►
From YouTube: JupyterLab Team Meeting - August 24, 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
Okay,
welcome
to
the
August
24th
weekly
Jupiter
lab
call.
Today
we
have
15
people
on
the
call
and
please
find
the
agenda
linked
in
the
chat.
Please
feel
free
to
add
your
name
if
you
would
like,
please
feel
free
to
add
a
an
update
to
the
agenda
section
and
if
there's
something
you'd
like
to
talk
about
that's
going
to
take
more
than
say
five
minutes,
please
add
it
at
the
bottom.
In
the
additional
discussion
section,
I
can
tell
you
right
now.
A
B
The
I
have
three
PRS
that
are
in
need
for
review,
so
if
people
have
sometimes,
we
appreciate
some
review
first,
one
is
small:
just
try
to
increase
some
coverage
for
translation.
B
The
second
one
is
a
features
that
was
implemented
by
Jessica
if
I'm
correct,
but
was
reverted
because
of
some
side
effects
so
did
is
to
to
prompt
the
user
for
renaming
the
Untitled
files
when
the
user
is
saving
the
files.
The
first
time
manually,
I
took
the
chance
also
to
add
or
to
address
another
issue.
B
That's
streaming
dialogue,
meaning
that
is
to
add
the
ability
of
having
a
kind
of
checkbox
in
dialog
that
you
can,
for
example,
use
for,
saying,
don't
show
me
again
this
dialog
and
you
can
use
the
the
result
from
the
dialog
to
get
the
value
of
the
dialogue,
but
also
the
value
of
the
checkbox
to
decide
what
to
do
with
and
the
last
one
is
the
new
extension
manager.
So
I've
already
done
some
modification
following
review
and
so
I'm
like
waiting
for
a
final
review
of
that
one
and
following
some
question
on
guitar.
B
A
nice
side
effect
of
the
new
extension
manager
is
that
there
isn't
there's
gonna,
be
no.
A
new
read-only
managers
that
a
user
can
use
and
the
read-only
manager
will
allow
the
people
to
list
all
extension
that
are
available
and
we'll
be
able
to
disable
or
enable
it,
but
without
the
ability
of
installing
any
new
extension.
So
it's
it's
a
bit
like
what
you
get
in
the
NB
extension
configurators
for
the
classical
notebook
in
which
all
extension
are
with
checkbox
that
you
can
check
in
and
check
and
then
refresh
the
page.
C
Yeah,
just
for
the
the
check
box
for
don't
ask
me
again
so
I
assume
that
that's
like
opt-in
for
all
the
dialog
boxes
you
want
to
use
I.
Did
you
only
add
it
for
this
dialog
box,
or
did
you
like
do
a
review
for
other
dialog
boxes?
Sorry,
that's
like
a
good
follow-up
to
figure
out
which
dialog
boxes
to
add
this
option
to.
C
B
B
I
would
say
for
for
cure,
but
I
definitely
agree
there
are.
There
are
a
couple
of
existing
dialogue
that
could
benefit
from
it
like.
Should
you
want
to
shut
down,
or
do
you
want
to
restart,
confirm
and
stuff
like
share
a
picture
yeah.
C
Definitely
agree,
that's
also
why
it's
in
that
issue
that
that
they
agree
is
it
would
be
fixed
by
this
PR,
but
there's
definitely
a
clear
follow-up
other
than
that
once
you
click
remember
what
the,
if
you
want
to
have
it
forget
what
this
remembers,
how
do
what
you
need
to
do.
B
C
D
C
Sure,
I
just
yeah
the
worst
case
scenario
is
that
if
it's
remembering
some
information,
then
you
cannot
clear
that
memory
somehow
without
you
know,
going
to
delete
all
your
settings
folders
and
then
searching
for
it.
So
as
long
as
it's
documented
how
to
clear
it
out
if
everything
goes
wrong
with
it,
that's
that's.
Okay,.
B
A
B
B
A
C
My
open
source
time
is
completely
filled
up
with
hyper-weighted
stuff.
This
coming.
A
C
A
C
A
Fair
again
the
agenda's
light
so
I
can
wait
uncomfortably
to
see
if
anyone
volunteers
to
review.
A
Cool,
please
could
you
look
at
the
second
one.
A
B
The
other
thing
are
more
information,
so
I
still
have
two
big
ones
that
are
I'm.
Reviewing
or
finishing
up
is
the
in
the
the
Takeover
of
the
pr
of
Kevin
about
removing
model
DB.
So
everything
is
green,
except
memory
leak.
B
So
when
we
test
creating
a
notebook
and
deleting
The
Notebook,
then
the
notebook
object
is
still
in
memory,
so
I
removed
that
a
couple
of
weeks
ago,
so
I
don't
want
an
appear
to
to
the
to
to
undo
that
or
to
to
conduct
to
some
memory
league
so
and
I'm
fighting
to
figure
out
what's
going
on,
but
eventually
I
will
find
that
and
then
I'm
fighting
with
force
of
the
rebase
of
the
windowing
of
the
notebook.
B
Because
now
there
is
this
part
of
the
LSP
features
that
are
in
core
and
so
I
need
to
rethink
a
bit
more.
Some
part
of
the
RSP
API.
The
reason
for
that
is
if
a
notebook
is
is
virtual
or
windowed.
You'd
have
the
editors
only
for
the
service
that
has
been
in
viewport
and
up
to?
No,
it
was
not
the
case,
so
the
LSP
extension
is
making
the
assumption
that
they
can
request
the
editors
for
every
cells
of
the
notebook
even
and
so
I
need
to
yeah
change
that
a
bit.
B
So
it's
gonna
take
some
time
and
those
who
and
those
two
are
important
for
V4.
That's.
Why
I'm
mentioning
that
here
and
the
last
thing
is
just
to
mention
so
next
week
is
our
sci-fi.
We
have
a
poster
for
presenting
features
of
Jupiter,
app,
4
and
notebook
7,
and
this
is
gonna,
be
plenty
of
talk
about
python
in
the
browser,
so
there's
gonna
be
one
by
Jeremy
and
I.
Think
thorsten
of
two
on
Jupiter
late
night.
B
A
A
A
Again,
I
will
exhort
you
if
there's
anything
you
want
to
talk
about.
Please
add
something
to
the
agenda
because
right
now,
it's
only
me
remaining
I
have
been
working
on
lumino
2
and
as
of
a
few
days
ago,
there
is
a
set
of
lumino2
Alphas
out.
So
the
last
day
or
two
I've
been
working
on
the
Jupiter
lab
for
luminal
migration.
A
It
is
ready
for
review
and
please
take
a
look,
because,
even
if
you
don't
want
to
do
a
review,
you
might
still
benefit
from
just
seeing
what
sort
of
changes
are
coming.
Lumino2
has
a
migration
guide
that
we're
working
on
as
we
change
things.
A
So
let
me
actually
link
that
migration
guide
as
well,
so
that
gives
a
partial
basis,
but
a
this
guy
doesn't
flux
MB,
it's
not
complete,
yet
it's
just
more!
So
we
don't
forget.
Where
is
the
minute?
Oh
here
we
go.
So
let
me
link
that.
A
In
the
meeting
minutes
as
well
so
lumino
2
vibration,
guide,
partial
and
I
guess
I
can
go
through
some
of
what's
different
because
it's
it's
a
fairly
big
PR,
but
actually
a
lot
of
it
is
the
same
Concepts
over
and
over.
A
So
there
are
two
pretty
commonly
used
functions
in
lumino
that
are
both
deprecated,
although
one
of
them
is
deprecated
only
in
a
PR
one
function
is
two
array
which
we
use
in
a
lot
of
places
and
the
other
function
is
each
which
helps
you
Loop
over
an
iterable
two
array
is
deprecated,
because
a
native
function
exists
array
Dot
from
that,
can
take
any
integerable
and
turn
it
into
an
array.
So
we
strictly
speaking,
don't
need
it,
but
I
would
also.
A
If
you
look
at
the
migration
guide,
I
would
also
recommend
not
really
using
a
radar
from
unless
you
truly
need
it.
There
are
some
places
where
you
actually
need
to
exhaust
the
contents
of
an
iterator,
because
you're
serializing
it
or
you're
converting
it
to
Json
or
you
need
some
array
method
or
you
know,
there's
lots
of
reasons,
but
there's
plenty
other
reasons,
plenty
other
cases
rather
where
you
can
just
iterate
through
the
object
itself,
without
converting
it
to
an
array.
A
So
if
you
look
both
at
lumino
and
at
the
lab,
PR
you'll
see
most
places
where
two
array
existed,
haven't
been
replaced
with
Arena
from
they've
just
been
removed.
The
other
one
is
each
which
lots
of
helper
libraries
prototype
low,
Dash
jQuery.
They
have
something
like
this,
but
now
that
iterators
and
iterables
are
native
and
there's
a
native
four
of
loop,
where
you
can
say
for
X
of
some
iterable
whether
that
iterable
is
an
array,
a
map,
a
set
or
some
custom
thing
that
you
created,
because
that
for
Loop
is
now
native.
A
It
just
really
takes
most
of
the
use
cases
away.
You
rarely
need
to
use
each
to
Loop
through
something
and
by
rarely
I
mean
pretty
much.
Never
because
it
won't
make
your
code
shorter
really
than
the
same.
For
Loop
arrays
already
have
a
for
each
built
in
and
yeah,
so
I
mean
anyway,
take
take
a
look
if
you're
curious
or
if
you
want
to
do
a
review
of
this
or,
if
you're,
just
looking
to
see
how
your
own
extension
is
going
to
need
to
change
to
be
compatible
with
luminode
too.
A
It
shouldn't
be
a
huge
change
for
you,
like
all
of
Jupiter
lab
with.
However
many
packages
it
has
and
and
the
the
pr
is
on
the.
What
is
this,
the
pr
is
about
on
the
order
of
like
12
1300
changes.
So
that's
pretty
small
on
the
grand
scheme
of
things
and
many
of
those
are
just
package.
Json
updates
the
other
big
change
in
luminal
2.
A
Is
that
or
one
other
big
change
in
Lumina
too,
is
that
the
data
store
package
is
gone,
but
I,
don't
think
that
really
affects
anyone,
people
who
aren't
using
it.
That
was
our
crdt
implementation,
and
now
that
Jupiter's
RTC
story
is
based
on
yjs
we're
not
using
that
anymore.
A
There's
a
lot
of
other
smaller
changes.
You
can
look
at
I've,
also
linked
to
the
lumino2
project.
You
can
look
at
what
issues
have
been
completed
and
what
issues
are
still
outstanding,
thanks
to
Fred
and
piyush
and
Gabriel
among
others,
who
are
working
on
this
yeah
I
guess
that's
all
I
really
have
to
say
about
that.
But
if
you
have
any
questions,
I'm
happy
to
try
to
answer
them.
D
A
Good
thanks
right
so
then
you're
up.
D
Yeah
I
just
added
a
small
item.
So,
as
you
know
there,
there
has
been
a
prerelease
of
lab
4
with
code
mirror
6.,
and
there
are
quite
a
few
extensions
that
will
probably
require
some
significant
changes
to
work
with
jupitab
4
and
code
number,
six,
one
of
which
is
LSP.
As
we
all
know,
I
mean
we've
discussed
this
before,
but
the
other
one
that
we
had
not
thought
of
earlier
is
Jupiter
lab
git,
because
there
is
currently
no
merge
view
plugin
for
code
mirror
6
that
will
have
to
make
one
yeah.
D
We
have
similar
issues
for
features
that
we
actually
dropped
when
we
upgraded
to
code
mirror
6.
For
example,
we
don't
have
latex
syntax
highlighting
so
if
people
feel
like
making
a
some
kind
of
latex
passing
for
Jupiter
lab
or
code
number
six,
they
should
you
know
manifest
themselves
and
because
I
think
there
is
significant
work
there.
So
yeah,
that's
it.
A
I
actually
did
think
of
one
other
thing:
I
brought
it
up
last
week
as
well,
bring
it
up
again,
and
that
is
that,
if
you
are
on
the
Jupiter
lab
Council-
and
you
would
like
to
nominate
someone
else
for
the
council,
please
do
if
you
would
like
to
nominate
someone
to
be
our
software
steering
Council
representative
for
this
project.
Please
do
we
are
trying
to
ramp
up
the
new
governance
quickly.
A
We
have
a
an
intermediate
deadline
in
October
for
the
sub
projects
to
all
have
reasonably
representative
councils.
I
think
Jupiter
lab
is
pretty
close,
I
mean
I.
It
was
just
counting
how
many
people
on
this
call
are
in
the
council
and
how
many
people
are
on
this
call,
and
it's
it's
it's
not.
A
It's
not
I,
think
complete.
Yet
we
should
probably
have
some
more
but
we're
pretty
close.
As
far
as
the
software
steering
Council
goes,
that
body
doesn't
exist
yet,
but
what
it
will
be
is
a
regular
meeting
with
representatives
from
every
Jupiter
sub
project
and
the
sum
of
the
working
groups
and
standing
committees
where
cross
project
things.
A
That
have
implications
on
someone
else's
project
are
discussed.
Sometimes
they
might
need
to
be
voted
on.
Usually
it's
probably
just
discussion
and
jeps
are
voted
on
by
this
body.
That's
Jupiter
enhancement
proposals
and
you
don't
need
any
additional
qualification
to
be
a
representative
for
a
project.
Then
you
do
just
to
be
a
voting
member
of
that
project.
On
that
project's
Council,
it's
not
meant
for
like
that
project's
mini
pdfl
or
its
technical
leader,
or
anything
like
that.
A
In
fact,
many
projects
don't
have
any
such
person,
and
so
the
idea
isn't
that
this
is
the
person
in
charge
of
the
sub
project.
It's
that
this
person
represents
the
subproject
and
knows
it
reasonably
well,
so
that
they
can
have
discussions
with
other
Jupiter
sub
projects
about
the
areas
where
they
overlap.
A
Json
grout
pasted
in
the
chat
the
section
of
governance
that
we
worked
on
last
week
that
describes
the
software
steering
Council
representatives
and
says
roughly
what
I
just
said,
but
probably
in
better
language,
because
we
thought
about
it
longer
in
that
call
yeah
any
questions
about
that.
E
I
have
one
question:
well,
yeah,
I!
Guess
that's
a
question
for
the
group
here
or
really
a
question
for
the
council.
Maybe
the
nomination
process
for
the
SSC
representative,
I
think
I
would
prefer
for
us
to
have
an
open
call
for
nominations
and
a
private
submission
process
so
that
if
someone
nominates
somebody
it
doesn't
intimidate
other
people
from
nominating
a
different
person.
Does
that
make
sense-
or
at
least
at
least
that
there's
a
deadline
by
which
people
are
nominated?
E
So
it's
not
like
the
first
person
nominated
is
okay,
let's
vote,
okay,
great,
so
that
we
can
have
some
choice.
A
So
that's
a
good
idea.
The
deadline
seems
really
easy,
because
no
one
has
to
do
it.
The
the
nomination
not
being
open
means
someone's
got
to
receive
those
nominations
right.
A
D
E
Mean
what
we
can
do
is
just
set
up
like
a
Google
form
and
then
on
the
deadline
date.
We
just
open
up
the
Google
form,
though
I
agree
that
it
makes
sense
for
someone
to
sort
of
manage
that
process,
but
it
can
be
a
lightweight
process.
Is
what
I'm
saying
I
just
oh,
create
a
Google
form
on
the
deadline
date
copy
those
things
in
an
email
to
the
council,
saying:
okay,
here's
our
nominations,
sure.
A
Yeah,
please
thank
you,
those
both
sound
like
good
ideas.
What's
a
good
time
frame,
do
you
think
when
should
that
deadline
be.
A
D
A
Right
so
why
don't
we
say?
Give
people
so
votes
only
take
a
week
right.
So
if
we
gave
people
a
whole
month
from
today,
that
still
gives
us
a
week
to
do
the
vote
and
that's
quite
a
long
time
to
do
nominations.
We
might
even
decide
that's
too
long.
Maybe
we
say
September
15th
or
something.
If
you
want
to
nominate
someone
for
the
software
steering
Council
to
represent
Jupiter
lab.
Do
it
by
September
15th
is.
E
That
okay
I
think
it's
fine
to
make
it
a
month
from
today,
because
remember,
the
council
might
be
growing
and
the
council
might
include
people
that
are
nominated
or
want
to
make
a
nomination.
So
I
think
it's
fine
to
say
a
week
before
our
deadline,
there's
a
deadline
and
can
we
hold
the
vote
and
then
October
3rd?
We
have
our
representative.
F
D
D
A
Have
two
items
in
additional
discussion:
I,
don't
think
they're
actually
meant
for
discussion,
but
we
may
as
well
go
over
them
who
added
the
iPod,
widgets
ones.
E
E
We
are
releasing
some
patch
releases
that
are
addressing
issues
that
people
are
are
coming
up,
but
one
of
the
first
things
that
one
of
the
first
patch
releases
we
did
for
iPad
widget
seven
is
we
realized
that
iPad
widget
7
did
not
appropriately
cap
the
dependency
on
Jupiter
lab
widgets.
E
So
if
you
install
iPad
widget
7
now
well,
if
you
installed
it
before
before
we
released
the
patch
release,
is
7.7.2
or
7.6.6,
then
you
would
end
up
with
IPI
widget
7
python
package,
but
Jupiter
lab
widgets,
eight
Jupiter
extension
package
and
those
weren't
compatible,
and
so
there
were
some
problems
with
that.
So
just
FYI,
if
you're
using
appetite,
widget,
7
and
all
of
a
sudden
things
broke
since
last
Thursday.
E
It
may
be
that
your
Jupiter
lab
underscore
widgets
package
is
the
iPad
widgets
eight
compatible
version,
so
you
can
either
pin
your
version
of
Jupiter
underscore
widgets
or
upgrade
to
iPad
widget,
7.7.2
or
7.7
6.6,
which
fixes
dependency
issue
so
a
bit
subtle.
But
if
you
notice
problems
since
last
Thursday
with
ipy
Widget
seven,
you
might
check
that.
F
D
Yeah
I
added
I,
don't
think
the
there
is
no
need
for
discussion,
but
just
to
notify
that
and
big
Raider
had
been
has
been
released
this
week
and
and
now
we
are
still
working
on
the
compatibility
with
notebooks
available.
D
Yeah
yeah
it's
right,
but
notebook
7
and
Jupiter
app
for
normally
soon.
A
F
One
of
the
things
I'd
love
to
see
is
some
explicit
consideration
for
accessibility
and
I
wondered
if
stuff
like
res
standards
or
Aria
specs
might
be
in
scope
for
doing
enriching
some
of
the
rendering
and
some
of
the
descriptors
there
I
don't
know
if
there
are
hooks
like
that
yet,
but
with
all
the
work
going
on
with
accessibility,
that
Frederick's
been
talking
about
and
work
at
the
Jupiter
lab,
retro
lab
level
I'm
curious.
If
lumino
has
the
glue
that
we
might
need
to
actually
unify
some
of
these
efforts.
A
Yeah,
so
this
is
something
we've
been
thinking
about
and
there's
a
few
different
answers
to
it,
but
before
I
fill
up
more
time
with
my
voice,
I
wonder
if
Gabriel
wants
to
say
anything
about
it.
D
So
good,
not
great
timing,
because
I'm
I'm,
currently
in
transit,
but
yeah
I,
there's
I,
don't
really
know
what
to
say
about
it
other
than
like
we,
we
there's
an
issue,
that's
currently
open
and
being
tracked
in
luno
to
specifically
about
accessibility
and
we've
had
conversations
about
it
and
we're
looking
into
the
code
base
to
see.
D
If
there's
you
know,
opportunities
with
this
upgrade
to
bring
anything
like
to
to
put
in
to
actually
sort
of
like
put
in
place
anything
that
would
be
useful
accessibility,
wise,
but
that's
all
I
can
that's
all
I
can
think
of.
A
Okay,
I'll
add
on
to
that
then.
So
today
we
merged
a
PR
that
Gabriel
had
opened.
A
That
fixes
the
fact
that
the
luminal
menus
will
capture
your
keyboard,
focus
and
tab
and
not
let
go
and
that's
been
resolved,
and
we
had
just
the
look
at
all
the
different
packages,
they're
sort
of
their
API
footprint
and
how
you
would
approach,
making
these
components
more
accessible
and
what
the
proximate
thing
we
wanted
to
resolve
was:
are
there
going
to
be
backward
incompatible,
API
changes
that
will
help
us,
because
now
is
the
time
for
that
and
it's
a
short
window.
A
We
have
for
that
and
then
the
next
question
is,
irrespective
of
whether
the
backward
incompatible
or
not,
what
sorts
of
things
do
we
need
to
offer.
So
much
of
lumino
can
be
augmented
to
be
more
accessible
in
user
land
right
because
it
uses
things
like
renderers,
and
you
could
just
give
a
better
renderer
that
has
these
attributes.
Some
of
it
probably
will
benefit
from
receiving
instantiation
options.
That
applies
things
for
you,
instead
of
you
having
to
do
it
after
the
fact,
and
so,
but
even
even
those
things
can
be
added
are
additive
right.
A
They
won't
necessarily
break
backward
compatibility.
So
some
of
the
pressure
to
get
a
whole
bunch
of
fixes
in
right
away
is
a
little
less
than
I
initially
feared.
However,
it's
definitely
a
really
good
time
to
be
looking
at
these
components
because
they're,
the
building
blocks
of
lab
and
notebook
and
so
improving
them
out
of
the
box
just
helps
everyone
I
think.
One
problem
we
have
is
that
so
few
people
in
the
project
work
on
lumino.
A
They
work
on
lots
of
other
aspects,
so
part
of
what
I
personally
am
trying
to
do
is
to
get
more
people
engaged
and
to
try
to
tell
them
what
I
know
about
it,
so
that
it's
not
just
contained
in
the
heads
of
a
few
people
who
worked
on
it
five
years
ago,
Mike
linked
to
an
issue
that
is
just
part
of
the
luminode
2
project,
which
is
specifically
about
this,
and
it's
a
it's
a
good
sort
of
place
to
not
quite
treat
as
a
wish
list.
D
A
Don't
know:
does
that
sort
of
give
you
a
basis
for
next
steps?
Is
there
like?
Is
there?
Is
there
some
aspect
of
it?
We
could
talk
about
that.
Would
that
would
help
I
I?
Think
one
thing
I
want
to
do
is
be
more
present
at
the
accessibility
calls,
because
not
that
many
people,
like
I,
said
no
luminos,
so
showing
up
there
and
trying
to
solicit
help
and
offering
my
own
knowledge
of
the
packages
might
might
be
beneficial
there.
F
I
mean
if
you're
around
today,
maybe
and
there's
time.
Maybe
we
can
continue
some
of
that
and
Gabriel.
If
you
have
some
time,
maybe
we
can
catch
up
about
it,
I
mean
I,
don't
know
lumino
all
that
well,
but
I
think
I've
seen
some
of
the
rubs
with
like
there's
some
accessibility,
things
that
are
like
internal
to
the
component,
and
then
there
are
some
accessibility,
things
that
are
application.
F
Specific
right
like
doing
Aria
labels
of,
say
the
notebook
in
different
applications
like
The
Notebook
widget
in
different
applications
might
need
to
change.
So
maybe
we
can
just
like
talk
a
little
bit
around
like
what
are
some
of
the
things
that
are
constant
and
some
of
the
things
that
might
need
to
change
and
it
might
help
find
a
common
ground,
and
thanks
for
this
issue
too,
after
that
discussion,
I'll
be
able
to
put
together
some
words
a
little
more.
A
Cool
great
okay,
then,
why
don't
we
go
to
demo
time.
D
B
But
there
is
no
more
point
to
discuss
so
at
first
it
doesn't
change
much.
So
this
is
the
new
panel,
the
the
main
difference
they
say
for
the
look
and
field
is
that
is
using
the
side
panel
with
a
caution.
That's
the
same
that
the
the
debugger
for
a
bit
more
homogeneous,
looking
field
and
the
back
end
is
is
done
with
Pi
Pi
so
and
everything
is
done
through
the
back
end.
So
if
I
look
for
something
here
and
I
should
have
open
network,
that's
my
fault
hope.
B
And
I
should
look
to
all
requests
and
not
okay,
so
everything
now
is
done
by
asking
the
the
backend.
So
we
are
it's,
maybe
a
bit
small.
So
we
are
asking
the
endpoint
API
with
the
query
and
the
server
is
responding
with.
What's
the
list
of
things,
of
course,
here
I
put
something
that
was
with
one
output.
If
I
put
more
I
will
get
more
output.
B
The
the
reason
for
doing
that
is
to
to
be
able
to
to
have
an
API,
that's
pluggable,
meaning
that
you
can
change
the
back
end
or
the
manager
on
the
back
end
by
a
thread
play,
and
you
can
Define
and
hook
your
own
by
just
defining
the
API
in
a
python
package
and
defining
an
entry
points
for
the
your
custom
extension
manager
to
be
available.
So,
for
example,
we
can
imagine
to
do
one
for
using
conda.
B
The
main
trouble
I
have
to
say
is
to
be
able
to
to
filter
the
the
list
of
available
package
to
be
meaningful.
So
for
now,
when
we
are
using
Pi
Pi,
what
we
are
tracking
is
all
the
extension
that
are
that
have
the
classifier
Jupiter
lab
and
pre-build.
B
So,
let's
bring
a
couple
of
limitations.
One
of
them
is
that
the
extension
itself
can
be
incompatible
with
the
current
version
of
Jupiter
lab.
So
we'll
see
how
to
deal
with
that
maybe
later.
Another
trouble
also
is
to
reconciliate
the
name
of
the
the
extension
in
Python
and
the
one
in
and
the
one
in
npm,
so
I
don't
have
it
installed,
but,
for
example,
NB
time
if
you
get
it
installed
for
its
name,
I
think
is
Jupiter
lab
nbdime
and
it's
missing
the
installed
Json
file.
B
So
you
cannot
figure
out
what's
the
exact
name
of
the
Python
package
to
uninstalled,
if
you
want
to
uninstall
it
from
the
from
the
interface,
so
those
those
are
the
kind
of
limitations
we
have
for
now,
so
that
was
the
Pi
Pi
version
and
if
I
just
restart
the
server
and
changing
the
out
the
the
manager
to
be
the
read-only
one.
B
What
will
happen
is
that
you
will
be
able
to
just
see
what
are
the
extension
installed
and
you
will
be
able
to
enable
them
or
disable
them,
but
you
won't
be
able
to
to
install
anything
or
search
some
distant
package
repository,
but
at
least
you
will
be
able
to
just
say:
Okay
I
want
to
disable
like
Corners
by
and
then
you
are
informed
that
you
need
to
refresh
that's
probably
a
follow-up
for
a
target
for
Jupiter
up
5
to
be
able
to
reload
some
plugins.
That
will
be
nice.
B
And
this
is
like
so
most
of
the
thing
is
not
in
the
the
UI
it's
more
in
the
in
the
API.
That
has
completely
changed
so
because
also
there
are
there
are
that
features
that
you
can
whitelist
or
Blacklist
extensions,
so
that
feature
is
still
available,
but
everything
is
done
in
the
back
end
before
replying
to
the
front
end
when
you
are
asking
for
a
list
of
extension,
and
so
those
features
are
still
available,
they
are
just
opening
in
another
place.
A
Yeah
there's
a
question
that,
from
like
saying,
should
we
show
the
package
manager,
name
and
installation
path?
Slash
environment
in
the
UI
users
might
get
a
breaking
environment
change
after
installation.
So
we
should
be
explicit
about
what
action
is
being
taken.
C
C
B
D
So
maybe
some
just
a
metadata
request
about
the
installation
method
or
or
for
people
which
are
the
six
prefix
and
cpip
or
and
for
npm
you
would
show
like
the
place
for,
or
you
know,
npm
packages
that
will
be
installed.
As
extension
Sim,
you
know
just
maybe
just
a
markdown
type.
You
know
like
something
you
can
display.
E
Nominally
the
install
that
Json
tries
to
record
the
information
of
where
that
extension
came
from
and
the
idea
was
you
don't
have
to
use
pip
to
install
it.
You
could
just
copy
files
there
from
wherever
you
find
them
on
the
Internet,
and
so,
if
we
had
some
place
to
get
the
files
say
a
CDN
that
was
showing
npm
packages
or
whatever,
then
your
extension
could
just
copy
those
files
into
the
right
directory
and
put
and
install.json
there.
E
That
indicated,
this
was
installed
by
the
trooper
lab
extension
manager
so
essentially
use
the
jupitab
extension
manager
to
uninstall
it.
If
you'd
like.
A
Thanks
for
demo,
that's
cool
one
thing
that
perennially
comes
up,
that
we
don't
have
a
good
answer
for,
but
it's
worth
pitching
to
16
people
is
ratings
or
reputation,
or
you
know
like
when,
when
you're
in
a
different
IDE,
whose
name
I
won't
say,
you'll
see
that
this
thing
has
X
number
of
stars
and
it
would
be
beneficial
to
have
the
the
the
sort
of
reassurance
that
oh
this
this
this
is
super
popular,
so
I'll
use
it
too
I
wonder
like
we're
talking
about
something
fairly
lightweight,
but
something
that
we'd
have
to
keep
track
of
and
we'd
have
to
host.
C
Yes,
today
on
topic:
first,
the
there
is
a
lot
of
work,
maintaining
those
kind
of
ranking
systems
and
having
hosting
for
it-
and
you
know
the
moment-
you
have
a
thing
that
calls
out
when
you
use
it,
I
think
the
previous
extension
manager
right.
It
had
something
like
that.
Yeah.
If
you
use
this
it
will,
you
know,
hit
it
told
you
what
you
would
do
and.
C
By
default
right,
so
if
something
is
enabled
by
default
and
it
hits
and
a
public
internet
endpoint,
that's
not
a
code
right,
so
you
want
something
had
to
be
something
you
opt
into,
but
yeah
toasting
and
maintenance
and
making
sure
that
nobody
is
trying
to
game
it
and
all
this
stuff.
C
It's
if
it
couldn't
be
crowdsources
crowdsourced
good,
if
it's,
if
not
and
if
it's
It's
You
Know
It's
Tricky.
C
Also,
just
a
side
point
for
the
install.json,
but
the
experience
for
the
previous
thing
was
that
if
you
need
metadata
for
the
extension
manager
to
work
once
the
extension
manager
is
out,
people
will
start
clustering.
Extension
authors
to
add
this
thing
so
that
that
will
come
surprisingly
quickly
after
it's
been
released.
So
it's.
D
F
E
B
The
archival
services
like
this
one
that
provides
a
simplified
API
for
the
Google
bigquery
table
on
Pi
Pi,
but
most
of
them
they
are
doing
throttling
for
a
good
reason,
and
so
definitely
what
Vidor
was
saying
is
the
is
true
is
if
we
really
want
to
have
that
kind
of
download
stat
like
on
Pi
Pi
for
the
Pi
Pi
manager,
we
would
probably
need
to
to
kind
of
having
a
repository
with
bot
that
are
just
like
creating
those
ranking
that
can
be
requested
more
heavily
than
those
Services,
because
those
ones
are
not
done
for
that.
A
I
mean
at
the
end
of
the
day
we're
talking
about
something
that
a
Raspberry
Pi
could
serve
right
because,
let's
say
there's
2
000
extensions
right,
put
them
in
a
sqlite
database
hammer
it
as
hard
as
you
want,
and
you
won't
take
down
a
Raspberry
Pi
so
like
it
isn't.
Actually
it's
more
the
persistence
and
the
hosting
and
all
that
other
stuff.
That
I
think
is
hard,
not
the
actual
amount
of
data
we're
going
to
be
sending
this
API
Jason.
What
do
you
think.
E
With
the
previous
extension
manager,
when
we
confronted
this
curation
and
rating
and
store,
you
know,
App
Store
kind
of
question,
our
first
cut
was
having
two
classes
of
extensions.
One
that
were
one
was
extensions
that
are
published
by
the
project
Jupiter,
and
so
there
was
this.
You
know
icon
of
little
Jupiter
off
to
the
corner,
and
that
was
an
easy
way
to
give
people
some
level
of
Judgment
at
least,
but
we
stood
behind
in
Project
Jupiter,
even
without
trying
to
solve
the
general.
E
Like
rate
everybody's
code
in
the
world
kind
of
problem,.
D
E
I
think
we
that
was
an
initial
rough
cut
but
I
think
there
was
a
hard-coded
list
because
there
may
be
some
stuff.
That's
not
in
the
org
like.
B
The
thing
is
in
the
API
of
the
reworked:
PRI
have
opened
I
kept.
This
thing
like
you,
can
add
the
metadata
to
the
packaging
it's
approved,
but
the
question
is
from
type
I
I
didn't
figure
out
what
could
be
a
good
way
of
saying?
Okay.
This
is
an
extension
that
by
default,
we
really
support
because
you
cannot
say
like
if
it
start
by
Jupiter
lab
it's
too
easy
to
hack.
B
B
C
B
B
C
If
it
starts
in
read-only
mode
and
then
the
warning
is
for
removing
read
only
mode,
then
at
least
the
base
level
is
more
informative
and
the
warning
is
for,
when
you
actually
take
saying
like.
Oh
I,
am
actually
going
to
change
stuff
around,
and
this
is
when
you
have
the
risk
of
executing
code.
Yeah.