►
From YouTube: IPFS Web and GUI Biweekly 2020-07-14
Description
Meeting notes: https://github.com/ipfs/team-mgmt/pull/1168
About IPFS GUI and Browsers Biweekly: https://github.com/ipfs/team-mgmt/issues/790
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
Okay,
welcome
to
Quian
in
web
browsers,
bi-weekly
call
for
14th
of
July
2020
this
week
on
our
agenda,
quick
update
about
pinging
services,
quick
update
on
subdomain
gateways,
performance
improvements
for
uploading,
web
UI
and
general
over
HTTP
client,
IP,
FS,
desktop,
update,
extravaganza
and
more.
That
was
me.
Reading
over
agenda.
I
thought
I
tried
that
so
yeah.
The
first
item
is
spinning
services
feel
free
to
add
your
agenda
items
while
I
go
over
this
one.
A
So
now
we
have
this
pinning
service
API
trapped,
it's
a
very
simple
API,
the
pin
object,
lifecycle
is
just
creating
modifying
and
removing
pins.
We've
tackled
problems
related
to
pinning
service
fetching
new
data
from
peers
that
are
behind
nuts
and
in
other
restrictive
network
topologies
by
adding
closed
provider
hints
to
the
response
on
our
client
has
an
information
that
hey.
This
is
a
multi
other
of
pinning
service.
You
should
connect
to
it
and
that
way,
pinning
service
is
able
to
find
the
content
faster.
A
It
does
not
need
to
wait
for
the
HT
and
we've
removed
a
bunch
of
fields
and
replaced
them
with
a
meta
dictionary
where
vendors
can
add
their
own
parameters,
which
are
about
the
specific
to
the
functionality.
So
it's
open-ended
spec
every
pinning
service
implements
the
minimum.
That's
in
this
document,
however,
there
are
free
to
add
as
much
as
they
want
to
the
meta
parameter.
So
the
list
is
listing.
Pin
objects
is
pretty
simple.
A
Pagination
is
built-in
and
we
have
some
hard
limits
for
the
number
of
results
that
are
returned.
It's
about
1k
of
results,
user
can
add
multiple
pins
in
a
single
request
and
then
I
can
get
status
of
specific
single
object
and
then
can
modify
that
pin,
object,
add
additional
metadata
or
change
providers
or
even
see
ID.
That
might
be
useful
in
our
case
for
web
UI
integration,
where
in
case
we
implement
pinning
policies,
the
root
of
amethyst
would
be
a
pin
that
is
constantly
modified
on
every
NFS
change.
A
And,
finally,
when
the
pin
is
no
longer
desired,
you
can
remove
it.
That's
it
that's
entire
API,
so
I
posted
links
to
this
working
progress,
spec
in
the
notes,
but
I
believe
by
the
time
people
watch
this
video.
It
will
be
more
or
less
finalized.
We
already
don't
expect
too
many
liquor
groundbreaking
changes.
Mostly,
we
may
specify
additional
limits
for
value
ranges
or
we
may
move
something
like
optional
fields
to
be
promoted
to
be
mandatory,
but
nothing
too
extreme.
A
So
we
encourage
people
to
visit
visit
this
repo
after
you
familiarize
yourself
with
this
spec
in
an
existing
shape
and
feel
that
something's
missing
or
something
is
not
clear
or
feel
free
to
open
issue
or
even
open
a
request.
Making
direct
change
propose
direct
changes
to
this
back.
We,
the
source
of
truth,
is
this
Yunel
file.
Here
it's
an
open,
API
format,
which
has
a
very
simple
docks
in
the
beginning,
and
then
it's
just
an
API,
spec
I.
A
Believe,
that's
it
for
my
part
with
the
API
and
we
hope
to
effectively
finalize
the
spec
this
week,
and
by
that
we
mean
we
will
give
a
green
light
to
cleaning
services
to
start
implementing
it.
And,
of
course
we
will
probably
gather
feedback
in
the
next
few
weeks
when
people
start
implementing
the
API
and
like
hidden
complexity,
surfaces
will
probably
a
continued
work
in
that
repo.
So
if
you
are
pinning
service
provider
or
are
interested
in
in
this
space,
you
may
want
to
watch
this
repo.
B
D
A
A
D
A
Apart
from
that,
there's
a
separate
discussion
in
the
web
UI
repo
regarding
stats,
I
filled
it
today
in
the
mock-ups
of
kibou
I,
think
screen.
We
had
column
for
transfer
and
storage
per
pinging
service
right
now
in
the
pinning
service
API,
we
don't
have
any
endpoint
for
fetching
stats
like
usage
stats
and
I
just
proposed
to
the
scope
that
from
MVP
I
just
mentioned
it
here
in
case.
Someone
feels
strongly
about
it,
but
I
don't
feel
it's
that
useful
to
users,
especially
like
given
in
different
pinning
services,
may
have
different
billing
strategies.
A
So
we
may
not
come
up
with
the
proper
way
of
visualizing
the
showing
visual
way
of
utilizing
different
services,
especially
they
have
different
pricing
strategies
they
may
have.
One
may
have
billing
per
storage,
another
for
transfer
or
something
else
so
I
would
propose.
I
propose
the
scoping
that
from
MVP.
So
if
anyone
has
any
stronger
opinions
about
that,
there's
issue
which
I'll
link
later
yeah.
C
A
C
I
think
I
think
really
within
within
the
UI
mock-ups
that
was
just
to
that
was
just
to
give
us
an
affordance
for
columns
in
that
space.
I
don't
you
know,
I
don't
feel
like
it's
any
wasted
work,
but
I
would
be
very
interested
in
keeping
tabs
on
that.
I
think
there's
going
to
be
some
sort
of
statistical
information
on
a
per
pinning
service
provider
that
will
be
useful
to
people.
I
just
don't
think
we
know
what
that
is
yet.
A
Yeah,
absolutely,
I
think
it's
better
to
hide
those
columns.
We
can
always
bring
them
back
later,
all
right,
a
quick
update
on
subdomain
gateways
in
case
someone
was
missing
that
topic,
so
interesting
peers
got
merged
this
week
last
week.
One
is
support
for
Edie
2f
19,
the
p2p
keys
in
subdomains.
That's
good
news,
because
we
are
planning
to
switch
to
those
inline
keys
in
next
releases
of
go
ifs,
and
this
enables
ask
to
use
them
in
subdomains,
which
means
IP.
A
Ns
websites
would
have
a
proper
and
valid
TLS
cert
on
public
gateways,
and
we
do
that
by
defaulting
all
subdomains
for
all
little
piggys
to
base
36
instead
of
32,
that
by
some
bias
as
enough
space
to
fit
in
under
dns
limit
of
63
characters,
and
the
second
PR
is
oh
and
in
case
CID
is
too
long
even
in
base
36.
The
subdomain
will
return
an
error
for
now
effectively
if
you
use
like
longer
hash
functions
like
sha-512.
A
D
Yes,
I
guess:
I
should
have
named
it
differently,
so
I
do.
We
did
had
a
meeting
last
week
with
the
current
limitations
folks
about
implementation
strategies.
We
had
to
decide
between
doing
fully
browser
for
native
implementation
versus
Samuel
upset.
We
decided
to
do
somewhat
native
and
maybe
in
the
future
we
will
reconsider
that
decision.
D
D
There
is
a
little
chance
that
some
changes
would
have
to
be
made
for
the
two
which
is
maybe
will
switch
to
the
fetch
blob
implementation
versus
Lobby
implementations
that
I
included
in
the
pull
request.
Problems
are
a
slope
has
bugs,
so
that
would
mean
we
would
have
to
fix.
Some
of
those
I
said
I
think
it
will
be
ready.
A
D
Actually
I
recalled
one
thing
that
I
should
have
captured
so
I
do
have
a
problem
with
my
web
UI
pull
request
said:
I
was
not
exactly
sure
how
to
so
that
is.
It
seems
that
I
think
there
used
to
be
some
things
that
track
the
progress
of
upload
and
it
rendered
it.
But
so
there
are
some
actions
immediate
by
the
uploading
saying
my
implementation
actually
does
a
progress
handling
and
amid
those
actions,
but
there's
nothing
listening
to
those
actions
except
to
when
it's
done,
which
is
fairly
bad.
D
When
you
have
multiple
gigs
of
files,
it
takes
quite
a
bit
I
think
1440
14
seconds.
I
was
a
time
so
having
no
visual
feedback
for
14
seconds
feels,
like
things
are
broken
all
he
mentions
are
there
were
some
things
that
removed
it
and
then
I
guess
it
got
lost,
and
some
of
these
things
were
left
some
I'm
not
too
familiar
with
the
code,
which
I'm
not
exactly
sure
where
to
wire
sinks.
Where
so
any
help
there
would
be
helpful.
B
B
D
B
A
A
B
B
Sorry,
a
plate
and
I
think
meanwhile
found
the
problem
and
fix
it,
but
that
will
only
work
from
this
version
onwards.
So
just
keep
in
mind
that
any
issue
that
will
appear
while
we
haven't
released
a
new
version.
It's
probably
really
dead.
However,
I'm
gonna
release
a
a
new
version
as
soon
as
I
finished
finished
grooming.
B
My
my
final
poll
request
of
the
on
the
pinning
settings
and
after
that
we
just
ask
feedback
again
for
the
users
just
check
if
everything
is
updating
properly,
but
I
really
think
that
once
we
have
the
time
after
pinning
services,
we
should
really
dedicate
some
time
into
update,
because
I
think
this
is
the
kind
of
features
that
are
bugs
in
this
case
that
we
can
lose
users
due
to.
If
they
get
really
mad
about
the
dating
issues,
they
will
just
stop
updating
or
stop
using
the
app.
A
That's
leverages,
ipfs
like
as
much
as
possible
yeah
for
that
this
specific
bug,
I
think
the
one
you
described
was
mostly
the
Windows
one,
which
was
a
problem
for
users
who
installed
aquifers
desktop
globally,
as
an
admin
in
like
globally
for
all
users
on
the
system
and
I.
Think
there's
a
bug
in
electron
builder,
where,
if
you
in
start-up
that
way
and
there's
an
update,
the
update
operation
will
uninstall
the
app.
So
that's
that's
interesting
way
of
breaking
Europe.
A
B
A
A
little
bit
like
skeptical
when
it
comes
to
trying
to
reproduce
and
I
think
on
Windows
platform
now
like
it
was
easier
when
you
had
like
Windows
7
Windows,
something
now
you
got
Windows
10,
but
there
are
like
dozens
of
Windows
stands
and
they
change
random
stuff
between
them
and
it's
now
it's
even
ridiculous.
It's
hard
to
tell
which
build
user
was
running.
A
B
So
can
you
see
my
screen
here
so
this
is
the
settings
or
just
to
quickly
show
you
what's
going
on
so
I
kind
of
want
to
fix
some
accessibility
issues
with
the
table
for
some
reason
after
I.
Did
the
accessibility
fixes
the
table
now
as
some
weird
focus
stuff
that
shouldn't
be
happening?
Why,
for
some
reason,
I
need
to
discuss
if
I
click
on
this
row
should
I
also
open
the
three
icons
menu
over
here?
No,
it
shouldn't
be
a
thing:
okay,
nice.
B
Once
we
have
the
repinning
API
providing
the
list,
we
can
also
have
some
URLs
over
here
or
some
custom
actions
to
add
opinion
services
in
a
pinning
service
it
instantly.
But
for
now
we
can
add
a
custom
one
I'm,
just
finishing
the
UI
liming
everything
and
spacing
and
whatnot,
but
we
can
just
add
a
custom
painting
services
over
here
and
we
also
need
a
link
for
the
learn.
How
right
now
we
just
linked
into
the
home
page.
B
C
B
A
It
yeah
some
some
updates
on,
like
the
Settings
aspect
of
pinning
services,
is
that
last
time
I
talked
with
a
Dean.
We
recognized
the
problem
of
storing
API
secrets
in
IPS
notes,
config,
and
we
already
have
a
convention
of
not
returning
private
keys.
When
you
read
config
over
API,
slash,
v-0,
convict
endpoint
effectively,
the
screens
the
settings
screen
are
in
web,
UI
will
be
reading,
pinning
services
map
from
the
config
and
the
way
it
will
work.
A
A
A
A
So
I
was
thinking
about.
He
was
even
thinking
about
adding
like
an
info
endpoint
we're
pinning
service
could
provides
all
those
additional
things
like
link
to
their
own
dogs
link
to
their
like
user
onboarding
web
like
pop
on
their
website
and
see
ID
of
their
logo.
But
that's
yeah,
I'm,
not
sure
if
we
want
to
persist
all
of
the
ads
in
in
the
API.
A
A
A
A
C
A
A
Got
a
huge
revamp,
Jessica
redid,
the
entire
main
menu,
and
now
it's
it
does
not
take
your
entire
screen,
and
it
shows
you
much
more
personally
I
like
those
copy
elements
which
now
show
you
what
you
will
be
copying
and
we
save
a
bunch
of
stuff
in
size.
A
bunch
of
language
tweaks
in
both
main
menu
and
preferences
pages
choose.
A
That's
available
in
for
Chrome
and
Firefox
interesting
fact:
Google
changed
something
in
the
way
they
calculate
stats.
We
had
over
30
k
of
users,
and
now
we
have
just
seven
and
it's
still
unclear
why
the
rapid
decrease
of
users
I
checked.
It
I've
been
looking
at
it
for
the
past
weeks
and
we
are
growing
around
like
between
50
and
100
of
new
users
each
week
after
this
drop
not
sure
why
maybe
Google
fighting
with
BOTS.
A
Maybe
we
got
some
fake
users
which
pretend
to
use
multiple
extensions,
no
idea,
that's
not
the
only
problem
people
had
with
it's,
not
just
us
having
this
problem.
Multiple
people
raised
that
issue
on
forums,
so
yeah
just
thought.
I
mentioned
that
in
case
people
wonder
why
20k
of
users
and
installed
IP
first
companion,
I,
think
we
also
landed
on
this
case.
Study.
I
am
not
sure
if
we
talked
about
this
on
this
call,
but
I
think
Jessica.
If
you
could,
like
briefly
mention
it,
yeah.
C
C
C
You
know
how
that
might
fit
into
their
overall
tech
stack
or
just
their
their
overall
metaphors
for
their
project
as
a
whole,
so
really
focused
on
the
value
that
ipfs
and
and
only
ipfs
was
able
to
provide
to
audience
like
I
said
this
is
gonna,
be
a
series
of
these.
They
will
live
canonically
in
the
docs,
but
we'll
keep
them
in
our
pocket
to
use
particularly
the
social
media
and
any
future
revamps
of
the
ipfs
dot
IO
site
next
on
tap
I.
Believe
our
open,
Bazaar,
fleek
and
IO
Tex
I.
A
C
D
A
Like
context
here
is
my
concern
was
we
are
improving
this
performance
of
uploading
files
over
HTTP
API,
but
the
risk
is
that
at
some
point
in
the
future,
maybe
this
performance
gain
may
be
lost
to
refactor
or
something
so
having
some
sort
of
regression
test
is
useful.
I
probably
won't
look
at
it
today,
but
I'll
try
to
talk
sometime
this
week.
Yeah.
Thank
you
folks,
I
think
we
will
end
this
sooner
I'll.
Stop.
Recording
now
see
you
in
two
weeks.