►
From YouTube: JupyterLab Team Meeting - 26 April 2023
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/170
A
A
If
you
have
something
you'd
like
to
talk
about,
please
add
it
to
the
agenda,
which
is
linked
in
the
chat
and,
if
you
don't
please
sign
in
and
if
you
don't
want
to
that's
okay
too,
but
with
all
that
being
said,
why
don't
we
get
started?
The
agenda
was
empty,
so
I
put
myself
first,
but
this
isn't
really
my
subject.
There
are
three
outstanding
issues
that
are
released
blockers
and
each
one
of
them
has
a
corresponding
PR.
A
So
I
guess
what
we
can
do
is
we
can
look
at
those
three
individually
and
if
the
people
who
authored
either
the
PRS
or
the
issues
are
on
the
call,
maybe
they
could
take
it
up.
So,
let's
open
this
up
and
see
so
the
first
one
is
about
search
and
replace
with
the
string
undefined
it
was
opened
up
by
Mike
and.
A
I
I
I'm,
actually
just
saying
the
words
and
the
pr
is
authored
by
Andrew
who
just
spoke
up,
so
why
don't
you
tell
us
what
the
current
state
is?
Please.
B
So
PR
is
ready
for
review
and
merge
as
far
as
I
can
say,
but
it
was
not
reviewed
yet
so
it's
in
the
agenda
under
entry
and
here
is
a
link
in
the
chat.
So
it's
it's
an
online
change,
so
this
power
again
would
solve
new
blocker
issue
created
by
Mike
yeah.
A
Because
that
one
line
seems
like
a
sound
line,
so
if
somebody
else
also
gives
it
a
Sandy
check
and
is
happy
to
merge,
it
I
think
that
that's
good,
because
all
that
was
happening
was
there
was
type
coercion
going
on
and
JavaScript
undefined
was
being
rendered
as
a
string
undefined,
so
initializing
it
to
empty
string,
seems
open
and
shut,
but
just
in
case
I'm
missing
something
a
second
pair
of
eyes,
maybe
but
otherwise
cool.
Thank
you.
A
Okay,
next
is
a
PR
opened
by
Brad,
and
this
is
for
scrolling
for
insert
cell
run
in
advance,
okay
cool.
It
looks
like
Mike
you've
taken
a
look
at
it.
There's
been
a
lot
of
back
and
forth.
Jason
also
has
looked
at
it.
Does
anyone
have
an
update
on
this?
Is
it
stuck
on
something.
B
So
there
was
a
visual
UI
regression
fail
like
the
similar
that
we
were
getting
on
the
Mars's
PR
and
the
one
I
was
getting
on
the
like
line
buttons
right
for
a
news
message,
so
so
in
terms
of
functionality.
It's
it's
a
separate
question
because
we
haven't
reached
part
here
with
the
3.7
or
3.6
rather
but
yeah
we
could
merge
this
version.
I
think
it's
just
that
the
visual
UI
regression
tests
are
failing.
A
Okay
and
it's
fully
caught
up
with
the
base
branch,
is
that
right,
so
there's
nothing
like
the
the
snapshots
will
automatically
fix
themselves.
Is
it
definitely
caused
by
this
PR.
B
So
Mike,
maybe
you
can
help
me
here,
expand
more
on
this.
So
it
was
a
situation.
But
like
last
night,
when
I
went
to
sleep
was
that
we
had
two
flaky
snapshots.
One
of
them
was
like
associated
with
a
new
cell
button
appearing
on
top
of
the
cell,
and
there
was
second
one
was
a
debugger
window,
I
think
with
variables
not
showing
and
Mike
indicated
those
to
me
and
I
have
like
rolled
them
back
to
the
previous
version
in
three
PRS.
B
So
two
of
them
like
this
critical
PRS
blocking
Powers,
the
Fredericks
one
and
Martin
Luther
one
and
the
third
is
my
other
player.
Let's
not
talk
about
it
and
after
that
the
tests
ran
again
and
the
server
still
fails.
So
I
just
looked
at
that
in
the
middle
of
the
night,
but
haven't
investigated
that
much
yet.
A
Yeah,
okay
Okay,
so
there
is
a
behavior
that
was
incorrect
and
it's
resolved
here,
but
the
behavioral
change
causes
the
visual
regression
test
to
fail.
I
think
we
can
create
a
secondary
issue
that
says,
fix
the
visual
regression,
but
the
release
maybe
doesn't
need
to
be
blocked
on
that,
because
the.
D
C
There
is
nothing
in
this
PR
which
causes
a
visual
regression.
There
is
just
that
one
of
the
pull
requests
which
added
click
to
ads
so
has
introduced
a
very
flaky
snapshots,
actually
two
snapshots
and
these
snapshots
are
causing
a
lot
of
noise.
When
we
update
snapshots
using
the
ports
to
read,
generates
snapshots
on
CI,
it
is
introducing
this
flaky
yeah,
problematic
structures
that
we
need
to
manually
invert
every
time
that
we
update
Solutions
now.
C
Yeah
as
much
as
I'm,
not
convinced
about
the
snapshots
that
change,
because
the
behavior
has
changed,
I
just
wanted
to
take
another
look
at
this,
as
if
anyone
else
could
also
try
it
to
see
how
the
new
Behavior,
because
this
request
plan
fixes
the
bug.
Second
changes:
the
behavior,
how
a
switching
to
active
cell
behaves,
so
it
adds
a
small
bug
in
it
at
the
bottom
and
I
think
this
is
probably
by
but
I
think.
It
also
benefits
someone
else
to
take
a
look
beside
me.
I'm
fine,
with
this
PR
over
this.
A
So
I
think
it's
reasonable
to
visit
the
question
of
what
we
want
the
behavior
to
be
I.
Just
don't
think
we
need
to
block
a
release
on
it.
So
maybe
can
someone
volunteer
to
write
an
issue
on
that
aspect
of
it
so
that
we
can
merge
this
PR
and
move
ahead
to
an
RC.
A
A
Sorry
I
opened
up
the
wrong
link,
which
is
about
additional
properties
in
the
settings,
editor
generating
a
validation
error
and
is
Martha
Nicole.
Is
there
any
okay
cool
hi.
E
Yeah,
so
it
the
problem
is
fixed
and
I.
Think
the
feedback
I
got
was
to
add
a
test
and
I
think
Fred
actually
added
one
for
me.
So
I
think
it's
ready
other
than
it
looks
like
some
tests
are
failing
in
the
CI
I'm,
not
sure
it
looks
like
Andrew
was
looking
into
that
I,
don't
know
if
you
have
anything
to
say
entry.
A
B
Is
a
same
issue
on
two
blocking
PRS
on
Frederick's
PR
and
UI
regression
test
yeah
fail.
A
Incidentally,
this
vs
codes,
Json
launch
file,
allow
comments
they're
using
a
version
Json
that
allows
line
comments.
I
guess
that's
neat.
I
didn't
know
that.
A
B
So
I
wanted
to
clarify
also
on
4.0.2
release
plan,
but
I
guess
we're.
You
know,
fixing
blocking
issues
as
soon
as
possible
and
then
doing
an
RC
and
then
waiting
two
weeks
and
then
doing
4.0.0
release.
Is
that
what
hand.
A
Will
probably
need
more
than
one
RC,
because
we
already
know
we
had
outstanding
issues
that
need
to
be
solved.
They
just
weren't
critical
release
blockers,
so
there'll
probably
be
an
RC
base
that
only
lasts
a
couple
weeks,
but
there
might
be
more
than
one
RC
within
that
phase.
I
think,
but
I'm
not
sure.
What's
what
is.
A
A
B
Oh
yeah,
so
do
is
RC
period
two
weeks.
My
first
question
before
is
the
four
point
of
pointer
release
and
the
second
question
I
want
to
double
check
that
we
can
still
merge
patches
into
the
RC
before
the
4.0.0
release.
A
Okay
question:
two:
the
answer
is
unambiguous.
Yes,
absolutely
question
one:
here's
how
I
would
frame
it
do
we
need
two
weeks
following
the
last
RC
before
the
final,
or
can
we
have
multiple
RC's
in
a
two-week
period
and
as
long
as
they're,
all
small
incremental
and
nothing
that
is
surprising.
We
stay
within
that
two-week
time
frame.
B
Can
I
ask
another
question
what
makes
an
like,
for
example,
rc1
versus
rc2
versus
rc1,
with
just
like
patches
merged
in
or
any
patch
makes
it
like,
rc2,
rc3
and.
G
B
A
A
H
Hi
nice
to
meet
you
I,
just
have
some
questions
about
Jupiter
light,
since
I
recently
started
to
incorporate
Jupiter
light
into
my
website.
H
A
A
D
By
separating
what
do
you
mean
exactly,
you
mean.
H
So
what
I
want
to
do
is
so
right
now
I'm,
showing
the
entire
Jupiter
light
website,
I'm
redirecting
to
Jupiter
light.
That's
hosted
on
my
GitHub
public
repo,
so
I
was
wondering
if
I
can
just
incorporate
the
jupiter-like
components
into
smoke
as
a
small
component
into
my
website.
H
I,
don't
know
if
that
made
sense
like
instead
of
redirecting
it
I
want
to
make
it
incorporate
to
my
to
high
level
web
component.
So
I
was
wondering
if
that
separation
was
possible.
A
So,
for
example,
there
are
projects
you
can
use
online
documentation,
embeds
a
Jupiter
code,
console
that
comes
from
Jupiter
like
but
isn't
like
the
full
application.
Is
that
what
you
mean.
H
Oh
I
think
you're
referring
to
Rebel
r-e-p-l,
so
I
actually
want
to
use
the
Jupiter
life
front
end.
H
F
It's
in
the
main
thread
of
of
the
UI
and
then
the
client
connects
to
that
and
overloads
its
service
worker
or
a
service
manager
which
handles
Fetch
and
websock,
and
all
that
stuff
yeah
there's
a
link
there,
it's
kind
of
out
of
date
the
kernels
themselves
actually
run
inside
of
web
workers
and
they
only
communicate
over
post
message,
there's
no
dirty
and
whatever
they'll.
You
know
that
could
be
on
another
domain
if
it
was
set
up
properly.
F
F
I
I
I,
don't
I,
don't
understand
what
you're
trying
to
remove
from
the
Jupiter
level
UI
exactly.
H
That's
right
so
I'm
just
trying
to
make
it
lightweight,
because
I
want
to
increase
the
speed
of
running
each
cell
and
I
saw
that
some
things
that
are
involved
I
try
to
remove
from
requirements.txt,
but
they
it
was
still
a
little
bit
slow
through,
like
that
and
all
I
want
and
also
I
I
want
to
have
some
control
over
what
kind
of
libraries
I
can
already
have
it
downloaded
and
that's
why
I
wanted
to
have
control
over.
I
So
I
think
maybe
then
the
answer
would
be
to
not
use
the
biodic
column
admit
like
Jeremy,
can
talk
to
that
and
use
this
other
kernel
that
we
have
based
on
juice
python,
which
allows
you
to
make
a
smaller
bundle
of
packages
that
you
host
then
So.
I
Currently,
if
you
use
the
buy
that
kernels,
everything
is
being
downloaded
from
a
CDN
and
if
you
use
shoes,
python
kernels,
you
would
deploy
your
own
environment
as
well
on
your
website,
including
all
the
python
packages,
but
it
may
be
smaller
if
you
pick
a
smaller
list,
so
Jeremy
can
maybe
explain.
I,
don't
know
whether,
where
this
will
be
documented.
D
Yeah,
it's
in
in
the
how-to
guides
and
I
also
just
posted
a
link
to
another
project
that
is
reusing
parts
of
the
HP
light
packages,
but
I
think
in
that
case
it's
a
bit
different.
They
only
used
server
part
by
server
I
mean
the
piece
of
software
that
runs
in
a
browser
to
mimic
the
regular
server
and
the
Chronos.
D
They
are
not
using
the
the
bright
front
end,
which
means
you
know,
Jupiter
lab
and
the
green
notebook
interfaces,
and
those
could
be
also
interesting
to
to
have
a
look
at
the
thing
there.
There
are
some
people
who
tried
also
to
have
a
web
component
for
crypto
Lite,
but
I
think
they
were
just
using
a
notebook
directly
I'm,
not
sure.
D
What's
this
what's
the
status
of
the
project,
but
you
can
probably
look
it
up
quickly
and
maybe
that
will
be
more
what
you
will
be
looking
for
like
if
you
want
to
integrate
it
in
your
host
page.
But
there
are
still
a
lot
of
things
that
you're
gonna
hit
I
think
like
loading
times,
I'm,
not
sure.
D
If
there
is
something
we
can
do
at
the
at
that
time,
because
you
still
need
to
download
quite
a
lot
of
things
from
the
internet
to
be
able
to
start
the
Chronos,
so
that
would
still
be
an
issue.
I
think.
H
Jeremy
somebody
asked
me
what
what
I
meant
by
speed
through
the
chat.
What
I
meant
is
that
the
first
cell
always
takes
about
between
10
seconds
to
40
seconds
to
run
and
the
following
cells
becomes
faster
and
faster
and
maybe
already
fast
by
then.
So
that's
what
I
meant
by
speed.
F
F
A
So
yeah
I
mean
I,
guess
I
guess
this
is
going
to
be
a
process
of
trying
out
some
of
the
different
combinations
and
seeing
if
any
of
them
gets
you
closer
to
what
you
want.
If
you
try
some
of
these
combinations
and
find
that
you
have
more
questions
both
here
and
the
Jupiter
I
repo
issues
are
good
places
to
follow
up
yeah,
but
thank
you
for
bringing
it
up
here.
It's
great
because
we
haven't
had
a
Jupiter
light
question
before
so.
That's
nice.
A
Okay,
next
person
on
the
agenda
is
Mike.
C
Books
was
a
little
bit
different,
a
little
bit
broken,
sometimes
quite
a
lot
broken
and
I
think
that
this
PR
finally
polishes
the
remaining
issues
to
level
at
which
we
for
users
would
be
really
happy
with
that
and
I
would
very
much
like
to
see
this
pull
requests
in
next
beta
or
risk
habitats.
Just
so
that
we
can
get
this
out
to
a
larger
user
base
to
test
and
highlight
any
remaining
issues
with
such
action
because
yeah.
That
was
a
feature
that
we
had.
A
Okay,
is
anyone
on
the
call
willing
to
volunteer
to
review
this.
I
A
Cool
and
Mike,
what
I
would
say
is
even
if
we
put
out
rc1
like
today
and
this
isn't
merged
until
Friday,
we
could
just
put
out
rc2
on
Friday.
We
don't
have
to
stall
very
long
between
those
if
we're
proposing
of
two-week
RC
period
in
general,.
B
Yeah
I
agree
with
that.
I
generally
think
we
can
even
create
some
kind
of
tag
to
prioritize
certain
issues
between
our
RC
release
and
4.0.0
release,
like
you
know,
not
maybe
blockers
but
I,
don't
know
first
priority
and
prioritize
those
for
work
and
review.
C
A
All
right,
Nick,
your
Flex.
F
Yo
yo,
it
would
be
cool
to
get
types
in
the
Jupiter
lab.
There's
a
lot
of
code
in
there.
A
lot
of
python
getting
in
there
would
be
very
helpful
I,
don't
think
it
should
block
anything
and
there's
a
few
upstreams.
Now
that
we've
adopted,
that
would
make
things
harder.
You
know,
so
we
can
discuss
that
after
the
fact
or
whatever,
but
there's
some
techniques
that
we
could
use
to
bootstrap.
Some
of
that
stuff.
F
Make
anything
faster
until
you
go
for
my
Pisces,
but
I,
don't
think
we
should
be
doing
that
on
Jupiter
lab
anytime
soon.
So
it's
not
going
to
help
anybody.
F
It's
not
going
to
necessarily
help
anybody
migrate
because
we
haven't
had
a
tremendous
amount
of
API
changes,
but
on
the
python
side,
but
I
feel
like
there's
I,
don't
know,
there's
some
stuff
that
does
make
it
harder
to
to
reason
about
if
your
stuff
that
has
deep
Integrations
is
going
to
work
between
versions
of
those
and
the
fact
we're
split
between
Jupiter
lab
Jupiter
server,
Distributors,
Jupiter,
lab
server
and
now
all
the
ydoc
stuff,
and
almost
none
of
that's
typed,
so
I
don't
know
it's
a
longer
discussion
give
people's
minds
around.
B
So
don't
talk
about
that
so
I
I
personally
want
to
attend
Jupiter,
Khan,
but
I,
think
and
I
know.
Maybe
I
don't
know.
Maybe
something
like
that
is
planned,
but
I
think
it
would
be
nice
to
have
a
4.0.0
RC
backbash
during
Jupiter
Khan
as
a
part
of
jupitercon
Sprints,
for
example,
we
could
concentrate
on
bug,
fixing,
maybe
also
bug
Discovery
but
I-
think
bug
Discovery
would
be
really
nice
and
important.
Oh,
we
had
a
bug.
B
Bash
meetings
within
our
teams
that
were
very
productive
yielded,
like
you
know,
the
thousands
of
bugs
per
session
and
the
ones
that
we
tried
to
organize
with
the
community
online
wasn't
very
well
attended.
So
I
thought
maybe
the
one
in
person
could
be.
You
know
more
popular,
but
there
is
no
question
in
that
bug.
Bashes
are
useful
and
they
help
to
discover
issues
early
and
you
know
discovering
issues.
Early,
helps,
improve
usability
and
not
block
releases.
For
example,.
A
There's
also
the
reality
that
we're
not
out
of
bugs
to
solve
for
Jupiter
lab
four
we've
just
punted
many
of
them
and
said
they're,
not
release
blockers
but
they're
real
bugs
that
need
solving
right
away.
So
I
think
a
bug
bash
that
bashes
bugs
is
better
than
a
bug
bash
that
finds
bugs.
G
B
Okay,
I'll
I'll
add
it
to
like
a
comment
to
that
and
like
bug
fixing
okay,
that
would
be
awesome.
So
if
anyone
wants
to
organize
it,
for
example,
someone
who
gives
a
4.0.0
talk
and
received
an
email
with
invitation
to
initial
like
make
a
Sprint
yeah,
it
would
be
awesome.
J
J
I
came
into
across
an
issue
where
just
the
status
of
the
extension
was
not
okay
and
that
led
me
to
to
realize
that
realizing
that
there
were
more
than
just
the
Jupiter
lab,
like
packages
that
were
being
like
compared
to
like,
at
least
in
from
the
packages
that
are
dependencies
in
my
extension,
were
being
compared
to
the
packages
in
in
Jupiter,
Lab,
Core,
I,
believe
and
I'm
just
wondering
if
it
is
correct
to
to
think
that
in
my,
if
I'm
writing,
my
thinking
that
maybe
all
the
dependencies
that
are
declared
for
an
extension
are
then
expected
to
match
if
they
are
in
Jupiter
Lab
Core.
J
And
just
I
guess
just
for
reference.
I
came
across
this
issue
because
I
was
doing
the
upgrade
extension
script
and
I
was
having
some
trouble
getting.
The
extension
I
was
trying
to
upgrade
to
work
so
I
tried
manually,
updating
that
extension
and
I
was
just
running
into
that
issue,
where
my
I
I
updated
all
the
the
Jupiter
lab
packages,
but
it
seemed
like
react
and
react.
Dom
were
also
supposed
to
be
updated
to
match
a
different
version,
and
presumably
the
version
used
in
Jupiter
lab
all
right.
A
I,
don't
where
do
you
encounter
these
statuses?
Where,
where,
where
do
they.
J
Show
up
sorry:
this
is
the
outcome
of
running
Jupiter
lab
extension
list
and,
seeing
like
is
my
extension
like
enabled,
or
okay
or
the
Red
X
you
get
when
something's
wrong
with
it
got.
A
A
A
Are
sound
dependencies
for
the
version
of
lab
that
you're
running
them
in
so
I?
Don't
know
if
that
answers
your
question,
but
I
think.
Yes,
if
you're
relying
on
a
dependency,
that
is
another
extension
and
it's
another
lab
extension
and
that
extension
itself
would
not
be
okay,
then
your
Ascension
is
not
going
to
be.
Okay.
Is
that
what
you're
asking.
J
I
think
yes,
along
those
lines
I
just
when
I
first
started
looking
into
like
upgrading
an
extension
I
was
thinking
if
the
where
my
focus
was
where
I
was
in
the
Jupiter
lab
packages
that
were
dependencies
in
my
extension
that
those
that
I
was
pulling
from
from
Jupiter
lab,
specifically
so
I
think
it
was
harder
for
me
to
realize.
Oh
there's
something
wrong
with
some
other
dependency
like
reactor
or
something
along
those
lines,
but
I
think
from
my
findings.
J
There's
just
there's
specific
functionality,
that's
called
and
it
just
it.
It
compares
the
dependencies
and
I
was
just
kind
of
I
guess:
I
could
do
a
little
bit
more
testing
just
to
figure
out
if
I
have
another
extension
installed.
If
it
has
some
other
dependency,
if
that
will,
if
that
will
compare
everything,
but
yeah
I
think
maybe
I
could
do
a
little
bit
more
digging
too
in
there.
C
A
Okay,
cool
yeah,
I'm,
sorry
that
I,
don't
have
something
smarter
to
say.
Does.
E
F
You
know
the
sensitive
packages
are
listed
in
in
cupert,
lab
staging
Json,
there's
a
whole
big
section
under
Singletons
and
those
are
the
ones
that
can
cause
you
trouble,
but
there's
other
ones
that
can
also
cause
you
trouble
like
code
mirror,
because
those
are
not
more
of
The
Singletons
for
whatever
reason,
so
it
really
depends
on
the
specific
package.
React
I,
don't
know,
I,
don't
know.
This
is
react.
J
Okay,
thank
you.
I'll.
Definitely
take
a
look
at
that
now.
Thank
you,
I'll
just
throw
out
that
I
have
some
some
documentation,
I'm,
adding
and
if
anybody
would
like
to
take
a
look
and
correct
anything,
I
I,
say
incorrectly
in
the
documentation.
I
would
really
appreciate
it,
but
thank
you
for
your
help.
Today.
G
Okay,
thank
you.
Do
we
want
to
keep
recording,
we
don't
usually
record
the
issue.
Triage
portion,
good.