►
From YouTube: JupyterLab Team Meeting - 12 July 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/205
A
So
if
you
would
like
to
talk
about
something
today's
the
day,
you
can
capitalize
and
if
not,
we'll,
probably
have
a
rather
short
call,
although
I
wonder
if
I
stall
longer,
if
I'll
see
Fred,
adding
things
to
the
things
that
he
wants
to
talk
about,
I
think
that
is
the
case.
Nonetheless,
I
will
pass
it
on
to
Fred,
so
he
can
actually
talk
about.
B
Thanks
Dorian,
it's
just
a
status
update
about
the
cleanup
of
the
GitHub
organization,
so
I
have
opened
the
issue
like
we
discussed
two
weeks
tricks
ago,
mentioning
that
there
is
a
two
weeks
notice
and
if
nobody
jump
in
I
would
I
will
archive
the
repository.
B
So
it's
that's
just
for
information
and
also
I
haven't
done
it
yet.
But
I
will
open
to
issue
on
the
team
Compass,
but
we
can
start
discussing.
If
nobody
has
anything
else
to
say
we
should.
B
We
should
Define,
probably
a
bit
more
strongly.
B
What
are
what
are
the
rights
of
the
members
on
GitHub,
because
basically
you
know
that
we
have
the
console
the
official
console
like
since
a
bit
more
than
a
year,
I
think
no,
there
is
a
team,
that's
the
console!
There
is
also
a
specific
team
that
has
the
power
to
release.
It's
not
the
same
team,
but
there
is
no
rules
for
who
is
in
the
release
team,
and
there
is
also
no
rules
to
Define.
B
The
ID
that
we
need
to
have
behind
that
discussion
is
that
the
console
is
too
big
for
everybody
to
be
owner
of
everything.
So
there
is
21
members,
I
think
on
the
console
and
for
security
reason.
We
should
not
have
so
many
people
having
higher
rights
for
publication
and
for
ownership
of
the
GitHub
organization
and
that's
the
reason
we
should
clarify
and
probably
write
down
somewhere.
B
What's
the
process
for
that,
and
and
like
maybe
having
some
something
like
for
the
software
steering
console
representative
to
have
some
kind
of
Elections
and
say
that
every
year
we
do,
we
decide
like
we
refresh
the
list
of
people
able
to
release
and
people
that
have
ownership
on
the
GitHub
organization.
B
So
that's
one
of
the
issue:
I
want
to
open
and
the
other
one
is
we
discuss
it
that
we
did
not
go
into
the
detail
of
changing
the
versioning
of
the
packages
with
Twitter
app
for
meaning
that
we
we
should
try
to
to
update
the
version
of
the
only
of
the
package
that
get
updated
a
bit
like
what
we
are
doing
for
liminal.
B
This
will
is
the
the
work
of
Downstream
extension
and
in
the
long
term,
we
can
hope
that
with
new
major
version,
most
of
the
API
of
core
won't
change,
and
so
it
is
also
upgrading
extension
to
a
new
major
version
of
Jupiter
lab.
B
But
we
definitely
should
discuss
a
bit
more,
the
detail
of
that,
because
it
has
eye
implication
on
the
on
the
tooling
we
are
using
for
now,
because,
for
instance,
every
time
we
do
a
new
mineral
release,
we
create
a
new
branches,
we
bump
old
version
and
all
versions
are
aligned
and
so
yeah.
There
are
lots
of
questions
for
like
handling
the
the
code
and
the
repositories
it's
itself.
If
we
change
the
how
we
do
the
versioning.
A
So,
with
relation
to
releasers
what?
If,
whenever
there's
a
release
upcoming,
we
asked
for
members
of
the
council
who
is
interested
in
doing
a
release
and,
however
volunteers
gets
put
in
that
list
and
then
the
policy
is
I,
don't
know
once
a
year
twice
a
year
or
something
what
we
do.
Is
we
just
check
with
every
member
of
that
list?
Hey
do
you
still
want
to
be
on
here
and
if
they
don't
respond
within
I,
don't
know
three
weeks
or
something,
then
they
automatically
drop
until
the
next
time
they
volunteer
again.
A
But
that
way
we
don't
have
to
nominate
anyone
or
anything.
It's
just
whoever
volunteers
to
do
a
release
and,
as
a
last
resort,
the
SSC
rep
always
is
on
that
list.
Just
to
make
sure
the
list
never
goes
to
zero,
but
really
it
shouldn't
go
to
zero,
because
we've
had
lots
of
people
volunteer
to
do
releases
in
the
past.
So
hopefully
we'll
continue
to
have
that
in
the
future.
A
A
A
A
B
A
If
we
see
not
enough
redundancy
there,
we
just
simply
say
Hey
look
is
anyone
who
does
not
work
at
company
X
interested
in
doing
this
release
this
time,
because
it
would
be
healthier
for
us
to
release
that
way,
because
we
might
all
take
the
day
off
or
go
to
a
company
thing,
or
you
know
the
bus
factor
and
all
that
stuff
right.
So
we
could
try
to
organically
address
this.
C
A
A
B
Yeah
then
I
start
to
or
start
a
PR
with
it
directly
and
with
those
those
ID
in
it,
yeah
cool.
So
sorry,
Eric
yeah.
A
Just
because
the
python
package
called
Jupiter
lab
was
getting
a
major
version,
but
we
were
only
going
to
do
that
when
it
became
absolutely
necessary
following
the
rules
of
semper,
because
those
are
Standalone
things
that
people
consume
independent
of
Jupiter
lab
and
if
they
don't
need
to
be
bumped,
then
everyone's
better
off
at
moving
slowly.
So
that
was
the
that
was
the
premise
now.
The
practice
of
it
isn't
just
one
simple
thing:
right:
it's
a
little
complex
because
of
dependencies
between
each
other
and
also
because
of
lumino,
and
so.
D
Yes,
now
please
go
ahead.
Yeah
the
the
I
ask
I,
sometimes
ask
myself.
As
an
extension
author
I
mean
I
kind
of
want
to
be
using
whatever
the
latest
versions
of
those
packages
are,
but
I.
Think
kind
of
right,
like
each
release
of
lab,
might
use
a
different
set
of
those
npm
packages.
A
A
D
A
C
C
C
There
is
also
this
tool
that
we
have
where
you
can
use?
What
is
it
built
in
the?
What
is
it
dependency
update
or
something
like
that?
It
builds
utils
triple
Cloud,
build
utils
I
think
has
this
one
tool
where
you
can
pump
generic
JavaScript
package
version
and
you
can
put
in
like
a
regex
to
match
all
the
Jupiter
lab
packages,
for
example
the
latest
that
could
be
expanded
upon.
A
I
wonder
if
we
could
provide
a
Jupiter,
a
jlpm
command
that
says
jlpm
ad,
well,
not
jlpm,
but
it
would
have
to
be
Jupiter
because
it's
we're
going
to
do
custom
stuff,
but
anyway,
add
Jupiter,
lab
slash.
Application
for
Jupiter
lab
equals
three,
or
something
like
that.
Like.
A
We
could
create
a
command
line
tool
that
lets
an
extension
off
the
target.
A
version
of
lab
I
mean
the
other.
The
other
approach
to
this
is
a
thing
that
we
haven't
made
that
much
progress
on
yet,
but
we've
been
at
least
kicking
the
ball
around
some,
which
is
writing
extensions
that
are
a
bit
less
dependent
on
these
package
versions
because
of
what
they
consume
and
what
they
export.
A
E
A
Not
doing
like
live
objects
being
passed
that
way
if
instead
you're
doing
your
application
construction
with
a
different
set
of
Primitives,
then
here's
a
live,
JavaScript
object,
so
it's
possible.
We
end
up
having
a
whole
different
kind
of
extension,
while
also
supporting
these
the
same
way
we
used
to
have
npm
only
extensions,
but
now
no
one
really
deals
with
the
npm
side.
A
D
D
I'll
also
add,
while
you're
talking
about
that,
if
you
imagine
that
as
kind
of
a
higher
level
API
because
I
know
I'm
sure
you're
like
super
excited
about
me,
throwing
another
thing
into
the
mix
but
like
I
I,
just
think
of
the
possibility
for
adding
other
language
endpoints
for
that
really
really
high
level
API
like
having
a
having
a
JavaScript
API
that
uses
that
very,
very
high
level
API
scheme.
You
know.
A
A
D
D
A
I
think
we
should
get
a
group
of
people
together,
I
think
Fred
tried
and
we
didn't
get
a
group
together.
So
we
could
try
again,
but
yeah
I
mean
not
just
thinking
about
lab
five,
but
also
thinking
about
sustainability
of
extensions
in
the
ecosystem
if
they
all
start
to
suffer
from
bit
raw
there's.
Also
the
concern
of
how
hard
it
might
be
to
support
lots
of
different
types
of
geometer
front
ends
because
that's
happening
and
so
yeah
other
people,
please
chime
in.
F
D
F
This
is
fortuitous
too,
with
some
of
the
discussions
we've
been
having
about
the
schemas
being
able
to
extend
those,
you
know
that
has
a
pretty
strong
sustainability
feature
to
it
if
they're
used
and
implemented
properly,
but
it's
exciting
that
y'all
are
talking
about
this.
G
F
Thing
yeah
yeah.
A
F
B
Good
question
for
versioning
I
don't
know,
but
you
definitely
need
to
to
think
of
it
a
bit
more
about
what
we
will
do,
because,
even
if
it's
just
not
bumping
the
major
version,
there
are
some
stuff
to
do
at
least
on
the
on
the
touring
of
the
project.
But
yeah.
D
Maybe
we
are
like
make
a
tool,
one
that
can
maybe
pull
data
out
of
the
meta
package
and
post
that
into
the
documentation
somewhere
or
maybe
the
change,
the
change
log,
because
we
already
do
a
change.
We
already
do
automated
change
logs
I,
wonder
if
you
could
at
at
a
minimum,
just
communicate
if
you're
not
going
to
bump
them
all
then
communicate
somewhere.
A
If
the,
if
there
could
be
a
job,
that
puts
a
table
like
this
in
the
extension,
author's
guide,
I
think
that's
the
ideal
place
and
that
sort
of
solves
the
problem
of
how
to
communicate
it
as
well
for
sure
I,
don't
know
how
hard
that
is
to
generate
that.
Doesn't
sound
that
hard,
because
the
branches
have
well-formed
names
versions
are
queryable
both
from
GitHub
and
from
Pi
Pi.
So.
B
B
I,
probably
some
kind
of
git
diff
between
two
known,
commit
and
and
then
printing
that
nicely
in
the
documentation
could
do
the
drop
here.
A
C
A
A
D
I
can
put
this
on
my
to-do
list,
though
I
know
that
I
probably
won't
be
doing
it
very
soon.
Well,.
H
A
Okay,
there's
a
note
to
remind
people
that
accessibility
is
15
minutes
after
this
call
ends.
Well,
no,
it
is
47
minutes
from
now
there
you
go,
because
this
call
I,
don't
know
what
it
does
and
the
next
person
up
is
Mike.
H
Oh
yeah
so
I
recently,
prompting
to
a
problem
of
after
porting,
an
extension
to
Jupiter
lab
fall.
I
wanted
to
avoid
introducing
dependency
on
lab
4
so
that
users
of
future
notebook
version
7
to
be
released
sooner,
don't
have
to
install
Jupiter
lab
when
installing
that
extension
and
and
that's
spectacular,
but
also
LSP
possibly
and
the
problem
is
that
when
you
use
Python
Packaging,
there
isn't
a
nice
way
to
do
that.
Declaratively
in
the
metadata,
because
if
we
say
that
we
Bond
version
all
the
other
like
older.
F
A
H
Sorry
new
version
then
free.
We
we
will
get
that
installs.
There
is
no
or-
and
there
is
no
indication
as
far
as
I
know
where
if
anyone
knows
how
to
work
around
that
and
open
the
propositions,
but
there
is
one
one
way
that
can
be
done
is
dynamically
Define
install
requires
or
on
your
equivalent,
which
might
be
very
bad
but
I
I
didn't
okay,
I.
Think
Nick
has
some
comments
in
the
chat.
H
Yeah,
so
what
do
you
all
think?
What
what
is
the
solution
here?
Should
we
try
to
do
that
dynamically
and
also
stream
argument,
hatch.
So
that's
it's
easy
to
say,
or
have
some
other
algorithmic
to
other
to
use
it
that
takes
station
will
not
work
with
Jupiter
lab
free
or
older,
and
it
will
not
work
with
notebook,
6
or
older
and
and
otherwise
we'll
either
install
a
new
version
of
lap
or
notebook
or
will
do
nothing.
E
Unfortunately,
so
it
is
this.
Is
it
purely
a
client-side
thing
and
you're,
just
you're,
just
riding
on
pip
to
do
the
dependency
resolution
which
you're
not
resolving
yeah?
It
won't
really
do
that
I
mean
conda
has
has
like
weird
optional
dependencies
that
are
different
from
extras.
You
can
make
an
extra
you
know,
recommended
lab,
recommended
notebook
or
something
like
that,
but
extras
don't
work
over
time
either.
So,
yes,.
H
E
E
Yeah
there
isn't
really
a
good
answer
for
that.
I
mean
you'd
have
to
look
at
the
specific
case
like
in
the
case
of
a
widget.
E
C
H
So
what
is
your
your
opinion
on
Dynamic
King
store
requires
what
what
is
the
downside
of
that
I
I?
Think.
E
Users
download
wheels
and
wheels.
Don't
like
you,
you
can
you
the
only
Dynamic
ones
that
work
in
wheels
are
ones
that
are
based
on
Extras,
which,
like
magnifies,
how
broken
things
are
so
yeah,
but
the
setup
I
thing
doesn't
do
not
assume
your
setup.
High
will
ever
be
run
on
an
end
user's
computer.
H
E
Like
we're,
never
gonna
get
a
we're,
never
gonna
get
a
a
specifier
flag
in
the
mini
language,
in
Packaging
and
in
PIP
and
in
six
other
places
where
it's
implemented
like
there's,
never
going
to
be
a
Jupiter
lab,
underscore
version
like
a
python
version.
It's
not
gonna
happen
so.
E
H
So,
in
spell
tracker
extension,
for
example,
if
I
declare
Jupiter
lab
for
or
newer
I
will
force
users
of
notebook
to
install
Jupiter
app.
H
Okay,
maybe
we
can
follow
up
applying
I'm,
not
sure.
E
H
A
A
It's
there
I
think
we're
moving
to
a
world
where
these
siblings
are
very
compatible
with
each
other,
but
yeah
I
mean
there
might
be
some
edge
cases
we're
missing
here
and
if
we
find,
as
people
build
more
applications
that
it
does
make
sense
to
basically
turn
the
Jupiter
lab
repo
into
more
of
a
specific
spin
of
a
Jupiter
front
end
and
to
move
a
whole
bunch
of
core
into
a
Jupiter
front
end.
We
could
do
that.
H
Yeah
that
sounds
like
a
very
like
reasonable
future,
possibly
Lab
5
or
something
all
right.
Thank
you
very
much.
That's
Source
my
question
on
kind
of
openness,
a
vision
for
the
future,
so
two
other
points
are
just
to
highlight
some
of
regressions
in
4.0,
which
users
complained
repetitively
or
didn't
receive
much
of
an
attention
and
unassigned.
H
H
So
that's
a
severe
progression,
syntax
highlighting
in
Python
built-in
functions
and
it
doesn't
work,
but
there
is
kind
of
a
draft
of
a
workaround
that
we
could
Implement.
If
anyone
is
interested,
I
can
give
more
hints
on
that.
One
and
finally,
horizontal
scrolling
doesn't
work
I'm,
not
sure
if
it's
only
Firefox,
but
there
are
some
issues
like
people
having
large
and
those
data
frames.
H
Right
and
I
highlight
that
free
requests
with
needs,
review,
label
and
I
think
that's.
One
of
those
can
be
closed
as
Frederick
is
commenting
and
to
other
are
pending
career.
Okay,.
A
Anyone
going
on
oh
look
at
that
got
some
volunteers
to
review
this
thanks.
Bro,
okay,
cool
well,
we've
reached.