►
From YouTube: JupyterLab Team Meeting - July 27, 2022
Description
A meeting to share and discuss features, ideas, issues, and pull requests in JupyterLab and other Jupyter frontends. This meeting is open to anyone and everyone.
Join future calls via the Jupyter community calendar: https://docs.jupyter.org/en/latest/community/content-community.html#jupyter-community-meetings
Notes for upcoming meetings can be found on the agenda: https://hackmd.io/Y7fBMQPSQ1C08SDGI-fwtg
Past notes can be found on the JupyterLab team compass: https://github.com/jupyterlab/team-compass/issues?q=is%3Aissue+label%3A%22Dev+Meeting+Minutes%22
A
Hello
and
welcome
to
the
July
27th
weekly
Jupiter
lab
call.
Today
we
have
21
people
on
the
call.
If
you
just
joined
us,
please
find
the
agenda
in
the
chat
and,
if
you
feel
like
it,
add
your
name
to
the
attendees
list.
If
you
have
something
to
discuss,
that's
only
in
the
five
minute
range
put
it
in
the
agenda
as
a
bullet
point.
If
you
have
something
that
requires
more
discussion,
wow
put
it
in
the
extensive
additional
discussion
section
we
have
this
week
cool
and
with
that
let's
get
started.
B
Hey
good
morning
to
those
of
you,
somewhere
west
of
the
east
coast
of
the
United
States,
like
myself,
good
evening
good
afternoon,
good
night
for
everyone
else,
so
here
just
doing
one
of
our
security
sub-project
world
tours.
Some
of
you
may
have
seen
the
blog
post
that
we
got
together
with
the
steering,
Council
and
talks
amongst
ourselves,
and
we
thought
that
four
official
Jupiter
security
sub-projects
it's
important
that
all
people
that
have
some
type
of
commit
owner
admin,
access
to
a
GitHub
repo
should
have
2fa
enabled
on
their
GitHub
accounts.
B
Most
of
you
are
familiar
with
why
this
is
a
good
idea,
but
it's
not
completely
straightforward.
In
particular,
there's
a
couple
of
a
little
bit
of
challenges
of
getting
the
word
out,
particularly
to
users,
and
so,
like
you
know,
how
do
we
communicate
with
them
without
telling
the
world
hey?
Here's
a
bunch
of
people
that
don't
have
2fa
on
their
account.
B
Maybe
you'd
like
to
try
random
passwords
on
their
GitHub
account.
So
this
is
kind
of
our
way
of
coming
in
saying
we're
here
to
help
and
exchange
some
ideas
in
particular,
and
that
challenge
of
you
know
how
much
should
we
say
in
a
public
forum.
Jupiter
lab
has
a
ton
of
contributors,
and
that
is
fantastic.
B
That
also
means
that
Jupiter
lab
your
GitHub
org
has
a
substantial
number
of
folks,
either
team
members
or
contributors
to
repositories
that
need
to
enable
2fa
and
it
can
I
want.
Frederick
you've
already
asked
the
right
question
on
the
issue:
I
dropped,
which
is
how
do
we
reach
out
to
them
so
I'm
just
going
to
throw
out
a
couple
of
ideas
and
then,
if
anyone
has
some
suggestions
for
Jupiter
lab
in
particular,
I'd
love
to
hear
them.
B
One
I
know
that
a
lot
of
the
contributors
are
probably
just
interns
from
past
projects
and
what
I'd
suggest
is,
if
you
are
the
admin
or
owner
of
a
repo,
and
there
was
an
intern
on
there
a
while
ago
and
you
can
still
reach
out
to
them.
Like
you
know,
their
email
address,
I
would
say,
take
action,
just
remove
their
access,
send
them
a
note
via
whatever
Forum
email
slack.
You
already
have
proactively
and
be
like
hey.
You
know
we're
doing
this.
B
If
you
want
to
come
back
and
contribute,
you
know,
that's
that'd
be
wonderful
and
we
would
encourage
you,
but
you
know
basically
do
a
little
bit
of
cleanup
if
some
of
you
are
admins
or
owners,
and
you
don't
have
to
enabled
please
do
that
soon.
Your
accounts
are
the
most
critical
and
then
once
we
get
through
that
phase,
so
we
have
in
a
private
repository
a
list
of
the
team
members.
B
The
GitHub
usernames
for
folks
from
each
org
that
don't
have
it
enabled
GitHub
does
not
make
it
easy
to
find
their
email
addresses,
but
usually
with
looking
at
commits
or
their
public
profiles
or
something
like
that
we
can
find
emails.
There
is
no
private
message
mechanism
in
GitHub,
so
we
could
do
that.
Yes,
Vidar.
C
B
Yes,
for
GitHub
at
the
GitHub
org
level,
you
can
enable
a
requirement
that
anyone
who
has
like
read
an
explicit
permission:
you're
an
admin
you're
an
owner.
You
have
read
access
to
a
private
repo
or
write
access
to
a
repo.
You
can
enable
that
at
the
GitHub
organization
level.
So
that's
what
we're
looking
to
do
is
get
everything
get
all
the
user
accounts
so
they're
in
line,
so
that
one
of
the
owners
can
go
click
and
then
it's
done
and
we
forget
about
it.
B
Okay
yeah,
so
maybe
I
should
have
prefaced
that
that's
the
goal
for
each
of
the
orgs
for
each
GitHub
org
that
hosts
official
Jupiter
sub
project,
we're
looking
to
get
to
the
point
where
an
owner
can
go
click
and
then
it's
a
done
deal
and
we
get
some
peace
of
mind.
Some
of
you
may
have
gotten
a
notice
from
Pi
Pi,
Pi
Pi
is
actually
requiring
this
for
critical
packages,
which
is
kind
of
the
next
step
for
us,
anyways
is
to
think
about
what.
B
Four
years
ago,
I
hope
you're
happy
in
your
new
job,
come
back
and
we'll
let
you
in,
if
you
enable
2fa
in
the
middle,
we're
willing
to
take
on
some
of
the
work
in
like
finding
email,
addresses
and
passing
them
on
to
you
privately
to
reach
out
or
help
in
with
an
email
notification.
So
basically
we
knew
that
that
would
be
some
workload
and
I'm
volunteering,
at
least
myself,
maybe
Jason,
because
he's
on
the
call
in
helping
with
stuff
like
that.
B
Well,
what
people
think
about
that
I'll,
respond
to
Michael's
question
Michael,
you're,
right
and
I'm?
Sorry,
if
I'm
mispronouncing
your
first
name,
but
you
are
correct.
Yes,
if
you
click
it,
you
will
kick
people
out.
We
don't
want
to
do
that
and
that's
why,
if
you
look
at
the
blog
post,
which
we
all
drop
in
the
chat
in
a
minute,
we're
trying
to
give
a
30-day
win
or
a
90-day
window
for
the
process?
Basically,
it's
now
the
end
of
September
that
we're
aiming
for.
B
We
want
time
to
notify
people
and
get
the
word
out
and
give
them
a
chance
and
have
the
communities
the
sub
projects
own
the
process,
so
that
when
you
do
turn
that
on,
like
people
have
had
it
been
communicated
to
and
they
don't
feel
like
they've
been
removed
from
the
community
and
appropriately
so
Mike
I
hope
that
open
answers.
Your
pro
your
question.
A
Rick
one
one
thing
that
is
worth
maybe
talking
about
with
maybe
on
the
server
call
or
if
you
happen
to
be
on
another
call
with
Steve
Sylvester,
is
Jupiter
releaser,
so
the
wage
you're
going
to
release
her
works.
Is
you
generate
tokens
to
do
publishing,
but
the
nature
of
publishing
with
tokens
for
npm
and
Pi
Pi?
A
Is
they
aren't
2fa
and
that's
probably
okay,
because
if
your
account
has
2fa
and
you
generated
a
token,
the
token
is
the
thing
that's
at
risk,
but
if
your
GitHub
account
is
also
behind
2fa,
then
presumably
you
can't
access
that
and
you
can't
read
a
secret
after
you've
written
it
anyway.
So
the
worst
thing
you
could
do
is
somehow
publish,
but
you'd
have
to
have
like
it's
convoluted,
so
it
seems
like
it
probably
doesn't
really
mess
with
the
security
profile
you
want
to
build,
but
it's
something
to
know
well,.
B
Thank
you
for
that.
Actually,
because
I,
one
of
the
reasons
I
ended
up
looking
at
stuff
like
this
is
a
history
dealing
with
like
identity
and
access
management,
and
that
thing
where
you
go
from
the
user
logged
in
with
2fa.
But
then
they
want
a
token
to
do
something,
and
just
look
at
the
AWS
CLI,
for
example,
or
even
GitHub,
with
their
token
mechanisms
it's
challenging,
but
doable
so
I'm,
really
glad
that
there
is
a
single
point.
I
can
now
go.
B
Look
at
and
I
think
somebody
has
as
an
issue
associated
with
that,
and
you
just
give
more
given
me
more
context
to
understand
it.
So
yeah
I'd
like
to
that'd,
be
a
great
place
for
us
to
say:
hey.
Let's
make
sure
our
package
updates
to
conda
npm
whatever
it
follows
a
similar
level
of
reliability
so
that
people
can
trust
what
we
do.
So.
Thank
you.
B
So
I'll
wait
for
updates
on
the
issue,
along
with
the
rest
of
the
team
and
just
let
us
know,
and
then
we'll
figure
out
a
way
to
pass
on
an
email
list
or
the
last.
So
let
me
finish
the
thought,
Isabella
and
then
I'll
I'll
stop
chat
running
my
mouth.
The
last
idea
I
had
was
open
a
private
repo
and
at
people
in
the
private
repo,
but
that's
also
messy
because
then
you're,
like
I
gotta,
add
20
collaborators
at
them
and
then
shut
down
and
delete
the
repo.
So,
okay,
thanks
Isabella.
D
B
E
B
Were
78
a
few
weeks
ago,
so
you've
done
you've
made
improvements
just
to
give
you
the
the
last
thing
taking
up
time
on
this.
We
looked
at
this
and
it
is
easier
on
the
or
the
sub
projects
and
orgs
with
fewer
members.
If
orgs,
like
Jupiter,
lab
tackle
theirs
first,
it
basically
clears
out
the
need
for
other
orgs
to
even
address
this,
so
we're
front
loading.
Some
of
the
activity.
A
A
B
More
along
the
lines
of
best
practices
and
maintaining
trust
in
the
community.
This
is
not
based
on
any
specific
thing,
but
if
you
look
at
like
the
actions
of
Pi
Pi
or
some
of
the
blog
posts
that
GitHub
puts
out,
you
know
we
lose
trust
in
our.
Our
community
may
lose
trust
in
us
if
they
can't
rely
on
the
code
and
I'll
drop.
A
link
to
a
recent
blog
post
by
a
law
firm
that
Eddie
within
the
United,
States
and
Europe.
B
E
But
just
for
you
to
be
deferred,
so
I
started
to
refactor
the
extension
manager
priority
V3.
Can
you
see
my
screen
yeah?
So
the
main
ID
here
it's
to
is
to
use
pipei,
so
the
refactoring
that
I
have
done
like
doesn't
look
like
much
things
like
on
the
UI
side.
E
I
only
switch
this
to
to
use
the
the
new
side
panel
with
a
caution
that
we
use,
for
example,
in
the
debugger,
because
yeah
it
makes
the
the
UI
a
bit
more
consistent
and
it
will
help
for
people
helping
with
accessibility
things
and
the
features
are
exactly
the
same,
so
you
can
install
and
install
enable
disable.
You
get
the
warning
still
and
you
have
the
ability
to
filter
things
like
it
was
before.
E
E
Can
I
find
it
quickly
forget
to
open
the
tab?
Sorry
for
that,
the
idea
is
to
to
get
everything
in
the
in
the
back
end,
so
for
people
that
are
not
aware
of
it.
The
current
version
is
using
npm
and
the
request
to
npm
are
done
from
the
front
end,
so
we
are
never
going
through
the
back
end
and
that's
brought
also
the
need
for
downloading
from
the
server
a
list
of
a
load
and
Diesel
loads
extension
and
actually,
even
from
a
security
point
of
view,
it's
better
to
have
everything
on
the
back
end.
E
So
the
new
infrastructure
moved
everything
to
the
back
end.
So
there
is
a
new
extension
manager
here.
It
is
that
people
can
extend,
and
so
I
I've
created
one
for
bypi
I've
created
one.
That's
just
read
only
without
capability
for
installing
things
and,
for
instance,
if
people
want
to
I,
don't
know
create
one
for
conda
or
one
for
NYX
or
whatever
fancy
package
manager
you
are
using,
then
you
should
be
able
to
do
it
and
they
come
with.
So
there
are
still
a
low
list
and
block
list
available.
E
A
E
For
the
Instagram
parallel
yeah,
this
one
is
tricky
because
a
pipei
doesn't
have
a
very
big
API
with
a
tars
that
is
still
open.
So
for
now
what
is
done
is
we
are
the
the
code
is
filtering
by
the
classifier
note
that
we
get
the
Jupiter
lab
extension
classifier,
or
this
is
available
so
I'm
basically
requesting
the
extension
that
get
that
classifier.
A
Boom
thanks
again,
okay
I
have
four
things
to
talk
about,
but
I'll
try
to
make
a
quick
one
is
I,
am
working
on
removing
the
deprecated
apis
and
styles
from
luminal
and
releasing
a
2.0
so
that
we
can
time
it
with
Jupiter
lab
for
Notebook
7,
because
it's
the
least
disruptive
time
people
will
already
be
upgrading
the
versions
in
their
package.
Jsons
for
Jupiter
lab
libraries.
They
may
as
well
upgrade
the
versions
for
lumino
libraries.
The
apis
for
lumino
will
remain
mostly
the
same.
However,
the
deprecated
bits
will
be
gone.
A
A
It
is
just
one
API
and
things
like
drag
and
drop
supported
both
Mouse
events
and
pointer
events
now
they're,
just
pointer
events,
although
that
means
the
tests
are
broken
and
I
I'm,
not
confident
I
know
how
to
fix
it
yet
but
yeah.
So
please
take
a
look
at
that
if
you're
interested
in
this
part
of
the
code
base,
if
you
have
comments
or
if
you
can
help
I
would
very
much
like
it.
A
Okay,
that's
one
thing,
and
it's
a
big
one.
So
if
you
want
to
talk
about
it,
I'm
happy
to
if
not
I'll,
move
on
to
the
other
three
things.
A
Very
well,
then,
number
two
I
wrote
an
issue
in
the
team
Compass
proposing
that,
instead
of
listing
out
all
of
the
names
of
the
members
of
the
Jupiter
lab
council
with
a
yes
or
a
no
or
abstained
next
to
their
name
and
waiting
X
number
of
days
for
votes
to
come
in
with
all
of
these
incoming
votes
being
publicly
viewable
that
we
should
just
publish
the
result
of
a
vote
hold
the
vote
in
a
private
GitHub
discussion
and
once
the
voting
period
has
elapsed,
publish
the
results-
and
this
is
in
essence-
I
mean
it's
a
secret
palette
and
there's
a
reason.
A
Secret
ballots
exist,
but
more
practically
I'm,
aware
of
at
least
once
I'm
sure
it's
happened
other
times
where
in
the
Jupiter
steering
Council,
where
we
also
have
the
same
mechanism
for
voting.
Someone
felt
uncomfortable
voting
publicly
so
chose
not
to
vote,
and
we
also
see
a
similar
Dynamic
where
people
often
don't
want
to
vote.
No,
because
it's
public
and
we
sometimes
see
a
dynamic
where
people
wait
to
see
what
someone
else
has
voted
before
they
vote.
A
Now
that
one
won't
100
go
away
with
my
proposal,
because
my
proposal
is
to
have
a
GitHub
discussion,
you
could
hover
over
the
thumbs
up
and
see
what
the
names
next
to
it
are.
But
a
I
mean
imagine
we
were
a
council
sitting
in
a
chamber
somewhere,
you'd
all
see
who
raised
their
hand.
So
that's
probably
okay
and
B.
It
remains
only
the
members
of
that
voting
body
who
would
see
it.
So
that
also
seems
like
less
less
of
a
burden
on
people
than
having
it
public
Peter.
C
So
you
know
normally
it's
the
other
people
in
the
project
whose
opinion
you
care
most
about
as
well
right
there
are.
There
are
different
cases
where
you
know
the
opinions
of
the
public,
or
maybe
your
employer,
or
something
like
this
is
something
you
care
about,
so
that
this
one
definitely
fixes
that.
But
you
know
in
the
case
of
let's
say,
two
positively
minded
pull
requests
competing
against
each
other.
You
know
it's.
C
There
is
an
argument
for
transparency,
saying
that
you
should
stand
behind
your
wall
and
there's
also
an
argument
for
anonymity
being
anonymous.
So
you
can
say
that,
okay,
you
can
vote
without
fear
of
appraisals.
A
We
could
possibly
hope
to
at
people's
once
a
week,
but
but
really
I
think
that
that
represents
the
active
Community
more
than
it
represents
something
like
a
supreme
court,
where
judges
write
their
names
down
and
there's
a
reason
why
democracies
have
secret
ballots:
I,
don't
I
don't
take
lightly.
An
actual
first-hand
report
of
someone
who
said
I
simply
didn't
vote
because
I
was
afraid
of
backlash.
I
think
that
that
is
so
important
to
resolve
that,
if
the
cost
of
it
is
you
don't
get
to
see
who
voted?
A
A
D
C
The
way
of
good
I'm
just
saying
that
if
somebody
else
had
an
idea
that
would
be
directed.
F
Sure
yeah
I
think
this
is
is
a
good
at
like
First
Step,
at
least,
if
there's
not,
if
there's
a
better
one,
there's
a
better
one,
but
it
at
least
adds
the
and
like
it
makes
it.
So.
You
can
vote
within
the
council
without
having
to
worry
about
outside
contributors.
F
Essentially,
judging
you
like,
there
could
be
cases
where
you
choose
to
vote
for
what
you
believe
the
project
should
be,
and
it
might
not
necessarily
line
up
with
like,
for
instance,
what
your
manager
thinks.
Your
role
in
that
project
should
be
it
like.
There
are
people
who
put
the
project
first
rather
than
their
job
first,
and
that
could
be
problematic
on
the
council.
They
might
just
end
up
not
voting
or
voting
against
what
they
actually
want,
because
their
boss
can
look
over
their
shoulder
and
tell
them.
F
No,
that's
not
what
I
wanted
you
to
vote
for
so
I
think
at
least
that
extra
security
is
worth
it
if
yeah,
if,
of
course,
if
somebody
pitches
a
way
to
make
it
100
Anonymous,
that's
even
better
yeah.
A
I
I
I'm,
not
even
100
sure
that
we
ought
to
strive
for
that,
because
someone
will
be
administering
it.
That
is
fine
and
also
because
it
seems
to
me
like
if
we
all
worked
at
the
same
company
and
we
had
a
regular
call
or
a
regular
meeting
or
whatever,
and
we
all
sat
around
a
table
to
try
to
arrive
at
consensus,
occasionally
didn't
and
raised
our
hands
and
voted
that
that
would
be
a
fine
system
and
yet
we'd
all
see
whose
hands
are
off
so
I
don't
know.
A
But
in
any
case,
if
you
have
comments
on
this,
please
comment
on
the
issue:
I,
don't
I
hope
this
won't
be
an
issue.
We
vote
on
I
hope
it's
an
issue.
We
come
to
consensus
with
and
just
start
doing,
but
if
it
is
an
issue
we
vote
on
we'll
vote
on
it
in
the
old
way,
presumably
right,
not
the
new
way
Okay.
A
So
why
don't
I
move
on
to
the
third
thing,
so
this
I
brought
up
last
week
there's
another
team
Compass
issue
about
voting
time
frame,
I
liked
the
suggestion
that
came
out
of
a
conversation
with
Jason
grout
and
the
people
who
were
on
last
week's
governance
call
which
was
leave
the
voting
period
for
a
week.
However,
any
voting
member
could
just
write
a
comment
right
there.
That
says:
hey.
Can
we
extend
it
another
week
and
as
long
as
one
other
person
says
cool,
then
we
extended
a
second
week.
A
A
All
right,
one
last
thing
this
one
I
haven't
written
anything
yet,
but
I
was
thinking
that
there's
lots
of
ways
to
contribute.
This
came
up
in
the
community
call
yesterday
having
people
testing
things
was
a
suggestion
that
Adam
made
oh
he's
on
a
call,
cool
and
I
think
there's
lots
of
ways.
You
can
contribute
that
aren't
obvious
to
someone
who's,
just
a
user
of
the
software
and
sort
of
cares
and
wants
to
start
participating
but
doesn't
really
feel
comfortable
or
want
to
I.
A
Don't
know
open,
PRS
or
whatever
so
I
I.
Think
I'm
gonna
write
an
issue
about
hosting
this
call
because
it
doesn't
have
to
be
me.
Every
week.
I've
asked
other
people
to
host
a
few
times,
but
really
maybe
lots
of
people
want
to
so
if
we
had
a
sign
up
sheet,
I'd
be
happy
to
be
the
default
fallback
in
case
nobody
signs
up,
but
maybe
other
people
want
to.
If
this
sounds
at
all
reasonable
to
you
can
I
get
a
thumbs
up.
A
That's
it
for
me.
Thanks
for
tolerating
a
long
one,
Isabella
you're
next.
D
Isabella,
that's
me
and
I'm
also
gonna
actually
say
thanks
once
more
Dairy
and
I
really
appreciate
all
your
updates
on
this
stuff,
but
yeah.
It's
super
helpful
for
me.
I
similarly
have
some
shorter
updates,
so
one
Jupiter
lab
accessibility
call
will
be
15
minutes
after
the
scheduled
end
of
this
one
hooray.
Everyone
is
welcome
to
that.
I
do
not
know
what
is
on
the
agenda
today.
D
So,
if
you
like
surprises-
and
you
don't
know
if
you
like
accessibility,
this
may
also
be
the
call
for
you
cool
and
then
also
we
had
Jupiter
Community
call.
Yesterday,
one
of
our
discussions
was
a
little
bit
on
hosting,
because
I
know
I've
brought
up
before
people
being
like
Oh.
We
should
rotate
similar
to
what
Darian
was
just
saying
about
this
call.
D
Rotating
hosts
are
having
different
time
zones.
We
had
a
few
different
calls
at
different
time
zones
for
the
past
few
months
and
I'm
not
I'm,
not
finding
like
an
emerging
pattern
of
different
things,
but
I
think
having
more
people
involved,
might
make
some
of
that
easier,
so
I'm,
starting
to
do
an
open
call
similar
to
the
security
groups,
World
Tour,
I'm
gonna.
Try
and
do
a
world
tour
for
that.
So,
if
you're
interested
at
all
reach
out
to
me,
however,
I
didn't
put
a
link.
There.
D
I
definitely
have
the
same
username
on
GitHub
and
getter.
So
that's
probably
the
best
way
to
find
me
but
yeah.
If
you're
interested
you
don't
even
have
to
host
alone,
I'm
really
open
to
options
here,
but
I
think
it'd
be
good
to
have
some
different
people
used
to
running
Community
calls
just
so
we
don't
everything
doesn't
implode
if
I
like
have
a
cold
one
day,
but
thank
you
very
much.
F
F
Just
yesterday,
everything
else
is
Awash
yeah.
That's.
D
How
it
feels
some
days
it's
going.
Thank
you
all
for
attending
I
do
enjoy
those
calls.
A
E
Yeah
sure
thanks.
So
the
idea
of
this
discussion
was
to
list
the
major
items
that
are
still
missing
in
core
Jupiter
lab
for
starting
better
releases.
The
focus
I
had
in
mind
when
writing
those
bullet
points
was
mainly
okay.
What
are
the
features
that
are
changing
the
API,
because
I
think
that's
those
that
we
would
like
to
have
as
soon
as
possible
for
starting
better,
but
I
must
admit.
E
So
the
first
one
is
probably
lumino
V2,
because
it's
gonna
update
the
major
changes
expected
in
quarries
the
upgrade
of
typescript
with
change
the
type
exported
by
some
packages
and,
if
I
remember
correctly
from
previous
trebles,
it's
going
to
be
mainly
due
to
the
widget
package
that
will
require
some.
Some
changes
in
in
Downstream
core
shouldn't
be
too
hard
to
to
get
it
in
when
Dorian
has
finished
the
tough
work
of
actually
doing
the
work
in
lumino
itself.
E
So
that's
that's
the
first
bullet
point
and
then
in
Jupiter
rapid
itself.
So
there
are
multiple
subjects
on
RTC,
I.
Think
the
only
big
one
missing
for
the
API
is
the
the
removal
of
the
model.
Db
I
think
Kevin
is
almost
done
with
that.
I
was
just
checking
during
the
call
seems
that
there
are
still
some
UI
tests
feeling,
but
for
what
I
remember
the
content
is
there,
so
hopefully
it
should
be
soon
in
Master
on
RTC
subject.
E
There
are
other
discussion
that
probably
can
happen
in
the
beta
phases
like
both
like
save
and
auto
save
features.
Do
we
need
stable,
save
button?
There
are
some
questions
about
what
does
trust
mean
in
RTC.
We
can
maybe
set
up
also
a
bit
later
on
that
and
the
second
main
subject
is
code.
Mirror
six
I
think
we
we
can
consider
that
we
are
good
for
beta
in
the
sense
that
the
work
has
been
mainly
merged.
E
There
is
a
follow-up
issue
listing
things
that
have
been
seen
that
need
either
to
be
fixed
or
features
that
have
been
tracked
for
the
first
PR,
but
I
think
we
are
on
on
a
good
track.
The
only
thing
missing
is,
it
was
planned
to
drop
the
code
editor
if
I
see
it
and
I
know
that
nobody
is
working
on
that,
and
so
on
that
subject,
that's
mainly
the
only
travel.
Let's
say
for
starting
better.
E
Then
there
is
the
extension
manager
I've
seen
I've
demonstrated,
so
the
pr
is
almost
ready.
Hopefully
it's
going
to
be
ready
to
forward
by
the
end
of
the
week,
and
there
is
the
windowing
return
notebook,
that's
still
waiting
for
review.
E
So
if
somebody
could
do
that,
that
would
be
very,
very
nice
because
it's
breaking
some
some
API
there
is
a
new
package
linked
to
LSP,
so
some
work
has
been
done
for
integrating
what
was
in
the
LSP
extension
as
part
of
core
that
one
I,
don't
think,
is
going
to
be
a
blocker
for
beta,
because
the
ID
as
it's
a
very
recent
integration
and
without
Direct.
E
Like
feature
available
for
the
end
user,
those
API
are
going
to
be
tagged
as
a
alpha
or
beta
or
just
to
say
they
are
experimental.
A
bit
like
we
did
for
the
debugger,
so
probably
is
not
a
showstopper,
and
then
there
are
some
side.
Questions
like
like
I,
don't
know
about
the
even
best
and
I,
don't
know
also
link
to
Jupiter
server,
but
the
identity
feature
so
because,
as
far
as
I
remember
the
that's
gonna
land
only
in
V2.
E
So
do
we
need
to
wait
for
Jupiter
server
V2
to
start
to
turn
up
V4
and
there
are
also
stuff
related
to
notebook
V7,
but
I
think
they
don't
they're
mostly
of
features
related.
Small
features
related,
so
probably
won't
change
the
API
so
much
so
we
can
keep
working
on
those
items.
Even
if
we
have
started
the
beta
cycle.
So
yeah
lots
of
information,
sorry
guys,
but
that
was
to
try
to
list
the
status.
A
A
A
A
Got
a
real
flood
of
enthusiasm
here
come
on
thanks,
Mike
Awesome.
A
A
So
we
have
this
abstraction
around
a
generic
code
editor
and
the
original
idea
was
that
maybe
one
day
we'll
support
multiple
editors
out
of
the
box
and
we'll
have
this
abstraction
layer
that
allows
you
swap
out
Monaco
for
code,
mirror
five
or
code
mirror
six
for
code
mirror
5
Etc,
and
it
just
turned
out
that
if
you
really
want
performance
from
an
editor
and
if
you
really
want
to
take
advantage
of
the
things
modern
editors
do
writing
this
abstraction
layer
is
not
a
healthy
pattern
and
we
should
probably
we
should
probably
discard
it
now
that
we're
changing
so
much
with
the
code
mirror
six
week
Factor
anyway.
E
Yeah
just
mentioned
I
did
I
forget
to
add
the
issue
related
to
that
one
in
that
issue,
Johan
that
the
work
of
reporting
to
the
code
to
code
mirror
6
is
made
some
good
arguments
saying
that
some
of
the
interfaces
at
least
should
probably
be
kept,
because
it's
probably
not
a
great
idea
to
just
expose
the
full
code
mirror
6
API.
E
A
Yeah
Rick:
what
were
you
thinking.
G
Just
just
the
thought
that
actually
it's
you
can
plug
in
an
editor,
just
stand
on
a
standalone
web
server
in
an
iframe,
and
what
you
really
need
is
the
ability
to
to
read
and
write
the
file
system
from
that
editor.
And
so
what
would
be
really
really
helpful,
I
think
and
go
a
long
way
towards
solving.
This
is
just
making
sure
that
the
that
the
model
model
Factory
Studio,
editor
and
editor
if
I,
remember
those
three
classes.
G
Well,
interface
is
really
well
documented,
because
I
don't
know
that
I'm
speaking
for
myself
and
I,
probably
didn't
do
I'm
sure
that
other
people
have
done
a
part.
Better
I
had
to
spend
a
fair
amount
of
time
going
through
examples
and
trying
to
figure
out
how
to
navigate
the
code
base.
A
Yeah,
there's
a
lot
of
complexity
there,
the
the
the
the
premise
that
you
could
have
it
be
sort
of
sandboxed
in
an
iframe,
though
it
it
works
for
some
use
cases.
But
if
you
want
real
integration
with,
say,
drag
and
drop,
or
you
want
integration
with
a
LSP
or
another
completer
or.
A
Other
things
that
are
sitting
in
the
main
application,
putting
things
in
your
iframe
sandboxes
it
in
a
way
that
disallows
those
sorts
of
richer
pieces
of
functionality
added
on
top
of
it.
So
we're
kind
of
stuck
needing
to
commit
to
a
specific
implementation
of
an
editor
and
putting
it
in
the
main
application.
Instead
of
putting
it
in
a
sequestered
iframe.
G
Well
sure,
no
I'm
not
arguing
that
that
the
the
second
doesn't
preclude
the
first
right,
I
mean
you:
can
you
can
have
your
cake
and
eat
it
too,
and
support
both,
for
example?
G
There's
no
particular
reason
why
vs
code
is
why
why
Scandal
an
instance
of
BS
code
couldn't
be
integrated,
which
gives
you
an
awful
lot
of
functionality
and
the
only
interface
there
is
the
only
interface
that
you
actually
need.
There
is
this
interaction
with
the
file
system.
A
G
That's
true
I,
but
that
doesn't
argue
no
you're
perfectly
right
and
and
for
that
reason,
having
an
integrated
code
editor
is
it
offers
you
a
bunch
of
features
that
you
don't
get
from
that
I'm,
not
saying
it's
either
or
right
doing
doing
the
allowing
people
to
run
other
things
in
iframes
doesn't
doesn't
preclude
us
from
putting
in
a
generic
editor
and
improving
it
right,
and
it
may
be
that
the
that
it
will
be
a
lovely
thing
if
all
those
other
editors
would
just
go
away,
but
maybe
that's
not
going
to
happen
tomorrow.
C
I
just
wanted
to
add,
maybe
riffing
a
bit
on
what
Frederick
said
as
well,
that
you
know
if
we
had
to
expose
the
entire
gold
mirror
5
interface,
I'm
sure
the
transition
to
code
number
six
would
have
been
a
lot
worse,
because
all
of
the
extensions
would
be
using
it
right
so
there.
That
is
the
argument
for
keeping
a
limited
interface
exposed,
while
I
think
it
makes
perfect
sense
to
let
the
extensions
that
needed
access
to
full
API
and
then
have
to
deal
with
the
changes
when
it
comes.
C
A
Yeah
I
think
that's
a
fair
point
and
I
think
you
know
the
person
who
has
spent
the
most
time
working
on
this
has
been
Johan
and
if
he
arrived
that
that
sort
of
conclusion
as
well
not
a
maximalist,
get
rid
of
the
abstraction
but
not
a
minimalist,
leave
all
the
abstractions
that
we
had
from
before.
Don't
modify
them
but
somewhere
in
between
then
that
that
sounds
like
through
somewhat
painful
experience.
He's
arrived
at
that
point
of
view
as
well.
A
So
yeah
I
mean
you
know:
you're
not
I,
I
facetiously
said
it's
mostly
deleting
Fred
corrected
that
and
yeah
I
I
retract
that
there
there
will
be
some
deleting,
hopefully
the
pr
that
does.
This
will
have
more
red
than
green,
but
yeah
totally.
A
Okay,
there
was
a
lot
of
other
stuff
in
this
discussion.
Is
there
any
other
aspect
caught
anyone's
eye
that
they
wanted
to
discuss.
E
I
mentioned,
but
I
don't
know
if
it's
needed
that
So.
Currently
we
are
we're
using
Jupiter,
Observer,
V2
I,
don't
know
if
we
need
to
release
a
V3
like
I
opened
my
first
PR
breaking
the
the
API
for
the
extension
manager,
but
even
that
one
it's
removing
features
that
will
not
be
needed
any
longer
and
not
changing
the
existing
one.
So
I
don't
know
if
it's
needed
or
not
but
I
don't
know.
Maybe
people
have
other
ideas
that
there
are
stuff
to
be
change.
That
would
require
a
new
major
version
of
Jupiter
server.
C
Are
we
going
to
require
Jupiter
server
to,
or
is
that
a
thing
where
we
said
like?
If
you
have
I,
remember
the
previous
discussion,
where
we
said
we
could
support
both
Jupiter
server
one
and
two,
but
have
limited
features
if
you're
on
Jupiter
server
one?
C
So
if
we're,
if
we
require
Jupiter
server
to,
does
that
interact
with
the
decision
for
troop
to
lab
a
server
major
version
at
all
or
is
those
are
those
orthogonal
decisions.
A
I,
don't
believe
they're
related
to
each
other
I
think
they
can
be
independently,
decided
and
I'm,
not
sure
whose
use
case
benefits
from
supporting
either
Jupiter
server,
one
or
Jupiter
server
2,
because
for
the
most
part,
if
you
have
Jupiter
server
on
your
machine,
it's
because
you
got
it
as
a
dependency
and
not
as
a
thing
you
installed
yourself.
So
it
seems
as
though
we
should
simply
pick
the
newer
version
of
jupyter
server.
That
has
the
endpoints
that
we
want
is
shipped
with
that.
A
I
mean
it
seems,
like
any
other
dependency
right.
If
you
want
to
install
the
newest
version
of
IPython
and
it
happens
to
ship
with
like
a
new
version
of
Jedi,
you
probably
won't
even
notice
it.
It
should
just
kind
of
happen
when
you
upgrade
IPython
right.
A
Okay,
any
other
thoughts
on
the
V4
roadmap.
This
is
actually
I,
think
a
really
nice
encapsulation
of
some
of
the
things
that
are
open
right
now.
So
thanks
for
writing
in
front.