on
me.
D
Sort
of
company-wide
version
of
the
community
team
is
to
help
employees
have
a
place,
an
ongoing
place
to
improve
their
speaking
skills.
So
one
of
the
ways
we'll
do
it
is
just
conversation
and
slack.
If
you
see
a
cool
article
about
slide
design
or
battling
an
impostor
syndrome
or
whatever
kind
of
thing
that
might
be,
and
then
I'm
also
planning
to
set
up
a
semi-monthly
call
where
we'll
have
an
opportunity
to
give
it
like.
If
you
have
a
presentation
coming
up,
you
can
book
it
on
the
calendar.
D
Give
your
presentation
on
the
call
get
feedback
from
other
people
who
are
hopefully
experienced
speakers
and
get
some
tips
on
how
to
move
forward.
The
equivalent
of
this
call
at
IBM
was
everyone
on
my
developer
advocacy
team's
favorite
call
of
the
week.
It
was
just
an
awesome
place
to
kind
of
brainstorm
together
and
share
ideas
and
in
a
group
where
that
was
very
forgiving
and
also
had
a
lot
of
good
feedback,
so
hope
that
people
will
join
that.
So
you
can
go
over
to
that
slack
channel.
D
There's
a
comment
there
about
the
call,
just
click
on
the
little
hands
up
icon.
If
you
want
to
be
included
with
me,
get
that
it'll
be
on
the
public
calendar
but
as
I
try
to
find
a
time
that
works,
I
will
consider
people
who
Clips
the
hand
emoji,
and
my
next,
like
my
big
priority
right
now,
is
building
that
files
tutorial
for
part
of
school
I
like
bi
and
I
need
to
work
on
a
road
map.
So
that's
what's
next
up
for
me.
Does
anybody
have
questions
on
it
and
stuff
I
just
talked
about.
B
D
E
E
E
Basically,
people
that
still
use
like
bubble
and
stuff
to
transpile
to
echo
script
files
and
have
a
hard
time
handling
stuff
that
ik
subclasses
negative
classes
like
URL,
so
I
needed
to
change
the
like
using
classics
time
to
extend
URL
to
use
the
delegation
so
bubble
and
create
react
and
stuff
like
that
doesn't
like
blow
up
with
the
with
the
class
extending
in
class
I
also
had
supports
either
supported
for
the
since
the,
for
course,
in
some
stuff
inside
the
test.
So
we
don't
need
to
like
disable
security
when
we
run
a
text
in
browsers.
E
Basically,
we
needed
to
create
custom,
browsers
or
custom
configuration
for
browsers
to
run
karma
to
deserve
a
web
security,
and
now
we
can
just
use
the
standard
browsers
that
dr.
karma
already
have.
This
is
related
to
the
CI
prototypes
that
I'm
finishing
also
the
bugs
and
others.
Another
thing
related
to
URL
that
a
struct
was
ready
to
ISO
URL,
but
it
isn't
basically
somewhere
in
the
IP
fresco,
probably
to
p2p
IP
offense
ipv6
addresses
are
not
kindly
using
brackets
when
they
build
an
URL.
So
that
gives
an
error.
E
Now
we
might
pick
it's
my
URL
package.
I,
don't
even
know
how
that
works
before
either.
We
only
now
are
starting
to
see
ipv6
I
piece
or
something
was
adding
good
stuff
before
I,
don't
know
which
one
is
it,
but
Velasco
isn't
trying
to
figure
out.
What's
happened
happening
there
and
fixing
that
stuff
also
related
to
the
window
size
for
requests,
while
trying
trying
to
fix
em
Plex
and
basically
gave
up
on
making
it
work.
E
E
So
I
already
talked
about
oh
I,
also
finished,
adding
cough
support
for
the
Kurama
set
up.
This
is
also
related
to
the
CI
prototypes
and
physically.
That's
it
I'm,
as
I
said
locked
on
the
ant
likes
and
switch
both
requests
a
lot
on
the
size
and
now
we'll
try
to
finish
tech
stuff
this
week.
Owls
so
I
will
also
try
to
finish
the
pull
string
to
string
and
CI
prototypes
questions.
E
Yeah
Allen
even
mention
it
that
is
already
being
in
some
places,
but
we
found
some
other
places
that
are
not
using
it
so
somewhere
we
need
to
either
start
using
it
or
some
find
some
other
way.
I
think
that
bugs
meets
that
it's
like
I,
don't
really
think.
We
are
only
now
started
to
see
ipv6
something
else,
change
yeah,
but
we
only
now
start
to
see
the
errors
so
something
weird
going
on.
G
Just
before
we
head
off
on
why
I've
been
doing
on
that
point
that
we're
just
talking
about
the
there
is
a
couple
of
instances
where
our
CC
and
WebSockets,
where
it
is
just
extract
like
they're
trying
to
dial
two
particular
addresses
and
it's
extracting
out
an
ipv6
address,
and
it's
not
placing
it
in
square
brackets.
So
that's
that's
likely
the
cause
and
I'm
guessing
Hugo
that
actually
your
library,
because
that
that's
the
thing
that
changed
is
better
at
validating
that
those
URLs
and
that's
probably
why
we
didn't
see
it
before
but
yeah
anyway.
E
E
So,
but
all
of
that
code
goes
through
either
you
know
the
URL
or
the
browser
URL.
That
now
are
basically
the
same.
So
I
guess
it's!
The
old
relative
package
was
doing
some
stuff.
That's
mask
the
error,
and
now
it's
kind
of
going
straight
in
and
either
the
browser
or
neural
just
errors
out,
because
it
doesn't
know
ipv6
without
brackets.
G
G
We
also
got
your
new
web
UI
as
as
new
as
it
is,
and
yeah.
There
was
a
lot
of
kind
of
API
changes
which
we've
already
talked
about
and
have
come
across
problems
in
the
website
and
stuff
object.
Changed
we've
got
like
files
stuff.
We
talked
about
the
file
stuff
being
some
some
of
these
things.
Moving
to
different
places,
we
added
some
cool
API
methods,
DHT
stuff,
yeah,
doesn't
matter
so
much
could
who
took
it
out
the
release
and
then
look.
G
We've
got
this
new
section
which
thanks
everyone
who
did
it
and
there
are
a
lot
of
people,
look
how
many
people
have
commit
and
have
contributed
to
this
release,
pretty
cool.
So
this
is
like
JSF
EFS,
but
also
like
Clippy,
to
be
independent
modules
in
the
ipfs
ecosystem,
so
not
just
JSI
their
best,
but
it
they
are
all
depended
on
or
used
by
Jess,
I
guess
so:
yeah
loads
of
fun,
fun
times,
fun,
people,
good
people,
doing
good
work
and
so
yeah.
What
else
cool?
G
Okay,
so
the
IP
OD
etherium
and
I
thought
he'd
get
libraries
didn't
have
their
browse
bills
enabled
basically
they
were
working
because
it
builds
them,
but
they
weren't
being
like
they
were
in
the
get
ignore
the
disk
directory.
So
when
they
don't
when
they
were
published
to
npm,
the
district
is
going
up
there
Eva.
So
that
was
a
fairly
easy
fix
to
add.
An
NPM
acknowledge
didn't
ignore
the
disk
folder.
So
that's
good
and
I
did
that,
because
I
saw
that
Hugo
was
is
doing.
G
Oh
I
saw
I
I
know
that
you
guys
doing
a
whole
bunch
of
work
on
making
the
bundle
size
smaller
in
the
browser
which
is
super
awesome
and
the
the
pull
request
against
ipfs
at
the
moment
removes
the
IP
ID
formats
that
are
not
the
default
ones:
dag
Seaboard,
dog,
peepee
and
sorry,
so
that
the
bundle
size
is
super
small.
But
then
you
don't
get
to
use
these
things
in
the
browser
and
I.
Remember:
David,
Diaz,
saying
something
about
loading.
G
It's
like
the
PI
PFS
CDN,
which
is
cool,
but
also
it
means
that
you
can
wecan
like
the
in
IBRD.
This
is
feature
called
load
format
which
which
I
added
that
it
allows
us
to
like
load
formats
dynamically.
So
if
I
PLD
doesn't
know
about
a
format,
it
will
call
this
function
and
you
can
asynchronously
like
go
get
that
format
and
and
then
provide
it
to
IPO
D
and
whatever
IP
I'll
D
was
doing.
It
can
continue
to
do
so.
G
G
So,
like
I,
there
is
a
pull
request.
Yes,
I,
professor
I
haven't
put
a
link
in
now
put
it
in
the
door
in
the
door,
but
I'm
interested
to
hear
your
like
opinions
on
it.
I
know,
Lytle
and
Hugo
have
already
commented
and
I
need
to
read
them,
but
yeah
I
sort
of
thought.
It
was
a
fun
fun
thing
to
do,
say,
yeah,
but
like
it's
kind
of
it's
there
I
don't
know
if
I
made
you
or
not
we'll
see
yeah
like
a
yeah
see
anyway.
G
G
G
It's
like
this
module
and
and
so
removing
them
I,
don't
see
any
like
I,
can't
think
of
any
reason
by
why
they
shouldn't
just
be
removed,
so
I
just
removed
them
and
just
running
the
test
now
and
hopefully
and
I
tested
it
with
meteor
and
it
works
now,
whereas
before
it
didn't
so
easy
fix.
Anyway,
that's
me:
that's
what
I
did
I'm
not
blocked
on
anything.
At
the
moment,
I
started
the
CID
basis,
32
migration
thing
with
starting
with
JSC
ID
change
the
default.
G
D
G
B
G
G
B
B
B
This
is
news
just
in
that
particularly
interesting
for
the
in
web
browsers
and
web
extension
crowd.
Chrome
has
made
some
proposals
about
where
they're
going
to
take
the
web
extension
its
API
next
and
the
work
is
being
done
under
the
banner
of
the
V
version.
Three
manifest
by
thinking
that
somewhat
distracts
from
the
core
of
the
changes
they're
proposing,
which
is
in
an
interest
of
making
web
extensions,
have
not
be
able
to
impact
performance
of
the
browser
requesting
and
making
HTTP
requests.
B
They
want
to
switch
from
allowing
extensions
to
have
programmatic
control
so
basically
have
the
browser
inform
web
web
extensions
about
every
outgoing
HTTP
request
and
then
say.
Is
this
one?
Is
this
a
web
request
that
you'd
wish
to
redirect
or
to
block
and
that
functionality
is
used
heavily
by
architects
companion
to
allow
redirecting
to
if
it
sees
a
valid
HTTP,
flavored
ipfs
address?
It
uses
this
fact
to
redirect
back
to
your
local,
your
local
IP,
this
HTTP
gateway,
also
and
blockers.
Any
form
of
web
extension
based
content.
B
Blocker
makes
Heather
use
of
this
functionality
to
go
and
check
kind
of
standards
lock
list
of
things.
The
proposal
from
chrome
is
that
they're,
seeing
a
significant
performance
impact
by
of
having
to
have
every
outbound
HTTP
requests,
go
through
a
list
of
well
extensions
and
say:
hey
well,
extensions.
Can
this?
Can
this
go?
It's
a
blocking
API,
so
each
one
extension
gets
a
chance
to
veto
the
request
so
from
a
perf,
an
goal.
B
It
makes
it's
understandable
that
this
proposal
has
come
through,
but
they're
they're
proposed
fix,
for
it
is
to
basically
take
away
programmatic
control
and
turn
it
into
a
declarative
API,
where
web
extensions
have
to
declare
in
advance
in
their
manifest,
so
statically
define
it
a
set
of
rules
for
which
urls
and
how
the
what
behavior
that
Brow's
I
should
do.
So
you,
you
tell
you,
give
chrome
a
recipe
where
you
declare
it.
Yes,
thanks
the
Clair
stuff,
so
you
want
to
handle
every
single
URL,
but
there's
the
other
problem.
B
But
the
bigger
issue
here
is
like
brave,
has
moved
to
the
chromium.
Api
edge
is
maybe
to
chromium
like
if
they
get
to
call
the
shots
on
what
the
web
extension
API
looks
like.
So
they
and
then
in
the
background,
like
Firefox
a
year
ago
or
two
years
ago,
switched
to
making
their
web
extension
API
more
like
a
crime
so
that
browser
web
extension.
Vendors
could
have
a
shot
at
making
portable
web
extensions,
and
we
moved
companion
to
that.
So
this
is
kinda.
This
is
pretty
bad
news.
It's
not
easy
to
sugarcoat
it
like.
B
We
could
still
I
think
we
can
statically
define
a
rule
that
like
identifies
an
IPS
path
and
legree
the
written
out
to
your
local
gateway,
but
we
lose
functionality
around
being
able
to
toggle
it
at
one
time
as
to
whether
you
want
to
go
to
the
public
gateway
or
a
local
gateway,
and
maybe
Martin
you
unpacked
a
few
more
issues
as
well.
So.
A
Like
like
morning
that
nodes
are
on
the
issue,
but
basically
we
do
much
more
right
now
done
looking
at
the
URL
and
looking
if
there's
like
ipfs
or
IP
NS
path
like
path
starting
with
this
segment.
We
expect
HTTP
headers
for
additional,
like
X
ipfs
path,
rather
as
additional
hint
that
the
website
might
support
Dennis
link,
and
then
we
do
DNS
link
lookup
to
confirm
that
and
the
web
request.
Api
is
not
blood.
It's
not
blocking
by
default
like
making
it
blocking
is
an
opt-in,
so
we
are
blocking
it
because
we
need
that.
A
But
there
are
other
extensions
that
don't
need
to
block.
They
can
just
a
synchronous,
observe
so
I'm
not
really
convinced
how
how
big
of
a
problem
is?
That's
like
realistically
for
us
like
we
will.
We
would
be
able
to
do
some
things
with
this
proposed
API,
but
we
would
lose
a
lot
of
the
single
seamless
experience
right.
A
Basically,
the
detection
of
the
NS
linked
websites
would
probably
go
away
ability
to
really
like
redirect
things
based
on
headers
or
like
do
header
based
opt-out
or
maybe
even
avoiding
adding
course
headers
to
your
local
Iranian
gateway
to
have
ipfs
companion
being
able
to
talk
to
it.
It's
like
sort
of
a
separate
issue,
but
a
related
one.
A
A
A
It's
like
ongoing
issue,
but
we
need
to
follow
it
to
ensure
we
like
to
to
make
sure
we
are
planning
accordingly,
because
it
would
be
a
shame
if
this
change
got
rolled
out
without
any
any
yeah
any
feedback
from
our
end
and
also
like
there
are
multiple
like
there
are
multiple
changes
that
could
potentially
maybe
not
break
ipfs
companion
but
would
require
us
to
basically
rewrite
a
big
chunk
of
it.
Basically,
there's
like
one
proposal
to
remove
the
concept
of
background
page.
A
It's
the
concept
is
that
there's
a
background
page
that
is
hidden,
it's
not
visible
to
user,
but
basically
there
is
a
Dom
HTML,
DOM
and
JavaScript
can
run
in
the
background
on
that
background
page,
and
they
want
to
remove
that
and
replace
that
by
serviceworker,
which
we
started
and
stopped
only
when
this
like
filter
list
hits
much
right.
Yeah.
B
Just
a
quick
rule,
but
I
think
we
explained
the
existential
threat
that
the
v3
man
of
the
thing
changes,
and
we
should
talk
very
quickly
about
brave
as
well.
These
will
just
unpack
a
little
bit
more
detail
of
the
cool.
D
A
Should
I
should
I
like
give
a
quick
background
on
general
plan?
Yeah
that'd
be
great
yeah.
So,
like
the
plan
is
in
the
past,
we
like
in
2018
the
best
we
could
do
was
to
redirect
to
locally
running
ipfs,
go
ipfs,
node,
so
and
brave
had
a
plan
to
ship
the
Tor
binary
with
the
browser
embedded
in
the
browser.
So
it
was
like
a
hammer
and
the
idea
was
to
yeah.
Basically
we
could
do
the
same
for
ipfs
and
in
2019.
A
Now
jpf
is
move
in
implementation
of
things
like
an
affairs,
the
HT
will
launched
on
IP
and
s
over
DHT
things
like
that.
So
there's
like
a
feature.
Parity
comment
on
Kleiss
knee
when
it
comes
to
JSA,
PFS
and
also
brave
themselves,
figured
out
a
framework
for
exposing
some
Chrome
OS
API.
So
those
are
more
powerful.
Ap
is
from
Chrome
OS
that
are
not
available
in
regular
chrome,
such
as
raw
sockets,
API,
TCP,
sockets
and
TDP,
and
you
are
able
to
start
your
I'll
TCP
server
or
listen
to
for
TCP
connections.
Things
like
that.
A
So
basically,
the
idea
is
that
if
we
have
access
to
that
dog's,
api's
from
ipfs
companion
and
we
run
chase
ipfs
in
ipfs
combine
a
context,
then
we
could
have
a
feature.
Parity
we've
got
EFS,
at
least
when
it
comes
to
starting
HTTP
gateway
and
redirecting
to
local
HTTP
Kotori,
just
like
we
do
for
the
NGO
activist.
But
the
difference
is
that
there
is
no
binary
to
worry
about
it's
much
easier
to
out
it.
It's
much
easier
to
cause
like
security
parameter
for
web
extension
is
already
defined.
A
B
B
B
A
So,
basically,
without
that
we
did,
this
plan
won't
work
if
we
are
not
able
to
expose
HDD
cage
way
from
embedded
Jase.
Ip
first
note
another
like
sort
of
a
prerequisite
for
that
is
to
add
defense
detection,
like
native
ipfs
detection,
to
break
itself.
So
brave
will
not
ship
with
IP
fast
companion
in
a
book
for
everyone,
but
the
first
time
user
is
visiting
website.
We
fight
give
us
resources
or
there's
like
ipfs
header
in
HTTP
response,
or
maybe
there's
a
DNS
link
or
maybe
user.
A
There's
a
JavaScript
call
to
window
ipfs
to
enable
eye
professor
API,
then
there's
like
a
prompt
to
user.
This
website
could
make
use
of
ipfs.
What
CIFS
would
you
like
to
enable
it
and
then
ipfs
is
started
and
automatically
J's.
Ipfs
note
starts
things
like
that,
so
for
this
plan
to
work,
we
need
at
least
to
expose
HTTP
gateway
in
some
way
and
those
low
level
so
consider
are
the
best
the
best
way
or
the
only
way,
I
see
right
now
to
do
without,
like
third-party
software.
B
Ok,
it's
nearly
only
our
I
would
it
would
be
nice
for
to
check
the
British
you
because
I
think
there
is.
There
are
some
things
to
unpack
on
the
pros
and
cons
of
using
an
embedded
case.
I'll
give
s
as
the
default,
so
we
could
to
get
people
source
on
that
as
once.
This
is
rolled
out
that
would
make
I
give
this
a
lot
more
visible
to
a
lot
more
people.
B
F
G
B
We
can
talk
a
bit
more
about
it
and
now
I
gotta
go
now.
Then
let's
have
a
call.
Please
read
the
issue
and
your
thoughts
there
and
we
all
showed
you
a
call
from
Baptists
sweet
thanks.
Everybody
is
ipfs
gooey
and
in
web
browsers,
weekly
sync
same
time
same
place.
Next
week's
brings
knit
your
GUI
and
in
web
browsers
ideas
and
we
will
talk
on
them.
Bye,
bye,.