►
From YouTube: JupyterLab Team Meeting - May 11, 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/co...
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
A
If
you
haven't
already,
please
take
a
look
at
the
agenda
which
is
in
the
chat
and
add
your
name
to
it.
If
you
wish
and
add
a
bullet
point,
if
there
is
something
you
would
like
to
discuss,
if
you
have
a
longer
thing,
then
say
roughly
five
minutes,
then
please
add
it
to
the
discussion
section
at
the
bottom
and
I
will
try
to
do
a
better
job
of
actually
keeping
people
to
the
five-minute
role
this
time.
A
B
Hey
everyone.
I
tried
to
not
take
as
long
as
last
week,
so
some
announcements
first
they're
gonna-
be
the
performance
meeting
after
this
one.
If
you
are
interested
so
just
stay
connected
I'll,
be
sharing
my
latest
work
on
the
virtual
notebook
because
I'm
back
to
that
subject,
if
you
want
to
to
join,
feel
free.
The
next
announcement
is
follow-up.
So,
since
a
couple
of
weeks
we
have
discussed
the
opportunity
of
introducing
a
ui
toolkit
in
jupiter
lab,
there's
been
a
vote
and
the
vote
on
the
team
compass
got
it.
B
So
I'm
gonna
probably
do
some
some
pr
in
the
coming
in
the
coming
time
like,
but
nothing
mandatory
for
for
oh,
like
it
can
be
differently
done
afterwards.
That's
not
priority
for,
for,
for
all
and
on
the
technical
side,
there's
been
some
discussion
and
I
think
the
subject
is
coming
back
a
couple
of
time
about
crypto
error.
B
B
So
on
the
lumino
site,
there's
been
a
a
pr
that
replaced
that
code.
So
basically
the
package
has
two
version:
he
has
a
version
for
node,
that's
using
crypto
and
he
has
a
version
for
the
browser.
So
it's
not
using
crypto
and
that
resulted
in
the
1.12.0
version
for
core
issues
and
I'm
wondering
if
the
error
people
are
seeing
stream
because
it's
usually
happening
when
they
are
trying
to
build
some
some
extension
or
some
customized
jupiter
lab
with
pre-built
extension.
B
I
think
I'm
wondering
if
the
reason
for
for
the
error
still
showing
up
it's,
because
we
are
not
constraining
the
version
of
the
qualities
to
be
1.12
but
to
be
1.11
in
the
three
dot
x
versions.
So
the
question
I
get
is
that,
should
we
bump
and
release
it.
B
Yeah
you're
right,
william,
the
the
alternative
is
definitely
people
can
either
customize
the
webpack
config
and
I
think
vidarf
has
tested.
Also
that's
a
possibility
to
to
do
that
directly
in
the
package
to
json.
I
think
you
can
also
do
some
kind
of
fallback
there.
So
yeah,
that's
that's
the
question
like.
Should
we
advertise
only
that
configuration
or
so
we'll
bumping
the
dependency
in
the
jupiter
lab
help?
Don't
stream
user
not
having
that
issue?
C
Yeah
no
worries.
I
had
a
bit
of
a
mixed
experience
trying
to
solve
this
for
my
package
and
I
didn't
go
to
the
full
root
of
it.
I
basically
made
it
work.
I
tried
different
things
until
something
stuck
and
worked,
and
then
I
didn't
fully
find
the
root
cause.
C
C
C
B
Yep,
as
I
said,
I'm
not
really
familiar
with
everything,
especially
because,
as
soon
as
there
are
extensions
involved,
we
start
to
be
in
uncontrolled,
even
less
controlled
environment
for
for
note
packaging
resolution.
So
it
can
be
very
tough,
so
yeah,
I'm
fine
to
say,
let's,
let's
advertise
on
the
issue
and
hopefully
with
time.
D
B
I
don't
know
I
mean
it
makes
sense,
I
think,
to
upgrade
it
probably
gonna
help
resolving
the
the
environment
of
downstream
user.
So
I
would
say
it's
it's
probably
the
right
thing
to
do.
D
Has
been
a
pretty
safe,
upgrade,
there's.
B
Very
minimal
changes,
and
it's
what
I
think
so,
but
I
was
thinking,
let's
bring
it
to
discussion
just
to
be
sure.
D
Like
if
there
was
a
concern,
we
can
patch
our
webpack
to
do
the
browser
multiple
fix,
which
is
what
william's
suggesting
and
other
people
on
the
issue
are
suggesting,
but
I
wouldn't
be
opposed
to
upgrading
the
lumino
dependency.
D
Okay,
in
fact,
since
the
limit
of
dependency
is
a
carrot,
all
we
have
to
do
is
really
just
bump
it
in
the
yarn
lock
right,
yeah.
A
D
A
D
There
is
a
browser
version
of
it,
which
is
this
crypto
dash
browser
for
zerify,
that's
mentioned
in
the
issue,
so
you
can
resolve
to
that.
To
get
the
crypto
library-
or
you
can
just
say,
look
if
we're
running
in
the
browser.
All
the
modern
browsers
that
we
have
these
days
support
the
the
crypto
random
number
generator.
So
we
shouldn't
call
it
anyway
and
that's
why
the
cryptocolon
false
works
as
well,
because
lumino,
I
think,
never
hits
that
fallback.
D
D
B
C
C
I
can
see
from
the
staging
yarn
lock
this
already
1.12,
which
so
it
may
be
that
they're
pulling
in
two
volts
of
a
version
of
jupiter
lab
right.
If
they're
trying
to
build
it
with
3.2
or
3.1,
maybe
they
have
the
old
version
locked
in
then
we
would
have
to
do
like
a
patch
version
of
all
of
these
other
older
versions
right.
B
But
that's
the
weird
thing,
because
the
issue
states
it's
jupiter
3.4.
So
but
I'm
wondering
if
there
are,
they
are
not
using
source
extension
or
those
extensions
that
may
pin
car.
It
is
at
another
version.
That's
that
I
don't
know.
E
B
A
Just
a
comment
on
your
first
point
about
the
ui
toolkit,
so
we
left
the
vote
open
longer
than
I
think
required,
but
we
had
at
the
moment
the
boat
was
open.
There
were
13
people
on
the
jupiter
lab
council.
There
are
now
14,
but
we
had
eight
out
of
13
people
vote.
They
all
voted
yes,
so
it's
not
particularly
controversial.
A
A
So
hopefully,
hopefully,
a
votes
are
kind
of
rare
because
we
arrive
at
things
by
consensus
but
b
next
time,
we'll
have
some
patterns
for
the
tags
and
the
the
labels
in
the
team
compass
and
if
there's
a
label
for
vote
and
there's
a
label
for
one
of
owen's
decision
made
that
comes
out
of
the
decision-making
guide
and
the
governance
talks,
and
hopefully
that
makes
it
easier
to
follow
issues
if
you
care
about
specific
decisions
being
made,
as
opposed
to
other
issues
that
are
in
the
team,
compass
and
yeah.
F
Hi,
so
it's
just
a
small
update
for
me.
I
have
posted
a
pr
that
I
submitted
against
the
3.4
branch
to
bump
the
version
of
a
library
called
mart.
F
It
was
detected
as
detected
as
a
vulnerable
so
and
it
was
used
in
rendermind
and
sales
packages
yeah
thanks
frederick,
for
guiding
me
with
the
failing
ci
tests
in
there.
I
just
have
one
small
failure
right
now.
F
I
think
it's
a
like
a
python
3.7
test,
I'm
not
sure
that's
flaky
test
or
what
so,
if
somebody
could
just
please
take
a
look,
and
let
me
know
if
there's
anything
else,
we
need
to
address
that'd
be
great
and
also,
if
we're
planning
a
patch
release
or
miner
for
304,
like
maybe
that
one
can
go
in.
A
Awesome,
thank
you.
I'm
gonna
re-run
this
job
because
it
seems
unlikely
that
that
would
be
related
to
your
change,
so
hopefully
it'll
pass
and,
if
not,
we'll,
take
a
look
and
try
to
figure
out
what's
wrong
with
it.
But
yeah
thanks
for
fixing
the
vulnerability.
B
Yeah
cara,
it
should
be
probably
just
flaky.
It
seems
that
the
registry
from
your
own
package,
just
failed
during
the
download
of
some
packages,
so
should
be
fine.
G
G
I
don't
think
she's
on
the
call
right
now,
but
in
any
case
I'm
curious
about
the
the
may
community
call
currently
on
the
calendar.
It's
scheduled
for
may
31st
at
8
am
which
does
not
conflict
with
any
other
jupiter
meeting,
although
it
is
the
day
after
memorial
day
in
the
u.s,
which
means
a
lot
of
people,
including
myself,
will
be
taking
additional
days
off
around
that.
So
I'm
curious
about
whether
folks
want
to
keep
it
there
or
move
it
around.
A
I
I
defer
to
whoever
is
going
to
host
it.
A
G
See
I
see
tony
said
I
can
relay
the
question.
Thank
you
tony.
So,
let's
we'll
we'll
I'll
defer
to
tony
to
continue
this
discussion.
Thank
you.
A
D
Thanks
yeah,
I've
started
using
my
github
handle
just
because
it
contains
my
last
name
and
it's
a
link
that
might
be
a
pattern
we
could
consider
for
for
these
names
in
the
agenda,
so
that
it
is
a
little
more
descriptive.
D
I
went
to
pycon
last
week
and
pycon
us
and
there
was
a
number
of
mentions
of
jupiter,
lab
and
interesting
things
that
I
thought
people
here
would
be
interested
in.
Knowing
we
had
a
jupiter
tutorial
by
mike
muller,
I
went
and
helped
out.
There
was
maybe
around
80
attendees.
It
was
packed
like
I
had
to
sit
on
the
floor
for
a
little
bit
type
of
thing
and
lots
of
good
discussion
and
lots
of
good
questions
and
a
bug
or
two
that
came
out
of
it
bug
reporter
too,
that
came
out
of
it.
D
The
jupiter
lab
in
the
expo
hall
you
wander
around
and
several
products
in
expo
hall
are
just
based
on
jupiter
lab.
You
can
see
it's
like
coiled
io,
of
course
uses
jupiter
lab
and
there
was
another
simulation
framework
that
was
totally
based
on
drupal
apps.
So
we
talked
about
coming
to
the
jupiter
meetings
and
being
part
of
the
community
is
a
good
conversation.
D
One
interesting
announcement
at
the
conference
was
the
pi
script
project.
So
many
people
here
are
aware
of
pyodide,
which
is
the
compiled
python
to
wasm
and
run
it
totally
in
your
browser.
It's
the
foundation
of
jupiter
light,
which
maybe
some
people
here
have
worked
on,
and
certainly
a
lot
of
people
have
seen
the
underlying
project
for
jupiter
light
is
pyodite
project
and
a
different
project,
that's
being
built
on
top
of
piyo.
Diet
is
called
pi
script,
so
pi
script
is
the
big
announcement
from
peter
wang
at
the
conference.
D
Anaconda
has
been
working
on
this
and
investing
in
it.
It's
essentially
python
in
the
web
browser
plus
dom
access,
so
it
adds
on
top
of
pio
died,
which
is
just
running
python
in
your
web.
D
Browser
dom
access
a
little
more
polish
around
wrapping
libraries
and
things
like
that
from
the
piodine
project,
try
to
just
make
it
a
little
more
polished
and
a
little
more
accessible
when
we
were
talking
about
what
ramifications
there
are
for
python
in
the
web,
browser
which,
by
the
way,
is
becoming,
as
I
understand
it,
an
official
target
for
python
see
python.
Very
soon,
people
are
working
on
upstreaming.
D
Basically,
the
python
packages
patches
to
corsi
python
so
see
python
has
a
is
supported
as
running
in
the
in
the
browser
under
wasm,
as
I
was
talking
to
the
iodine,
maintainers
and
others
about
like
what
does
it
look
like
to
have
python
in
the
browser
you
know.
Is
this
the
year
of
python
in
the
browser
one
thought
that
came
to
us
was:
wouldn't
it
be
nice
if
we
could
write
superlab
extensions
in
python?
D
Can
we
offer
them
the
ability
to
customize
their
environment
in
python,
and
so
so
I
started
working
on.
You
know
converting
the
plug-in
playground
to
run
a
python.
You
know
expose
the
jupyter
lab
api
as
python
objects
and
let
you
just
write
an
extension
in
python
that
connects
to
the
various
services
in
jupiter.
D
I
got
just
a
few
minutes
into
it
and
realized
that
the
pyodide
project
doesn't
compile
on
webpack
and
everything,
and
I
spent
the
rest
of
the
sprints
working
on
iodide
and
working
with
the
piodide
folks.
But
I
think
this
is
a
really
interesting
possibility.
Right,
jupiter,
extensions
in
python
and
and
pio
died.
D
Piedid
makes
it,
I
think
possible
to
do
that
and
to
experiment
with
it.
So
anyway,
I
don't
have
time
to
like
devote
my
full
time
to
it
right
now,
but
this
is
a
project
that
I'm
interested
in
pursuing
sort
of
in
in
spare
time.
If
anybody
else
is
interested
in
looking
at
it
I'll
open
an
issue
and
jupiter
up
and
we
can
discuss
there,
and
maybe
we
can
arrange
a
time
to
maybe
sprint
together
on
it
or
something
I
guess
so.
D
One
other
thing
I'll
mention
so
st
carol
willing
was
there
steve
sylvester
was
there,
so
we
had
a
a
jupiter
office
hour
that
a
few
people
came
in
and
and
in
general
there
were
various
other
conversations
at
the
conference
about
using
jupiter
using
widgets,
using
jupiter
lab
and
in
in
in
their
workflow
as
opposed
to
the
pets.
Pycon
has
typically
been
a
very
general
conference.
There
were
a
ton
of
people
there
these
days
in
data
science,
and
there
was
talk
about
next
time,
maybe
having
a
whole
data
science
track.
D
So
it
might
be
a
conference
that
you
would
like
to
put
on
your
radar
for
next
year.
It's
in
salt
lake.
In
april
I
was
at
salt
lake
city
this
year,
it'll
be
in
salt
lake
city
next
year.
Anybody
that
comes
I'm
happy
to
host
a
tailgate
party
at
my
house.
I'm
now
local
in
salt
lake
it'd
be
a
lot
of
fun
to
see
people
there,
but
I
think
it'll
be
a
good
conference
like
rising
to
the
level
of
scipy
as
far
as
importance
or
pi
data
conferences.
D
H
Okay,
so
these
are
some
questions
that
came
up
internally
as
they're
upgrading
from
jl1
to
gl3.
The
first
one
is
around
ipad
kernel.
Six
and
I
didn't
know
how
to
answer
that.
To
them.
Are
there
any
known
problems?
Is
it
okay
to
ship
jupiter
lab
3
with
ipad
kernel,
6
versus
5.?
H
Are
there
any
known
problems
around
that
vegetable
3?
Does
anyone
know.
H
Okay,
the
second
one
is
around
a
customization,
so
the
the
thing
that
they
want
to
achieve
is
that
when
you
have
a
remote
kernel
and
you
want
to
have
a
terminal
open
with
a
remote
kernel,
the
current
behavior
in
gl3
is
that
it'll
just
hang
on
for
the
websocket
until
the
kernel
is
available.
So
that
takes
about
a
couple
of
minutes
and
there's
no
terminal
window
available
in
jl1.
H
There
was
a
way
to
customize
that
by
overriding
some
of
the
functions
so
that
you
can
create
a
dummy
terminal
without
any
name
and
then
update
the
connection
once
the
term.
You
know
once
the
connection
was
available,
I'm
just
wondering
if
anybody
else
had
tackled
this
problem
in
jupiter
lab
3.
Is
it
possible
to
have
a
better
experience
in
triple
b3
yeah.
H
I
think
the
ones
then
are
you
mute?
Oh
sorry,.
A
D
A
Okay,
I
had
a
quick
announcement
next
tuesday
in
the
hour
before
this
hour,
so
8
am
pacific
time.
4
pm
uk
time
is
the
second
notebook
v6
to
v7
extension
migration
office
hours.
That's
a
that's
a
mouthful,
but
the
idea
is
if
an
extension
author
who
wrote
an
extension
for
the
classic
notebook,
is
interested
in
having
some
guidance
or
has
some
questions
and
how
to
make
their
extension
work
in
notebook
version,
7
or
jupyter
lab
using
the
new
apis.
A
We
are
there
to
help
last
week.
No
was
it
last
week
last
time
we
held
it
whether
it
was
last
week
or
not.
It
was
last
week
we
actually
ended
up
talking
about
an
extension
that
wasn't
a
classic
one,
but
it
was
a
helpful
conversation
anyway,
and
this
coming
one
please
come
if
you
have
an
interest
in
any
of
those
things,
whether
or
not
you're
migrating
an
extension,
because
I
guarantee
I'll
be
there,
but
I
may
be
there
by
myself,
yeah.
C
Vidarn
so
for
those
who
have
extensions
which
distribute
both
the
classic
nba
extension
and
the
jupiter
lab
extension.
C
Well,
we
don't.
We
won't
have
to
do
anything
about
the
classic,
nb
extension
stuff
right
that
there's
no
shims.
That
will
try
to
pick
this
up
or
anything
like
this
by
default.
We
can
just
leave
that,
as
is
to
retain
support
for
version
six
correct,
so
if
we
have
it
put
the
classic
in
the
lab
extension,
the
lab
extension
will
probably
work
if
it
integrates
well
with
the
the
ui,
the
lack
of
like
specific
zones
in
the
areas.
A
That's
right
so,
if
you've
already
written
a
lab
extension,
you
might
want
to
give
some
thought
to
the
user
experience
of
what
that
lab
extension
feels
like
in
notebook
version
7,
because,
as
you
say,
it
has
a
different
application.
Shell
and
different
zones
where
your
extension
can
reside
and
but
you
shouldn't
have
to
re-author
much
of
your
code
if
any
at
all,.
B
Yes
about
this
issue,
let's
say
of.
B
Okay,
so
let
me
zoom
a
bit:
the
animation
won't
be
bigger,
but
the
the
screencast
is
not
fully
accurate,
but
I've
been
working
on
a
pr
to
be
able
to
customize
the
positioning
of
panels
or
widgets
from
different
factories
in
different
area
than
what
an
extension
is
proposing
by
default.
B
The
idea
is
to
to
use
that
features
in
the
coming
notebook
v7
so
that,
for
instance,
if
the
extension
you
are
implementing
in
jupyter
app
shows
up
in
a
sidebar,
but
you
want
it
to
be
on
like
the
tree
view,
like
the
the
the
view
in
which
you
get
the
file
browser
in
the
notebook,
it
will
be
as
simple
as
defining
the
settings
for
the
notebook
and
with
the
fallout
pr
for
notebook
7.
I
will
also
propose
something
to
introduce
some
some
area
for
the
left
right
and
probably
done
by
not
too.
D
Cool
jason-
I
was
just
following
up
on
vidar's
question
because
I
know
a
lot
of
ipi
widgets
maintainers
will
have
this
question
about
jupiter
lab
jupiter
notebook
version
7..
My
understanding
is
ipad.
Widgets
already
worked
in
retro
lab,
so
picking
up
their
jupiter
lab
extension,
so
it
sounds
like
widget
maintainers.
Just
don't
have
to
do
anything.
It'll
automatically
work
in
notebook,
seven
using
a
drupal,
app
extension
is
that
should
that
be
generally
true,
because
they're
not
doing
anything
sort
of
complicated
or.
A
A
A
Okay,
the
first
discussion
bullet
point
says:
pi
data
london,
sprint
question
mark
who
added
that.
D
I
added
that,
but
I
was
hoping
you
would
talk
about
it
darien.
I
can
give
a
brief
overview,
but
you're
probably
more
suited
to
talk
about
it.
Okay,.
A
Cool
so
pi
data
london
is
happening
in
the
middle
of
june
and
on
the
saturday
in
the
middle
of
it
saturday,
the
18th
we
are
going
to
have
a
sprint,
a
two-hour
sprint
on
jupiter
lab
issues
and
vidar,
and
I
are
going
to
be
hosting
it
and
I
think
the
plan
is
that
we'll
do
a
tutorial
that
is
similar
to
the
office
hours,
that
jupiter
server
has
been
doing
the
contributing
hours.
A
Sorry
we're
basically
we'll
show
people
how
to
do
a
small
pull
request
from
scratch
against
jupiter
lab,
and
then
we
will
share
the
a
curated
list
of
issues
so
good.
First
issues
help
wanted
with
the
participants
and
we
will
support
them
as
they
work
on
those
issues,
and
the
hope
is
that
we
get
some
new
contributors
from
this
event.
A
If
you
come
across
an
issue
whether
you're
coming
to
the
triage
calls-
or
you
are
just
seeing
them,
come
in
and
you
think
it's
appropriate
for
a
first
contribution.
Please
use
the
good
first
issue
label
that
we
have,
because
that's
super
helpful.
We're
often
short
of
those.
The
event
again
is
on
the
18th
of
june.
A
A
I
don't
think
the
schedule
is
on
the
site
yet,
but
when
it
is
it'll
be
on
there,
I
don't
believe
I
don't
believe
any
portion
of
the
sprint
ends
up
recorded
or
on
zoom
or
anything.
I
think
it's
an
in-person
event
and
meant
to
be
like
this
isn't
a.
A
Is
just
bringing
in
people
who
are
interested
in
contributing
to
the
project
and
again
that
is
in
london.
It's
near
tower
bridge.
I
forgot
where
exactly,
but
I
think
I
checked
them
out.
It's
a
kind
of
a
scenic
place
and
london
in
june
is
nice,
so
you
should
come.
D
So,
if
you'd
like
to
help
either
curate
a
good
first
issue
or
create
a
new
good
first
issue,
I
think
the
best
good
first
issues
are
ones
where
essentially,
most
of
or
even
all
of
the
debugging
has
been
done
for
the
person
and
essentially
it's
here's
the
problem.
Here's
kind
of
the
background
of
what's
going
on
behind
this
problem,
here's
the
lines
of
code
that
are
relevant
to
this
problem
and
then
like.
D
If
you
really
want
to
give
them
a
helping
hand,
here's
the
change
I
would
make,
or
here's
kind
of
the
idea
of
the
change
I
would
make
so
that
the
good
first
issue
like
they're
already
pretty
overwhelmed.
Often
with
you
know,
cloning
jupiter
lab
getting
a
dev
environment
set
up
and
everything
if
the
actual
code
change
that
they
make
is
extremely
simple
or
even
just
told
basically
to
them,
and
they
can
just
make
that
change
and
then
go
through
the
process
of
submitting
a
pdr.
D
Often
the
good
first
issues
are
more
about
helping
them
through
the
pr
process,
rather
than
diving
deep
into
the
internals
of
jupiter
lab
and
figuring
out
where
five
different
connections
live
and
everything.
If
we
can
do
that
work
for
the
people,
then
they
can
concentrate
on
the
process
of
becoming
a
contributor
and.
C
D
Look
for
another
issue,
that's
a
little
more
complicated
that
they're
going
to
have
to
dive
into
the
code
a
little
deeper
and
that's
a
good
second
issue.
So
anyway,
if
you'd
like
to
help
out
with
curating
for
good
first
issues,
that
would
be
an
excellent
way
to
improve
the
quality
of
our
good
first
issues.
A
Yeah
for
sure,
thanks,
jason
adam,
what
are
you
thinking.
I
A
A
D
F
D
A
I
just
thought
of
something
sort
of
fun
and
like
okay,
so
next
tuesday
is
call
right
if
it
turns
out
that
no
one
shows
up
for
help
migrating
their
extension
from
classic
to
v7.
It
seems
like
a
good
use
of
that.
Time
might
be
to
do
a
small
pull
request
against
jupiter
lab,
because
it's
the
same
skill
set
and
might
be
beneficial.
A
So
that's
one
time
we
could
do
it.
We
could
maybe,
as
part
of
the
documentation
blitz
that
comes
out
of
jupiter
lab
4
notebook
v7,
make
a
short
video,
that's
similar,
because
the
sort
of
content
we're
talking
about
aside
from
you
know,
bandwidth
aside,
it
should
really
be
kind
of
shorter
than
10
minutes
really.
So
the
idea
is,
you
already
have
a
github
account.
Let's
say
you,
you
fork
the
repo.
You
pull
down
your
clone
to
your
local
machine.
You
create
a
new
branch
and
get
you
make
your
change.
A
I
I
just
want
to
say
what
what
I've
always
been
confused
about
is
which
how
do
you
know
where
to
put
the
pr2?
Those
are
like
questions.
I
don't
understand.
I
don't
understand,
there's
so
many
branches
there.
Maybe
it's
just
because
I
don't
know
how
vr's
work
very
well.
Maybe
that's
the
problem,
but
that's
been
my
confusion.
I'm
afraid
to
do
anything
because
there's
it's
a
complex
project,
so
it's
one
of
the
barriers
I've
had.
I
D
Saul
played
around
with
the
idea
of
twitch
streaming
jupiter
lab
development,
and,
just
you
know
the
debugging.
This
is
a
little
bit,
maybe
above
the
level
of
what
you're
thinking
of
just
start
to
end.
Here's
how
to
submit
a
pr
which
I
think
we
should
have
as
well,
but
if,
if
anybody's
willing
to
record
themselves,
just
diving
in
and
and
figuring
out
a
jupiter
issue,
that's
also
helpful
yeah.
I
D
I
Example
on
issue
example:
maybe
just
something
very
simple:
you
know
and
then
once
once
that
is
done,
then
I
think
anyone
could
extrapolate
on
that,
because
then
it
becomes
a
programming
issue
and
not
a
how
to
how
to
work
in
this
environment
issue.
You
know
how
to
you
know
all
those
code
testing
things
you
got
to
run
your
code
through
when
you
do
a
pr.
Those
things
can
be
confusing.
I
A
Of
this,
oh,
that
that
was,
that
was
just
me
trying
to
figure
out
what
to
do
next
week.
In
case
nobody
comes
for
the
office
hours
for
notebook,
v67
extension
migration,
but
if
it
turns
out
that
that
is
not
an
appropriate
time,
I
think
I'll
give
it
some
thought,
and
we
can.
We
can
schedule
a
time
to
do
it
because
it
seems
like
if
we
record
this
it'll
be
beneficial.
I
C
So
if
you
have
written
documentation,
there
is
at
least
a
much
lower
threshold
for
fixing
docs.
Once
they
become
out
of
date.
They
will
be
grappable
these
kind
of
things
so
it
like.
For
the
the
moment,
we
moved
from
the
master
branch
to
the
main
branch,
whatever
instructions
you
have,
which
branch
your
target
will
be
outdated
right.
So
it's
those
hopefully
won't
happen
too
much,
but
you
can
constantly
change
stuff
about
the
which
tests
we
run
and
how
things
fails
and
what
you
should
run
locally
right.
C
So
it's
it
might
be
worth
having
it
you've
written
instead
of
a
video,
but
a
video
might
be
a
good
place
to
start
from
for
somebody
to
make
a
written
instruction.
A
Yeah,
maybe
maybe
a
blog
post
that
has
the
text
in
it
and
a
screencast
and
the
blog
post
has
the
quality
of
being
a
moment
in
time.
It
falls
into
a
timeline.
So,
if
you're
looking
at
a
blog
post
from
two
years
ago-
and
it
says,
hey
this
targets-
you
know
jupiter
lab
4.x
before
it's
been
released,
and
this
was
the
process,
then,
at
the
very
least
you
won't
get.
A
A
What
other
discussion
items
we
have,
though,
scipy?
What's
new
in
jupiter
lab
proposals
accepted
as
a
poster
with
the
possibility
of
a
backup
talk
who
wrote
that
and.
D
I
think
I
wrote
the
next
two
items,
so
just
fyi
we've
been
accepted,
martha
and
alex,
and
myself,
I
think,
were
the
ones
that
submitted
the
talk.
D
It
was
accepted
as
a
poster,
but
it's
a
given
a
backup
slot,
which
means
we
may
or
may
not
be
able
to
actually
talk
there,
but
anyway,
scipy
just
put
it
on
people's
radar
again.
Another
conference
middle
of
july
saipai
and
in
austin
it'd
be
a
fun
place
to
see
people.
J
Yeah
as
a
reminder,
we
also
have
a
tutorial
at
that
conference.
That
was
a
backup
and
did
get
accepted.
So
there
is
precedence
for
backup
to
get
accepted
at
that
conference.
D
D
A
A
B
B
B
So,
on
the
benchmark
repository,
I
have
set
up
some
automatic
tests
for
generating
the
html
dom
from
command
markdown
tests
using
jupiter
renderer
some
weeks
ago.
We
discussed
about
that
and
the
idea
was
trying
to
figure
out
what
are
the
source
of
the
discrepancy
if
they
are
desired
or
if
there
are
there
is
something
we
need
to
fix.
B
So
that's
what
I'm
trying
to
report
here.
There
is
three
things:
three
discrepancies
that
are
expected,
let's
say
or
want
it,
the
in
jupiter
app.
Every
time
we
are
converting.
B
A
heading
we
are
adding
it
some
id,
that's
computed
automatically
from
the
text
of
the
title,
and
we
are
also
creating
a
link
for
it.
So
that's
something,
of
course:
it's
not
done
by
command
mark
and
the
other
thing
that
we
are
doing
that.
I
didn't
know.
Actually
it's
because
you
are
using
a
code
mirror
for
highlighting
syntax
in
code
block.
We
are
having
some
additional
elements.
B
I
don't
think
I've
printed
print
it,
or
did
I
yeah
the
in
the
table
below
that
this
desired
discrepancy?
The
the
second
test,
the
10112.
So
when
there
is
a
language
here,
it's
ruby,
you
would
get.
If
you
you
go
on
the
right,
you
will
get
like
a
class
that
say
we
are
using
command
code,
mirror
jupiter
special
team
for
for
ruby.
B
So,
and
so
we
are
getting
much
more
dumb
note
that
what
command
mark
is
doing
by
default
because
he's
just
like
using
a
code
block,
so
the
the
those
stores
are
expected
and
the
other
one
that's
expected
is
we
are
using
a
sanitizer
by
default,
so
some
stuff
get
converted
or
suppressed
in
jupiter.
B
Up
like
the
main
example
here
is,
there
is
a
test
with
a
script
tag,
and
so
the
script
tag
is
removed
for
us
and
there
are
some
some
styling
or
so
that
that
got
removed
by
time
to
time,
so
that
that
I
presume
it's
fine
and
also,
I
think
a
consequence
of
the
sanitization
is
in
the
test.
There
are
a
couple
of
tests.
B
That
are,
and
it's
like
in
the
second
part,
below
the
first
table
like
the
two
fixed
discrepancies,
there
are
a
couple
of
tests
that
result
officially
in
unbalanced,
html
block
and
in
jupyter
lab
that
get
fixed
somehow,
and
I
think
it's
the
sanitization
step,
that's
converting
or
closing
some
div
that
are
partially
open.
So
I'm
not
sure
is
it's
something
to
be
fixed.
I
think
it's
expected,
but
I
won't
like
I'm
not
sure.
B
So
yeah,
that's
what's
come
out
of
it.
I
also
realized
that
so
we
are
using
the
javascript
mark,
gs,
library
that
we
discussed
a
bit
earlier
for
converting
markdown
syntax
and
the
indoor
cid
are
also
running
against
those
tests
of
command
mark.
It's
just
like
they're,
not
covering
all
of
it,
but
they
are.
They
basically
are
covering
quite
lots
of
it,
like
I
think,
like
90,
95
percent
of
the
tests
are
passing
directly
by
margies,
so
that's
kind
of
a
good
news
on
that
front.
B
And
so
I
just
focused
really
on
common
mark.
I
know
that
I
I've
read
also
I
I
looked
at
nb
convert
because
it's
basically
also
one
of
the
big
tools
in
jupiter
ecosystem
that
is
important
to
align
for
rendering
mordon
and
they're
using
a
python
library.
That's
called
mistune
that
by
design
decided
to
not
be
compliant
to
all
tricky
case
of
command
mark
and
it's
written
like
on
the
front
page.
B
We
decided
that
there
are.
There
were
some
some
of
those
tests
that
are
like
normal
people.
Writing
more
down.
Won't
won't
try
to
to
use
those
kind
of
syntax,
so
we
decided
it
was
simplifying
the
parser.
So
we
decided
to
drop
it
to
something
efficient
in
term
of
fast
to
to
to
parse
so,
and
so
I
think
it
will
be.
It
will
be
tough
for
to
keep
that
that
library
on
nb
convert,
but
that's
another
discussion.
There,
probably.
A
B
The
main
trouble
I'm
seeing
is
if
we
want
to
have
alignment,
especially
between
nb,
convert
and
and
jupiter
labs.
So
that's
mean
you
are
talking
about
tools
in
different
language.
B
Keeping
mark
gs
may
be
tough
to
align
with
some
some
parser
on
the
python
site.
For
what
I
understood,
because
smart
gs
is
leaking
extension
like
plugin
system.
So
that's
that's.
Maybe
the
the
next
question
like
what
what
should
be
the?
What
could
be
a
good
candidate
like
lots
of
people,
have
heard
of
marked
on
it,
of
course,
but
I
know
that
some
people
have
trouble
with
it
in
the
sense
that
if
I
cut
it
correctly,
it's
sometimes
tough
to
have
feedbacks
from
the
the
maintainer.
B
And
other
than
that,
I
don't
have
neither
lots
of
experience
with
modern
parser.
I
know
that
there
are
lots
of
them
in
javascript
and
probably
on
byte
inside
too,
so
I
don't
have
like
a
preference
or
something
to
say
yeah.
This
one
looks
very
much
better
than
another.
One.
E
Yeah-
and
I
really
I
really
like
it,
what
it's
worth-
it's
cool
that
there's
a
python
implementation
of
it
that
tf
pointed
out.
I
didn't
know
that,
and
it
looks
like
it's
still
actively
maintained
by
the
maintainer
there's
new
commits
weather
coming
in
here
is
the
thing
I'm
most
interested
in
in
addition
to
the
cool
stuff,
you've
done
that
you
just
showed
us
is
precisely
when
jupiter
recognizes
something
as
being
math
jacks
and
should
be
processed
by
math
jackson.
E
You
know
which
dollar
signs
count,
which
begin
and
environments
count,
what
shifts
it
over
to
math.