►
From YouTube: JupyterLab Team Meeting - November 2, 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/152
A
Today
we
have
18
people
on
the
call
and
hopefully
more
will
join
if
you
haven't
already
please
sign
in
to
the
agenda
which
is
linked
in
the
chat
just
in
case
I'll
link
it
one
more
time
and
if
there's
anything
you'd
like
to
discuss,
please
add
it
to
the
agenda
as
a
bullet
point
and
if
it's
a
bit
longer
than
just
a
few
minutes,
then
maybe
put
it
in
the
additional
discussion
section
at
the
bottom.
A
Today,
at
the
top
of
the
agenda,
we
have
some
items
from
Jeremy
who's
unable
to
join.
Is
there
anyone
who
would
like
to
go
through
these
I
can
read
them,
but
other
voices
are
nice
to
hear.
B
If
you
want
yeah
go
for
it,
so
Jeremy
is
proposing
to
add
deprecation
notice
if
people
are
using
the
command
line,
interface
of
adding
or
installing
and
installing
lab
extension
and
tidbi
in
it
is
that
since
Jupiter
3,
we
are
now
using
and
advertising
people
to
use
pre-built
extensions.
So
the
idea
is
to
bring
that
deprecation
warning
as
soon
as
possible
so
that
people
can
prepare
for
it.
B
C
I
thought
we
talked
about
this
like
back
in
Spring
and
decided.
We
were
gonna,
do
this
and
I
I
guess
in
the
back
of
my
mind,
I
thought
somebody
just
picked
it
up
and
done
it
already
and
for
those
else
that
remembered
did
we
not
make
a
decision
or
do
we
just
defer
it
and
forget
about
it
or
because
I
could
have
sworn
we
had
decided
to
do
exactly
this,
like
literally
like.
C
A
A
C
Oh
sorry,
I
was
gonna,
say
if
I
remember
correctly,
we
still
will
support
them.
There
just
won't
be
part
of
like
the
manager.
It's
like
a
soft
retirement
like
it.
They
they
still
work
they
just
we
won't.
They
won't
be
fully
supported
like
through
the
manager
like
you
can't
just
what's
the
what
was
the
term
that
they
used
in
the
like
lab
extension
install?
Wouldn't
you
couldn't
do
that
way,
but
you
could
like
still
pip
install
them.
A
C
Correct
that's
how
a
Lyra
actually
used
to
install
all
of
our
extensions.
A
year
ago,
we
didn't
actually
use
any
of
the
command
line
tools.
We
everything
up
individually
and
manually,
put
it
in
the
correct
place.
E
Another
possibility
is
the
lab
extension
install
installs,
a
pre-built
extension.
You
know
just
oh
hey.
We
have
a
visitor
who's
who's
enamored,
with
seeing
herself
on
the
camera.
E
We
could
have
Jupiter
lab
extension,
install
a
pre-built
extension
just
by
downloading
the
correct
files
from
say
a
CDN
somewhere
and
and
putting
them
on
disks,
so
there's
other
possibilities
for
a
lab
extension
install
command
that
lets.
You
install
extensions
that
aren't
necessarily
python
pit
packages
API
packages
so.
F
D
At
the
inner
version,
so
if
and
it
has
all
this
Logic
for
checking
whether
it's
compatible
with
the
Jupiter
version
and
updating
it
as
everything.
So
if
we
had
a
standard
way
of
Distributing
rebuilt
lab
extensions
over
npm
registry,
then
we
could
have
those
CLI
apis
just
work
as
they
do.
But
instead
of
saying,
like
oh
I'm,
taking
the
source
and
I'm,
including
this
in
the
build
and
whatever
it's
just
saying,
like
oh
I'm,
going
to
go
there
see
if
it
supports
this
pre-built
stuff
and
if
it
does
just
pull
it
in.
E
Yeah
yeah
or
maybe
maybe
we
don't
want
to
make
that
this
exact
CLI
that
that
sort
of
method
of
installing,
because
it
would
break
people
that
are
jumping
from
three
to
six
or
whatever.
Maybe
maybe
we
do
make
it
a
separate
CLI
command
but
I'm
just
saying:
there's
a
there's
another
Avenue
for
package,
installation
that
we've
long
contemplated
and
that
is
not
as
python
packages
but
just
as
pull
the
files
from
somewhere
and
put
them
in
the
desk
in
the
right
place,
but
as
a
precursor
to
that
plus
one
to
deprecating
The
Source
packages.
E
I,
don't
think,
we've
seen
any
pushback
from
the
community
about
needing
Source
packages
instead
of
pre-built
packages,
so
hooray
success,
yeah.
D
I
think
the
last
thing
we
said
there
was
that
it
would
be
good
if
some,
if
we
actually
got
some
numbers
on
whether
pre-built
packages
end
up
loading
a
lot
more,
a
lot
less
or
roughly
the
same
amount
of
data
in
the
browser
versus
a
compiled.
One.
D
E
G
G
I
mean
these
the,
which
is
that
it's
it's
and
I've
put
this
on
for
additional
discussion
at
the
end,
because
I'm
sure
most
people
know
how
to
do
this
and
think
it's
a
boring,
stupid,
newbie
request
and
I
I
apologize,
but
seriously
a
guide
to
to
publishing
Ascension
in
a
registry,
packaging
and
extension
and
pre-building
extension
will
be
terrific
and
better
debugging
information
on
why
this
thing
would
also
be
great.
A
A
G
Yes,
thank
you
18
and
I'm
happy
to
do
whatever
the
community
recommends.
Just
Adam
and
I'm
also
happy
to
write
up
the
experience
for
future
for
future
Extension
Office
I
I
may
be
exceptionally,
I
may
be
exceptionally
slow,
but
I'll
bet
you.
There
are
people
who
are
even
slower
than
I
am
you
might
have
to
look
hard
to
find
them.
D
A
F
A
H
Hey
everyone:
my
name
is
Andrew
I'm
working
in
iws
and
today
this
week,
I'm
gonna
call
so
because
of
that.
The
first
first
item
that
I'd
like
to
ask
help
with
is
related
to
that.
So
there
is
a
bug
that
Fernando
have
identified,
that
I
or
as
far
as
I
know,
Nolan
else
was
able
to
produce
yet,
but
that
potentially
is
a
big
problem.
It's
like
a
markdown
Lineback
and
this
line
is
not
shown
and
anything
that
is
like
in
the
same
cell
is
not
shown.
H
So
if
there
are
20
people
present
here
on
the
call
I
wanted
to
maybe
share
my
screen
and
show
how
the
bug
could
be
reproduced
and
if
you
guys
could
try
reproducing
it
it
would
help
and
if
no
one
can
reproduce
it
I
would
love
to
get
any
advice
from
the
community
on
how
to
like.
You
know
proceed
in
this
like
in
this
situation,
because
the
bug
is
potentially
critical,
but
we
can't
reproduce
it
so
far.
H
So
the
idea
is
that
you
open
any
cell.
You
change
the
type
of
the
cell
to
the
markdown,
you
type
three
dashes,
which
is
a
line
and
any
text
below
it
and
the
text
should
the
the
text
should
not
be
like
should
not
appear,
but
on
my
end
it
does
appears
so.
H
H
So
if
you
guys
could
take
a
minute,
try
doing
that
on
your
end,.
A
Yeah
got
it:
yeah
I
also
was
unable
to
produce
it,
but
and
I
tried
multiple
versions
of
lab,
just
in
case
surrounding
the
one
that
Fernando
was
using.
He
was
using
3.5.0
I
tried
in
3.4.7
and
3.5.0
and
I
tried
just
the
default
Branch
just
to
see.
If
this
is
like
an
ongoing
thing,
I
couldn't
get
to
break
in
any
of
those,
so
yeah.
If
anyone
there's
20
people
on
the
call.
A
Now,
if
anyone
manages
to
produce
this,
please
make
note
of
your
environment
and
please
write
a
comment
in
the
pr
it.
D
Would
be
good
for
to
ask
Fernando
whether
he
has
any
lab
extensions
that
are
non-vanilla
installed
that
could
potentially.
A
Yeah,
that's
right.
We
probably
should
have
asked
them
about
his
extensions.
We
tried
to
blame
his
browser
extension
setup,
but
it
turns
out
that
he
tried
it
in
all
his
browsers
and
manage
to
produce
the
bug
which
was
wild,
so
it
does
probably
indicate
yeah.
Maybe
one
of
his
actual
Jupiter
lab
extensions
is
doing
it.
H
Is
there
so
potentially
one
could
ask,
are
there
any
other
commands
to
get
like
the
environment?
The
list
of
currently
installed
extensions
with
one
command,
another
lab
extension
list,
Jupiter
server
extension
list?
There
is
Jupiter
version
and
so
on.
That's
probably
the
best
approach
right
to
get
the
environment
yeah.
D
H
E
One
great
comment
on
this
I
did
a
quick
search
on
for
the
horizontal
line,
the
words
horizontal
line
and
Jupiter
lab
issues
see
if
we've
ever
encountered
anything
like
this
before
and
at
least
one
pointed
to
the
virtual
notebook
rendering
so
I'm
curious.
If
what
settings
he
has
for
the
notebook
and
in
particular,
does
he
have
virtual
notebook
cells
on
or
there
may
be
other
settings
and
I
think
did
he
ever
give
us
the
Jupiter
lab
version
that
he
was
dealing
with?
E
There
have
been
bugs
that
look
like
this,
as
in
cells
kind
of
Disappearing
into
a
horizontal
line.
Oh
sorry,
that
word
that
were
old
versions
of
3.1,
something
or
other.
E
Yeah
so
another
place
to
look:
is
his
notebook
settings
and
making
sure
that
whoever's
reproducing
has
the
same
notebook
settings
in
particular
the
virtual
rendering
or
something
like
that,
and
it
would
be
helpful
if
he
just
recorded
a
short
video
if
he
could
reproduce
it
with
a
clean
notebook
with
a
single
cell
Etc
I
mean
that's
what
we're
trying
to
reproduce
or
if
there's
like
a
number
of
things
that
he's
doing
before
this
happens,
it
would
be
helpful
if
you
just
had
a
short
video,
so
we
could
see
exactly
what
steps
he's
doing.
H
Thank
you
Jason.
Could
you
send
me
a
link
to
this
issue
if
it
is
open
for
you?
If
it's
not
it's
fine,
the
one
you
were
talking
about
the
similar
one,
So
yeah.
Thank
you.
H
So
if
no
one
has
any
inputs
in
terms
of
that,
I
wanted
to
ask
about
advice
on
the
other
thing,
so
the
other
thing
is
I'm
trying
to
implement,
drag
something
using
notebook,
drag
and
drop
capabilities
in
Jupiter
lab
and
I'm
seeking
good
good,
but
not
too
complex
of
an
example
of
an
extension
or
of
Jupiter
lab
like
core
package
that
uses
notebook,
drag
and
drop
capabilities
in
not
to
off
a
complicated
way.
If
anyone
has
any
advice
on
this,
please
let
me
know.
I
I
might
be
able
to
I
guess
recently,
I
bumped
him
to
a
park
in
lumino,
where
there
are
several
packages
in
Jupiter
lab
which
reuse
lumino,
data
track
and
drop
functionality.
So-
and
these
are
not
only
in
the
notebook
but
also
the
file
browser
uses
dots.
So
maybe
the
file
browser
has
more
self-contained
example.
So
there
is
a
an
issue
which
I
filed
that
it
actually
turns
out
to
be
very
the
degrading
performance
on
chromium
browsers.
I
So
maybe
we
could
actually
look
into
re-implementing
track
and
drop
using
native
browser
mechanisms
rather
than
using
the
luminal
stuff,
which
is
not
wrong
in
itself.
It's
bugs
in
Chromium
browser
which
are
contributing
to
the
sound
dollar.
I
D
That
is
just
my:
has
there
been
any
improvements
on
the
browser
native
drag
and
drop
that
we
think
we
can
replace
the
Lumina
one,
though.
I
There
are
limitations
of
the
Native
support.
You
can
do
a
lot
of
things
with
the
native
drivable
attributes.
You
can
listen
to
them
on
drop
drug,
started
and
track
motion.
But
during
the
motion
you
cannot
render
elements
the
browser
takes
control
of
that.
So
you
can
only
replace
the
ghost
image
with
another
image
rather
than
the
one
generated
by
the
browser,
but
you
wouldn't
have
the
full
control
like
in
Lumina
when
you
are
moving
it
up
to
another
dock
area.
I
You
get
this
blue
outline,
which
is
nicely
showing
where
the
top
will
land
we
cannot
get
that
easily,
but
for
file
browser
where
you
are
just
moving
a
node
with
the
file
name
and
an
icon,
you
can
get
the
same
behavior
from
the
Native
browser
support.
I
H
Okay,
good,
thank
you
for
the
discussion.
I
will
take
a
look
at
the
file
browser,
for
example.
So
that's
it
for
me.
Thank
you
very
much.
A
Cool
thanks,
Angie
I
have
a
quick
one.
So
if
you
are
a
member
of
the
Jupiter
lab
Council
or
any
other
sub-project
Council
or
a
member
of
a
standing
committee
or
working
group
that
has
a
named
Council
in
Jupiter,
then
you
probably
already
got
an
email
from
Fernando
talking
about
the
Jupiter
executive,
Council
election
and
nominations
for
it.
A
I
just
wanted
to
describe
a
little
bit
what
that
is
and
encourage
you
to
apply
if
you've
been
holding
back.
So
the
new
governance
structure
of
Jupiter
as
I'm
sure
you're,
probably
tired
of
hearing
by
this
point,
but
some
of
you
may
not
have
heard,
is
broken
up
into
multiple
groups
that
are
to
reflect
the
separation
of
concerns
that
it
exists
in
the
project.
So
there
are
named
sub
projects
like
Jupiter
lab
or
Jupiter
Hub
or
Jupiter
notebook.
A
There
are
standing
committees,
there
are
some
working
groups
and
each
of
these
has
their
own
group
of
people
who
work
on
them
regularly.
So
those
people
are
have
always
unofficially
been
the
decision
makers
of
those
projects
and
now
there's
just
a
bit
of
official
veneer
on
top
of
it,
to
make
it
clear
to
someone
who
may
not
be
familiar
with
the
project,
how
did
decisions
get
made
and
for
people
who
can't
come
to
every
single
call?
A
It's
clear
who's
going
to
vote
on
things
and
every
person
who
is
involved
in
one
of
these
projects
and
is
in
one
of
these
voting
bodies
is
eligible
to
run
for
the
Jupiter
executive
Council,
which
is
the
group
of
people
that
the
standing
committees
and
working
groups
are
managed
by,
and
is
the
group
of
people
who
are
in
charge
of
things
like
fundraising
or
putting
on
events
or
handling
code
of
conduct,
violations
or
lots
of
project-wide
things
that
are
not
strictly
software
or
technical
decision
making,
and
that
body
is
to
have
six
people
in
it.
A
A
That's
why
I'm
saying
to
us
and
three
people
are
meant
to
be
elected
and
for
the
very
first
election,
the
people
who
are
elected
will
have
one-year
terms
after
that
everyone
will
have
two-year
terms
so
that
each
year
half
of
the
council
is
up
for
election,
and
so
the
commitment
that
we're
asking
people
who
are
elected
to
this
office
make
is
is
fairly
modest.
It's
two
to
four
hours
a
week
and
it's
for
one
year
and
the
nomination
form
is
fairly
terse
as
well.
A
A
So
if
you
are
one
of
the
people
who
is
eligible
to
run
for
this
and
you're
interested
and
you've
been
holding
back,
please
stop,
please
feel
free
to
nominate
yourself
if
you
want
a
little
bit
of
help
or
you
want
some
questions
that
have
not
been
answered
and
what
I
just
said
or
in
the
emails.
Please
feel
free
to
contact
me
I'm,
happy
to
help
you
with
it
and
if
there's
any
so
the
same
thing
I
can
answer
now
any
questions
that
have
come
up
in
your
mind,
I'm
happy
to.
B
E
I
noticed
in
the
notes:
there's
no
mention
of
Jupiter
Khan.
So
if
you
haven't
heard
already
Jupiter
con
2023
is
on
there's
a
website.
It's
a
preliminary
website.
You
can
see
in
the
notes,
jupitercon.com
or
jupiter.org
has
a
banner
for
Jupiter
con
as
well.
So
it's
in
Paris,
May
10th
through
12th
next
year,
I
really
hope
to
see
everyone
there.
It
would
be
really
fun
to
get
together
and
it
should
be
a
great
time
to
share
and
and
and
participate
and
learn
from
the
entire
community.
E
The
second
just
follow-up
from
last
week,
I
created
the
repo
that
we
talked
about
last
week,
invite
anybody
that
wants
to
contribute,
or
even
just
have
better
suggestions
for
the
plug-in
name
or
the
mime
type.
Thanks.
B
F
B
B
So
that's
one,
that's
the
like
the
main
change
and
the
other
one
and
from
that,
incidentally,
is
that
we
should
allow
empty
notebook,
so
empty
notebook
are
may
happen
only
if
so
in
3.6.
Only
if
the
user
is
in
collaboration
mode
and
only
if
one
or
multiple
users
decided
that
all
cells
should
be
deleted,
so
it
can
be.
B
If
it's
the
same
user,
that
is
deleting
all
cells,
then
there
will
be
an
additional
cell
that
will
be
added,
but
the
empty
notebook
can
happen
if,
like
there
is
two
sides
and
the
two
user
delete
the
separate
cells
and
then
that
will
be
an
empty
notebook
and
so
the
the
user.
We
see
a
small
node
instead
of
just
having
a
blank
page
to
say:
okay,
the
notebook
is
empty,
please
add
a
cell,
but
it
should
be
a
rare
case
as
also
creating
a
new
notebook
will
be.
The
new
notebook
will
be
with
an.
F
B
Cell
as
it
used
to
be
so
that's
the
main
changes,
there's
been
another
changes
and
related
to
Jupiter
server,
2..
So
to
the
collaboration.
We
need
a
better
support
for
ident
identity,
and
this
is
provided
by
the
coming
Jupiter
server,
V2
and
so
there's
been
an
extension
of
the
service
package
to
cover
that
new
API
part
of
Jupiter
server.
So
this
is
just
to
be
compatible
with
Jupiter
server
V2.
But
so
that's
the
question
I
have
for
later
about
which
version
of
jupyter
server
do
we
want
to
to
support
for
3.6.0?
B
And
finally,
there
has
been
a
teeny
changes
related
to
the
usage
of
fi
ID
service,
so
for
people
that
are
not
aware
of,
it
has
been
a
work,
funded
and
and
done
by
AWS
and
many
David,
but
I,
don't
think
he's
on
the
call.
B
The
the
main
trouble
when
you
are
doing
collaboration
is
if
the
if
a
file
you
are
working
on
is
rename
and
we
need
to
be
able
to
catch
that
and
fire
ID
Services
the
way
around
it.
So
basically,
the
the
server
is
tracking
a
file
by
a
unique
ID,
and
this
ID
is
independent
of
the
of
the
file.
One
is
when
a
session
is
is
live,
let's
say
so
that's,
but
it
doesn't
impact
the
the
API
in
Jupiter
lab
itself.
B
It's
just
the
there
were
some
tiny
changes
needed
for
that,
so
there
is
no
breakage
of
the
API.
For
that
specific
thing,
oh
no
I'm
wrong.
There
is
one
the
we
have
dropped,
the
rename
the
rename
features
of
the
doc
provider,
but,
okay,
all
all
that
work
is
done
with
one
issue
that
I
may
have
inserted
on
the
evening.
B
So
there's
one
thing
that
I
need
to
look
at,
but
so
far
the
latest
of
to
the
alpha
one
that
has
been
released
apparently
is
quite
stable,
and
so
that
work
is
mainly
done
and
there
has
been
other
PR
listed
here.
So
the
notification
queue
and
display
of
those
notification,
so
I
have
implemented
it
in
master.
B
The
idea
is
to
backport
it
to
3.6.0
and
to
add,
on
top
of
that
and
announcement
features
that
I
will
just
present
a
bit
in
detail
afterwards
with
a
demo,
and
there
is
this
PR
that
Mike
has
opened
I.
Think
already
two
weeks
ago,
yeah
October
11,
but
I
think
we
should
move
forward
with.
Ivory
is
also
a
new
version
of
luminos.
So
this
one
but
performance
issue
could
make
it
too
easily.
There's
been
other
small
small
thing
like
Microsoft.
B
Take
the
chance
of
reviewing
some
stars
to
use
more
icons
in
the
settings
to
make
the
plugin
stands
out
a
bit
more
than
just
having
so
many
with
with
a
cop
and
there's
been
small
things
about
style
for
the
border
and
there's
been
some
work
done
by
Nicola
from
constack
for
improving
a
bit
some
behavior
when
you
are
doing
single
document
opening.
So
basically
the
idea
is
to
not
load
all
documents.
B
So
there's
been
this
deprecation
that
is
spoken
about.
There
is
one
new
contributors
that
want
to
add
some
options
for
the
sanitizer
I
think
it's
gonna
be
able
to
to
make
it
for
3.6
and
I.
Think
this
one
is
also
by
Mike
about
Improvement
for
the
performances
and
the
CSS
style.
So
this
is
the
list.
If
you
are
thinking
about
something,
that's
missing
and
you
want
it
into
3.6.
B
I
would
like
people
to
edit
like
before
the
end
of
this
week,
because
ideally
I
would
like
to
start
doing
features
for
them
so
betas
next
week
and
because
otherwise
it's
gonna
be
difficult
to
release
3.6
before
end
of
the
year.
So
that
means
that
it's
going
to
be
difficult
to
have
four
after
that,
before
the
end
of
the
year,
so
tid3
that
3.6
will
be
the
the
last
major
version
on
the
3x
branches,
and
so
coming
back
on
my
point.
So
the
there's
three
questions
to
that.
B
So
the
first
one
is:
should
we
pins
a
Jupiter
server
2
for
with
3.6.0
and
Pro?
Is
that
for
collaboration?
We
need
the
identity
API.
So
that
would
be
nice
if
we
enforce
that.
But
as
it's
the
collaboration
entry,
it's
behind
the
flag,
the
other
possibility
would
be
to
implement
a
runtime
dependency
check
and
say:
okay,
if
you
want
to
use
jupyter
lab
with
collaboration,
probably
you
should
use
those
version
and
we
should
stop
for
that
specific
case,
but
still
keeping
the
the
compatibility
with
chapter
server.
B
One
probably
as
Jason
is
writing
yeah.
The
idea
would
be
to
to
keep
compatibility
with
V1
the
the
advantage
of
that
is
that
people
will
be
able
to
to
update
to
patch
release,
also
with
bug,
fixes
and
stuff
like
that,
without
without
fear,
so
I
think
we
probably
cannot
go
the
second
path,
but
I
would
like
for
people
to
some
some
feedback
and
then
I'll
go
back
to
the
other
two
points.
H
E
Theater
does
have
a
good
point
here
that
it
may
not
be
need
to
be
a
requirement,
but
perhaps
we
could
say
there's
a
warning
in
3.6
if
you
have
an
old
version
of
Jupiter
server,
that
renaming
doesn't
work,
but
other
things
should
work
the
same
as
before.
In
other
words,
your
experience
isn't
degraded
yeah.
B
I
I
want
maybe
a
bit
far
a
fast
story
here,
for
that,
so
definitely
for
the
file.
Id
thing
it's
true
that
we
can
just
make
a
warning.
The
trouble
is
more
with
the
Jupiter
server
version
that,
if
you
don't
have
identity
provider,
then
it's
gonna
be
weird
for
the
user.
If
he's
using
collaboration.
D
B
I
Like
how
would
that
potentially
affect
rookdale,
Hub
and
binder,
because
I
know
that's
previously,
when
RTC
was
released
for
the
first
time
there
was
some
questions
on
whether
it's
compatible
and
pinning
a
Jupiter
server
might
potentially
cause
some
issues.
Was
that
tested.
B
It's
not
compatible.
The
trouble
of
binder
is
that
it's
assuming
that's
notebook
is
executed.
So
actually,
if
you
want
to
use
Jupiter's
server
2,
you
cannot
use
binder
easily
and
let
me
go
back
to
Jupiter
lab
I
had
to
to
modify
our
binder
settings.
So
basically
we
need
to
trick
the
start
of
Jupiter
notebook.
To
say
we
don't
start
Jupiter
notebook,
but
we
start
guitar
lab
because
yeah
Jupiter
binder
doesn't
work
with
Jupiter
server,
V2.
B
B
B
Have
other
issues
right
yeah,
then
the
other
thing
is
so
for
now
we
have
the
short
models
in
inside
Jupiter
lab
core,
and
the
trouble
is
that
now
we
have
two
place
where
we
Define
the
schema
of
a
document.
So
we
have
the
typescript
version
that
are
in
short
model,
but
we
have
also
a
python
version.
B
That's
in
the
back
end
in
Jupiter
white
Dock,
and
what
I
would
like
to
propose
is
to
extract
those
sharp
model
to
be
at
the
same
place
than
the
python
description
of
the
schema,
so
that
we
can
more
easily
maintain
the
consistency
between
the
two
schemas,
the
other
ID.
Also
behind
it
is.
It
will
be
easier
for
third
party
to
connect.
There's.
A
B
A
Don't
in
theory,
think
that's
a
bad
idea,
but
in
practice
the
Jupiter
org
just
has
a
lot
of
things
that
are
difficult
to
organize
in
it.
So
the
fact
that
it's
in
server
now
kind
of
makes
it
easier
to
figure
out
and
I
would
suggest,
leaving
that
there,
but
as
far
as
moving
shared
models
to
server
I
think
that's
not
unreasonable
and
I.
Imagine
well
I'll.
Ask
again:
does
anyone
have
an
objection
to
this.
E
I
would
mention
one
other
thing.
Do
we
expect
this
to
become
a
standard
on
the
level
of
Jupiter
notebook
format
and
Jupiter
kernel
messaging
protocol.
A
E
So
eventually,
there
is
a
new
subproject,
the
Jupiter
standard
subproject,
which
nominally
maintains
the
repos
Jupiter
client
and
B
format.
These
repos
that
sort
of
are
encoding
our
standards,
and
so
maybe
eventually
it
makes
it
into
the
Jupiter
standards
project
as
sort
of
a
lower
level
standard
but
yeah.
If
it's
not
to
that
point
yet
then
carry.
B
On
yeah
I
would
say
it's
not
yet
at
that
point
actors:
okay,
thanks!
So
then
I
I
ask
on
the
Jupiter
server
side
tomorrow
and
then
finally,
just
have
a
couple
of
minutes.
Sorry,
it's
long,
but
just
a
demo,
so
I
hope
you
can
all
see
my
screen.
So
some
weeks
back,
I
have
implemented
the
ability
of
doing
notification.
B
So
basically
you
get
a
cure
in
which
you
stack
notification
and
there
is
a
plugin,
that's
listening
to
that
queue
and
you
can
either
display
directly
in
with
those
the
notification
or
having
them
display
in
a
in
a
kind
of
Center
notification
center.
And
so
what
I
have
implemented
is
an
announcement
features.
So
here
it
is
when
a
Jupiter
app
is
launched.
B
So
by
default
it
will
fetch
an
RSS
feed
and
for
for
all
post
it
will
create
a
notification
and
it
will
need
for
it
to
tune
if
it's
on
all
posts
or
whatever,
but
that's
the
ID,
and
the
idea
is
to
use
that
to
be
able
to
connect
more
easily
with
our
community
like
to
mention
that
there
is
a
big
new
version,
that's
coming
and
like
if
we
want
to
to
do
another
survey,
as
we
did
like
two
or
two
years
back
now,
and
to
be
to
be
nice
to
to
use
that
features
too,
because
it's
it's
one
of
the
trouble
we
always
have.
B
We
have
plenty
of
communication
kind
of
channels,
but
probably
the
best
is
in
the
application
directly
and
yes,
as
Jason
is
written
on
the
note
that
I'm,
seeing
in
parallel,
it
is
what's
missing,
is
what
I
want
to
add
a
button
to
say:
yeah,
don't
show
me
that
again
and
that's
the
thing
that
I
want
to
show-
and
there
is
another
thing
that's
not
displayed
here
is
that
is
checking
for
updates
too.
B
And
the
idea
is
that
all
of
those
are
customizable,
so
that's
mean
that
administrator
can
set
up
server
settings
to
turn
the
both
features
off
and
user
can
decide
that
I
don't
want
to
be
buffered
by
that
any
longer
and
so
to
just
give
a
an
ID.
So
this
is
the
notification
center.
In
case
you
haven't,
you
have
missed
the
peer
about
it.
F
B
D
D
B
I'm
not
sure
because
you
are
just
asking
for
RSS
feed,
so
there
is
no
way
we
can
store
any
information.
The.
A
Sure,
but
we
could
decide
to
host
it
on
jupiter.org
or
something
that
we
know
isn't
collecting
it.
But
the
point
the
point
of
opting
in
is
good,
even
if
it
had
zero
privacy
implications,
because
you
might
want
everything
inside
your
Jupiter
instance
to
be
air
gapped
and
not
connected
to
the
internet
and
all
that
so
I.
Think
opt-in
in
general
is
a
good
idea.
E
D
Just
wanted
to
add
one
last
feedback
that
that
that
question
should
probably
then
be
a
a
notification,
as
he
said,
and
not
the
dialogue
right,
because
everybody,
whoever
starts
up
a
Jupiter
light
or
a
binder,
will
see
this
question
every
time
right,
because
it's
a
fresh
instance
without
any
instance.
So
it
should
be
a
discrete
question.
A
A
Okay,
just
in
the
interest
of
at
least
getting
to
hear
Rick
out,
even
though
we
don't
have
enough
time
for
discussion.
Sorry
Rick,
but
you
are.
G
Okay,
very
briefly,
so
here's
the
extension
and
what
it
does,
a
click
it
probably
should
have
opened
ahead
of
time.
What
it
does
is
it
loads,
an
external
dashboard
editor
which
we
wrote,
the
the
and,
as
you
can
see
as
it
opens
it,
opens
this
up.
It
follows
the
editor
protocol
and
opens
up
a
gd.json
file
in
the
in
the
in
in
the
file
browser
and,
as
you
update
it,
it
continuously
updates
that
here's
a
more
fully
formed
one
now.
What
the
so
that's
the
extension.
G
It's
it's
it's
a
couple
of
things.
It's
pure
typescript
front
end
only
there's,
no
there's
no
back
end
or
python
stuff.
Here,
that's
it's!
As
you
can
see
it,
it
does
menus.
It
also
puts
an
icon
on
the
launcher
and
it's
generally
designed
actually
that
the
only
thing
that
you
have
to
do
to
add
any
other
external
editor
to
this
thing
is
provide
a
new
icon
and
provide
a
and
provide
a
URL
to
provide
a
new
icon,
provide
a
file
type.
G
You
know
a
file
extension
and
three
provide
a
new
web
page
where
the,
where
the
actual
editor
lives
now
I've
been
I've,
been
installing
I've
been
running.
First
of
all,
it's
it's
completely
open
source,
I'm,
entirely
happy
to
publish
it
to
the
world.
I
just
need
to
to
I
just
need
guidance
on
on
exactly
how
to
do
that
and
what
the
end
and
how
to
and
how
we
want
to
be
good
members
of
the
community
and
I
want
to
be
to
and
I
just
want
to
follow
best
practices.
G
Here,
that's
one
two:
we've
I've
had
the
thing
compiling
it.
For
a
year,
I
haven't
been
able
to
do
a
pre-built
extension,
install
it
I
have
to
build
the
thing
in
and
and
and
install
it
from
Source
every
time,
but
as
of
just
yesterday,
it
suddenly
stopped
working,
and
the
error
message
I
got
back
was
on
the
build
well
build's,
fine
and
then,
when
I
try
to
to
run
it
that
says
I
the
metadata
wasn't
found
in
the
registry.
G
So
clearly
our
our
publishing,
our
publishing
and
packaging
of
this
extension
is
poor.
I
would
like
to
do
it
properly
and
I
would
also
like,
assuming
that
I'm
not
uniquely
cursed
to
to
write
up
to
how
to
do
it,
so
that
other
people
don't
run
into
the
same
problems.
I
have
and
I'd
like
help
just
like
guidance
and
and
oh
gee.
Here's
what
you're
doing
wrong
and
I
do
promise
to
give
back
by
writing
it
up.
That's
it.
A
Cool
so
I
think
in
this
particular
case.
If,
if,
if,
if
there's
anyone
on
the
call
who's
recently
done
this,
perhaps
someone
could
volunteer
to
look
at
it
with
you
in
in
the
general
case,
for
people
who
are
starting
from
scratch,
the
cookie
cutters
have
improved
now
they
ship
with
a
release,
dot
MD
file.
That
shows
like
what
how
to
release
the
newly
created
package
that
you've
made,
but
that
doesn't
help
you
in
your
current
circumstances,
because
you've
already
built
this
thing.
It's
already
working.
Now
it's
stopped.
A
So
that's
that's
sort
of
like
it
would
be
kind
of
a
waste
of
time
to
start
rebuilding
the
wheel
here.
A
There's
an
extension
migration
guide
that
Fred
just
linked
in
the
chat
that
might
be
helpful
for
you,
especially
because
there
was
a
a
script
that
we
wrote
to
help
people
migrate,
their
extensions
from
old
style
to
new
style
that
might
be
beneficial
to
look
at
as
well,
but
I
don't
know.
Is
there
anyone
on
the
call
who
maybe
has
some
time
to
look
at
this
with
Rick
or
if
not
to
is
there?
Is
there
a
place?
A
G
If
I'm
I
may
just
start
just
put
it
into
the
extension
report,
but
you
know
following
your
advice
earlier
in
the
call
I've
downloaded
the
the
the
new
cookie
cutter
and
what
I
think
I
might
just
do
is
migrate.
The
sources
over
to
that
and
the
package.json
file
and
see
what
and
see
if
I
can
get
it
to
work.
A
G
A
Yeah
I
mean
I
I.
Think
that,
honestly,
if
I
were
in
in
your
situation,
I
I
think
that's
what
I
would
do,
but
if
it
turns
out
that
you
get
stuck,
there
then
come
on
back.
The
other
thing
I
would
suggest
is
and
I
know
it's
not
actually
meant
for
Jupiter
lab,
but
I'm
sort
of
co-opting.
Anyway,
at
this
time
on
Thursdays
on
the
same
channel,
there's
a
Jupiter
contributing
hour
that
typically
we
do
Jupiter
server
stuff.
G
Thank
you,
while
trying
so
I've
got
I'm
stacked
up
with
meetings
today,
but
I'll
try
and
and
solve
it
on
my
own
today.
If
I
can't
get
it
done,
I'll
well,
I'll
probably
start
coming
to
that
meeting
anyway,
because
we
have
other
Jupiter
server
issues
that
we'd
like
to
that
we'd
like
to
discuss
and
I'd
actually
like
some
of
the
folks
on
my
team
to
come
to
the
well
one
of
the
folks
on
my
team
would
not
only
Not
only
would
I
like
him
to
come
to
it.
G
He'd
love
to
come
to
it
cool.
By
the
way
I
noticed
there
was
an
accessibility
meeting
just
15
minutes
after
this.
G
This
is
completely
off
topic,
but
I
was
talking
to
Shauna
dark
who's,
the
head
of
undergraduate
CIO
for
undergraduate
instruction
at
UC,
Berkeley
they've,
taken
over
maintaining
the
Berkeley
data
Hub
and
one
of
the
things
she
told
me
was
that
she
didn't
want
to
be
responsible
for
doing
accessibility
there
and
she
didn't
so
I
suggested
a
way
that
she
talked
to
her
team
and
somebody
should
I
mean
I
can
reach
out
to.
Is
there
a
person
that
I
could
Point
Shauna
at
who
knows
how
to
do
this
stuff?
A
No
is
a
strong
term,
but
that's
I
would
suggest
actually
attending
one
of
our
accessibility
calls
and
and
pitching
it
there,
but
if
not
I'm
I'm
in
that
group,
I'm
happy
to
answer
questions
Gabriel
on
this
call
he's
in
that
group.
I,
don't
know
who
else
on
this
call
right
now
is:
oh
Martha
is
there's
a
few
of
us
who
and,
and
so
is
Ely
like
a
a
group
of
us
on
this
call
are
usually
on
that
call
too
so,
but
yeah
the
best
place
to
get
a
sort
of
live.
A
Rapid
feedback
is
to
actually
just
come
to
the
call,
but
if
not,
you
could,
if
you
have
a
specific
issue,
you
want
to
talk
about
open,
initial
and
tag.
The
group
of
people
that
I
just
mentioned,
or
you
could
put
you-
could
drop
something
in
the
accessibility
team
Compass
because
we
do
have
a
repo
for
communication
and
yeah
I.
Don't
know
this.
G
Is
an
issue
I'm
passing
on
for
print,
though
I
guess
we
should
be
concerned
about
it
too.
So
I
will
give
her
those
reporters
and
thank
you.
Okay,.