►
From YouTube: JupyterLab Team Meeting - July 20, 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
july
20th
weekly
jupiter
lab
call.
Today
we
have
20
people,
which
is
great
21.
As
I
said
nice
in
the
chat,
please
find
the
link
to
the
agenda
and
add
your
name
if
you
wish
and
add
an
agenda
item,
if
there's
something
you'd
like
to
discuss,
that
is
on
the
order
of
five
minutes
or
so,
if
there's
something
longer
than
that,
please
add
it
to
the
additional
discussion
section.
A
B
A
Looks
like
the
discussion
items,
or
rather
the
agenda
items
are
going
to
exceed
five
minutes,
but
hey
we'll
do
what
we
have
to
all
right
with
that.
Why
don't
we
get
started?
The
first
person
on
the
agenda
this
week
is
friday.
C
So
I
have
2pr
and
if
somebody
feels
like
having
half
an
hour,
something
like
that
to
the
look
they
are
not
too
complex,
so
I
will.
It
would
be
much
appreciated
if
somebody
can
review
them,
especially
the
first
one,
because
it
would
be
nice
to
have
it
in
the
patch
fix
release.
I
would
like
to
do
tomorrow
for
3.4
thanks
daryn,
and
it
was
it
by
student.
I'm
I'm
monitoring
on
jupiter
upgrades
yeah
for
poor
student.
They
need
to
use
windows.
C
New
person
on
the
project
and
the
idea
was
to
it's
to
add
some
documentation
about
using
tools
for
automation,
of
setting
up
a
development
environment
for
a
jupiter
lap.
So
the
tools
he
is
explaining
his
vagrant.
C
So
as
I
wrote
in
the
notes,
I'm
not
I
mean
I
really
appreciate
the
the
user
sharing
the
the
code
and
the
script
he
has
done
and
and
I
would
rather
prefer
people
managing
those
kind
of
setups
to
be
in
a
guest
or
in
a
separate
github
repository
rather
than
having
it
officially
in
our
documentation,
because
I
have
always
the
impression
that
if
it's
in
our
documentation,
then
we
need
to
maintain
it
and
stuff
like
that,
and
we
already
what
was
the
name
of
the
the
service,
like
the
not
a
gitlab
kit,
but
gitbot
setup
was
in
our
official
documentation,
but
nobody
was
using
it
anymore,
and
so
it
was
broken
and
stuff
like
that.
C
So
I
I
would
rather
reply
that
thanks
a
lot,
but
please,
if
you
update
doc,
maybe
just
point
to
your
gist
or
something
like
that,
and
I
was
wanting
to
know
to
have
feedback
of
what
others
are
thinking
about.
That.
D
So
I
replied
on
the
original
issue,
suggesting
to
indeed
just
link
to
the
repo,
because
I
think
he
did
some
good
job
setting
up
some
example
repo,
and
that
would
be
enough-
and
maybe
the
request
is
a
bit
too
elaborated,
like
you
said
like
there's,
maybe
too
much
content
there
and
a
link
would
be
enough.
I
guess
right
and
if
the
link
becomes
dead,
then
we
will
catch
it
with
the
ci
and
we
could
remove
it
in
case
they
they
remove
the
ripples.
E
So
the
main
motivation
for
this
right
is
to
make
it
easy
for
people
like
on
spins
or
whatever,
to
get
quickly
set
up
with
a
dev
environment
where
you,
you
can
use
troop
to
lab
right
and
there's,
of
course,
always
some
changes
on
that.
So
the
advantage
of
taking
it.
A
E
C
A
Out
of
curiosity,
since
we
have
24
people
on
the
call
who
are
at
least
somewhat
interested
in
this
sort
of
thing,
could
you
do
maybe
a
thumbs
up
or
any
emoji?
Do
you
use
vagrant
or
have
you
used
vegan?
This
is
something
you
would
use.
B
A
That
x
means
no
right,
yeah,
okay,
cool!
I
have
used
favorite.
I
think
it's
all
right.
I
don't
have
a
strong
opinion,
though,
and
I
certainly
didn't
have
an
opinion
one
way
or
the
other,
but
this
is
interesting
cool.
A
A
Yeah,
that
sounds
that
sounds.
A
B
A
A
The
context
here
is
vidar
and
I
were
at
a
sprint
that
we
were
hosting
and
someone
who
had
a
windows
machine.
A
couple.
People
had
windows,
machines
have
difficulty
getting
a
dev
environment
set
up,
and
one
person
said
what,
if
you
had
a
vm
that
we
could
develop
on
and
that
person
happens
to
use
vagrant.
So
that
was
their
solution
to
this
and
we
said
cool
open,
an
issue
talk
about
it,
and
this
is
us
talking
about
that
issue.
Yeah,
so
cool.
B
B
Like
one
of
the
things
that
has
helped
us
with
this
is
git
pod,
I'm
not
really
like
all
that
big
on
it
personally,
but
like
getting
a
dev
environment
set
up
and
making
prs.
That's
been
really
successful
in
like
quite
a
few
big
projects
for
us
and
it
kind
of
fits
some
of
this
architecture.
So
if
you
don't
need
to
do
it
locally,
getpod
is
an
option.
E
So
I
would
say
that
here
was
the
contributor
who
who
came
with
the
solution.
If
somebody
else
says
we
should
use
this
other
thing.
Instead,
I
don't
think
it's
fair
on
that
contributor
to
say:
hey.
Can
you
do
this
again
with
another
tool
either
we
should
have
give.
Somebody
else
should
give
another
option,
or
we
should
accept
this
contribution
in
some
form
or
another.
Then,
of
course
taking
into
account
frederick's
question
of.
Can
we
make
this
smaller
or
or
somehow
you
know
reduce
the
amount
of.
E
G
You
know
me
I'll:
throw
out,
do
it
with
conda
works
on
windows,
doesn't
need
admin
doesn't
need,
a
virtual
machine
doesn't
need
docker.
I
don't
think
putting
getting
a
working
virtual
machine
system
or
docker
in
front
of
someone
on
windows
that
shows
up
as
a
workshop
is,
is
making
anything
better.
A
G
G
G
We
would
need
an
actual
package
of
all
of
our
offline
build
chain,
and
that
would
be
good,
but
yeah
I
mean
I
I
I
have
empirical
data
that
I
was
able
to
get
an
environment
that
complicated
plus
the
full
browser
environment
plus
an
automated
acceptance,
testing
environment
and
got
it
working
on.
You
know
two
versions
of
windows,
four
distros
of
of
linux
and
two
different
mac
versions
from
usb
sticks
in
a
workshop
setting
in
under
20
minutes.
G
The
shelf
life
on
that
was
worked
for
that
work
show
right.
So
it's
it's!
How?
What
is
the
shelf
life
on
the
jupiter
lab
desktop
because
it
uses
exactly
the
same,
build
mechanism
for
that
right?
It's
a
construct.yaml
that
describes
your
requirements,
it's
just
like
an
environment.yaml
and
it
just
poops
out
installers.
At
the
other
end.
G
It's
writing
software's.
Writing
software.
So
yeah
I
mean
you.
Would
you
would
get
this
installer
you
would
on?
You
would
run
it
and
then,
on
your
you
know:
zero
privileges,
windows,
computer
would
have
git
and
the
browser
binaries
that
you
needed
and
node.js
of
a
specific
version
that
actually
worked
that
wasn't
you
know
crazy
from
the
future
or
old
and
broken.
G
G
So
if
the
challenging
thing
is
getting
a
working
outer
environment
that
you
are
running
stuff
into,
because
I
don't
even
think
we
have
like
instructions
on
how
to
do
it
with
virtual
invert
pie
and
verb,
pippin
or
poetry,
or
you
know
the
n
thousand
things
right,
yeah,
I'm
just
saying
it's
an
option.
We
have
lots
of
experience,
doing
it
and
we've
helped
other
people
do
that
for
specifically
for
the
use
case
of
onboarding
people
quickly.
In
a
workshop
setting.
A
So
I
think,
hey,
we
could
probably
talk
about
this
a
long
time,
and
maybe
we
should
time
box
it,
but
I
think
a
solution
that
doesn't
involve
a
large
download
like
a
git
pod
solution
might
be
really
good,
although
it
does
require
good
internet
to
use
it
persistently,
because
another
issue
we
often
run
into
is
everyone
downloading
all
of
the
jupiter
lab
dependencies?
A
H
H
Yeah,
so
the
question
is
I
you
know
I
do
this.
I
I
do
this
all
the
time.
On
my
own
laptop,
I
mean
I
could
con
I'll,
take
a
look
and
see
what
my
what
my
docker
file
that
I
use
looks
like
and
put
it
back,
but
frankly
this
should
be.
This
should
be
more.
H
F
I
know:
can
your
environment
download
and
build
jupiter
lab.
H
H
H
Yeah
so
yeah
putting
together
a
docker
container
that
that
builds
jupiter
lab
would
be
or
ideally
has
it
pre-built
right.
Well,
I
guess
you,
I
guess
you
do
need
to.
If
you
people
are
doing
development
on
core,
they
need
to
rebuild
this
stuff
right.
Yeah,
usually.
H
Should
you
I
mean
one
of
the
one
of
the
issues
that
that
I
have
in
doing
builds
in
when
I'm
working
on
my
extension?
Is
that
is
you
get
packaged?
Is
you
get
package
changes?
Should
we
should
this
thing
have
package
lock?
Where
you
know
I
we
guarantee
this
thing
will
build.
You
know
the
libraries
may
not
be
the
most,
and
packages
may
not
be
the
most
up-to-date
version,
but
on
the
other
hand,
they
they
do
work
yeah.
No,
I.
B
H
That's
fine,
okay!
I
will
contribute
the
I'll
disentangle,
the
stuff
we
put
it,
and
where
should
I
put
the
docker
file
for
people
to
look
at.
B
H
C
A
C
So
it's
a
it's
a
big
change,
so
we
are
not
going
to
release
enough
of
just
right
now.
The
idea
is,
if
people
that
are
contributors
could
try
already
master
branch
even
on
binder
and
try
to
report
as
much
as
possible.
I've
linked
the
issue,
that's
a
listing
for
things
that
are
known
right
now,
so
feel
free
to
to
add
some
observation
that
you
are.
You
are
seeing
and
stuff
like
that,
and
so
we
can
start
back.
C
Fixing
things
and
the
follow-up
also
on
codemirror6
is
that
it
was
planned
to
drop
a
code
editor,
so
the
phase
8,
but
that
probably
can
move
to
discussion
because
that's
going
to
require
some
discussion.
As
I
just
mentioned
there.
There
is,
for
example,
an
aspect
that
codemirror6
is
no
more
using
positioning
with
line
and
columns
is
using
offset
directly
in
absolute
in
the
in
a
string
and
stuff
like
that.
So
that
may
require
a
bit
more
discussion
for
what
what
do
we
do?
A
This
is
a
huge
amount
of
work.
We've
talked
about
a
few
times,
but
thank
you,
that's
awesome
and
everyone
else
who
is
even
marginally
curious.
Please
take
a
look
because
it's
a
deep
one,
so
we
should
find
out
all
the
things
that
are
different
need,
addressing
any
any
any
any
sort
of
edge
cases
that
may
have
been
missed
et.
A
One
aspect
of
this
is
performance.
One
aspect
of
this
is
accessibility.
A
F
F
Like
just
like
a
warning
sign,
I
don't
see
anything
so
I'm
going
to
assume
that's
never
going
to
work
all
right.
So
basically,
the
issue
is
that
when
a
document
or
notebook
is
read
only
it's
not
necessarily
immediately
obvious.
So
if
you
try
to
save
you
know
using
the
keyboard
command
sort
of
nothing
happens,
since
you
know,
maybe
that's
what's
going
on.
So
there
was
a.
There
was
a
request
from
a
long
while
back
to
make
the
we've
only
status.
F
The
document
cleared
so
taking
an
initial
stab
at
that,
based
on
some
of
isabella's
old
old
ui
designs,
so
the
toolbar
has
changed
a
little
bit
since,
like
three
years
ago,
when
isabella
wrote
up
those
designs
started
to
tweak
some
things.
F
So
I
think
what
I'm
going
to
ultimately
do
is
make
the
save
like
the
save
button
in
the
toolbar,
which
now
gets
just
grayed
out
when
documents
read
only
give
it
like
a
more
noticeable
background
and,
like
a
clear
label
same
like
document
is
read
only
or
something
make
it.
You
know
super
obvious.
F
What's
going
on
and
as
a
follow-up
that
work,
I
also
want
to
add
a
system
for
toast
notifications
of
some
kind
to
jupiter
live
so
that
like
when
somebody
tries
to
save
it,
can
just
pop
up
the
toast,
which
can
then
kind
of
you
know
unobtrusively
fade
away,
not
not
like
you
know,
neighbor's
way,
I
think
the
the
best
way
to
do
this
is
probably
to
use
some
complicated
components,
probably
best
use
some
of
the
existing
implementation.
F
There's
many
good
ones
out
there,
so
I
just
wanted
to
get.
I
guess
everyone's
feedback.
You
know
maybe
yay
support
on
using
one
of
these
packages.
I
I
am
currently
using
the
material
ui
snack
bar
in
alaira
in
a
few
locations.
I'm
a
big
fan
of
it.
I'm.
I
Yeah,
I'm
I've
been
planning
on
doing
exactly
what
you're
talking
about
like
adding
a
generic
way
of
doing
toast
notifications
in
jupiter
lab
and
just
haven't,
had
the
bandwidth
to
do
it
for
like
a
year.
Well,
somebody.
F
The
groups
talked
a
lot
back
and
forth
over
a
long
while
about
you
know,
should
we
add
material
garage
we
had
was
the
other
one
blueprints,
and
so
I
think
it's
reasonable
to
add
the
interior
toast
here,
because
at
least
at
this
point
theoretically
tree
shaking
has
gotten
good
and
consistent
enough
that
it
should
only
add
the
instead
of
the
whole
500k
of
the
ui
package
should
just
add
the
13k
of
the
the
snack
bar
package
sub
package.
J
Sorry,
two
things
in
terms
of
material
ui,
the
jupiter
lab
council
a
few
months
ago,
took
a
boat
to
adopt
fast
ui
for
our
leaf
react
and
web
components.
A
lot
of
that
was
driven
by
the
accessibility.
J
F
If
there's
a
missing
component
from
fast,
if
there's
a
toast
component
in
fast,
certainly,
is
that
what
I
said.
J
F
This
is,
this
is
also
an
issue
that
has
been
proposed.
I
think
so
that
the
issue
number
on
the
original
issue
about
toasts
is
below
1000,
so
it
was
proposed,
I
think,
eight
years
ago
or
something
so.
This
is
an
issue.
That's
been
brought
up
many
many
times.
I
think
in
the
community
and
sort
of
gets
stalled
on
like
super
easily,
so
I
think
the
the
closest
we
can
use
to
just
using
a
ready-made
solution.
Probably
the
more
realistic
like
anything
done,
will
be.
J
One
is
what
ui
component
should
we
use
for
the
the
toast
part.
The
other
is
the
broader
question
of
how
we
use
that
in
a
notification
system
and
a
couple
of
people
darien
here
andre,
the
jupiter
server
group
has
been
working
on
a
broader
effort
around
notifications
in
jupiter
lab
and
so
max.
I
I
would
encourage
you
to
plug
into
that.
There's
a
lot.
There's
been
a
lot
going
on
and
there's
other
people
that
also
have
funding
to
work
on
that.
So
please,
please
coordinate
sure.
C
I
just
want
to
mention
for
the
read-only
features.
You
are
saying
that
you
plan
to
tune
the
the
save
button
style.
There
are
some
questions
about
them.
C
No,
no!
It's
just
to
mention
that
in
the
frame
of
rtc,
that
button
may
disappear
in
the
sense
that
the
saving
would
be
automatic
automatically
down
on
the
server.
So
the
save
button
by
itself
may
not
make
sense.
Yeah.
F
It
has
rtc
tackled
read
only
at
this
point.
C
No,
no,
I
mean
it
won't
tackle
the
read
only
by
a
sense.
I
think
it's
still
neat,
I
mean
that's,
that's
just
to
bring
like
to
bring
the
discussion
about
that.
F
B
Yeah
ms
onenote,
for
example,
maps
control
s
or
command
s
to
sync,
so
even
if
things
are
auto
sa
are
normally
auto
saved
periodically.
If
you
hit
the
save
button
you
like
force
it,
which
I
personally
appreciate.
K
It's
like
there's
an
auto
save
configuration
option
that
the
user
can
configure
and
it
only
says
to
disk
when
you
explicitly
save
or
you
hit
the
auto
save
timeout
or
you
just
like
close
the
tab.
Then
it
also
says
to
disk
and
and
what's
your
default,
auto
save
is
30
seconds.
Something
like
that.
Okay,.
I
K
In
sharp
contrast,
if
you
look
at
vs
code,
it
tries,
or
at
least
code
server.
I
think
it
tries
to
save
to
disk
as
quickly
as
possible
and
very
frequently
so
it
has
a
different
model,
but
that's
vs
code
without
real
time.
Sync,
I
don't
know
what
it's
like
with
real
time.
Sync.
A
No,
I'm
sorry,
I
just
because
it
might
get
lost
in
the
chat.
I
also
so
so.
The
extension
max
that
brian
mentioned
the
the
notifications
one
it's
currently
in
the
jupiter
cal
poly
orc,
but
what
it's
using
is
the
jupiter
lab
contrib
jupiter
lab
notifications.
E
Okay,
yeah.
I
also
put
a
link
to
that
on
the
issue.
B
F
I
will
keep
looking
into
and
prototyping
toasts
trying
to
coordinate
with
everyone
else.
While
I
do
so.
A
Cool,
so
one
one
thing
to
consider
here
is
you
know
that
there's
there's
been
a
a
few
times
now,
where
we
have
actually
put
things
to
a
vote,
and
this
is
new,
so
it's
it's
not
actually
familiar.
I
think
for
any
of
us
yet
how
to
navigate
around
that,
like
there's
a
there's,
a
vote
that
closed
and
then
discussion
continued
on
that
vote
anyway.
A
Recently
right
and-
and
this
is
another
one
of
those
cases
where
we
had
a
lot
of
discussion
about
fast
ui
versus
other
toolkits,
and
we
eventually
held
a
vote
and
arrived
at
a
decision.
It's.
A
Well,
the
decision
was
to
build
a
ui
toolkit,
and
the
underlying
technology
of
the
ui
toolkit
would
be
fast
ui.
So,
okay,
I'm
not
exactly
sure
like
what
the
dynamics
that
that
around
that
sort
of
decision
ought
to
be
and
when
somebody
in
good
faith
produces
like
some
cool
new
functionality,
but
it's,
but
it's
introducing
a
new
dependency
or
different
dependency
than
one
we've
sort
of
discussed
and
voted
on.
I'm
not
exactly
sure
how
to
handle
it.
A
But
in
this
particular
case,
because
we're
only
talking
about
a
specific
widget,
let's
find
out
if
it's
there,
and
if
it's
not,
let's
find
out
how
hard
it
is
to
add
it
with
that
paradigm.
I
would
that's
so
I
I.
F
A
F
Sure
I
can
see
that
too
so
yeah
I'd
like
to
just
look
into
it
and
see
you
know,
cool
hey
what
products
has
already
been
made
on
those
those
components
you
pointed
out
and
you
know
b
what
what
can
what
exists
in
fast
ui
that
would
be
compatible.
J
And
I
do,
I
do
think
in
the
the
decision
to
adopt
fast,
there's
still,
there's
still
a
lot
of
details
to
work
out
right
and
yeah.
I
think
there's
there's
a
number
of
different
ways
that
we
can
leverage
fast
from
using
the
built-in
components
to
building
our
own
components
with
the
underlying
framework,
so
that
I
don't
think
we
work
through
all
the
details,
or
at
least
I'm
not
aware
of
the
if
we
work
through
all
those
details.
Yet
I.
F
Mean
one
one
approach
of
course
would
be
just
to
have.
You
know
shims
for
specific
libraries
so
that
they
present
interfaces
that
are
similar
to
whatever
you
know.
The
built-in
ui
library
presents.
F
Cool
it's
it's
like
fairly
lightweight.
It's
like
these
cute
little
rectangles
in
the
corner.
There's
a
optional
plug-in
to
stack
them.
If
you
really
want
to
be
able
to
receive
multiple.
F
C
B
J
Another
another
dimension
of
this
that's
coming
up
in
the
notification
work
is
that
when
you
add
real-time
collaboration,
notifications
become
much
more
complex
than
we
had
ever
imagined
and
so.
J
So
a
great
example
of
this
is
if
the
server
does
something
while
you're
well,
you
have
the
ui
closed.
C
F
J
And
a
lot
of
this
actually
is
the
event
bus,
not
rtc,
that
that's
been
designed
with
this
and
a
couple
other
use
cases
in
mind,
and
so
I
I
realized
that
there
is.
There
is
a
world
where
toast
notifications
are
simple.
I
think
jupiter
lab
is
is
no
longer
in
that
space.
E
So
I
I
think
what
you're
talking
about
brian
is
different
things
that
can
produce
notifications
right,
but
which
is
a
separate
thing
that,
from
a
extension
or
token
that
produces
a
way
of
displaying
notifications
right.
B
J
The
other
part
of
this
that
we
have
been
working
a
lot
on
is
you
know
ensuring
that
the
user
experience
overall
is
good,
so,
for
example,
having
some
notion
of
notification
types
that
a
user
can
dismiss.
So
don't
show
me
these
notifications
again
again,
like
I
there's
a
lot
of
outstanding
questions
here.
I
just
want
to
make
sure
we're
all
people
who
want
to
work
on
this
work
together
on
this.
That's
that's
really
amazing.
J
E
A
Not
really
because,
right
now,
we've
been
so
focused
on
the
server
side
of
things
of
building
what
the
event
bus
is
and
the
jupiter
lab
api
connectivity
side
that
we
haven't
actually
focused
or
refocused
on
the
notifications
ui.
Yet
so
there
isn't
a
central
repository
for
a
conversation
about
something
like
this
right
now.
Oh.
F
I
Quick
quick
clarifying
question
just
for
my
own
sanity,
my
understanding
of
what
max
was
proposing
was
solely
ui,
essentially
to
give
us
a
way
to
trigger
toast
notifications,
what
figures
them
and
how
those
get
triggered
can
be
vastly
varied
and
could
be
leveraged
by
this
new
notification
system.
But
my
understanding
was
the
goal
was
just
to
add
the
ability
to
pop
up
toast
notifications
in
jupiter
lab
natively
and
have
them
look
good.
I
So
so
the
idea
is
even
no
matter
how
these
are
used
in
the
future
by
this
notification
system.
Just
adding
the
ability
to
cleanly
visualize
toast
notifications
in
jupiter
lab
is
just
a
really
good
ad
like
we,
we
can
discuss
how
we
want
it
to
plug
into
the
notification
system
after
it's
been
implemented.
A
Go
ahead
thanks
brian,
so
so
these
words
are
ambiguous
in
this
context,
but
from
the
way
we've
been
talking
about
it,
as
as,
as
the
implementers
of
some
of
these
things,
we
we
and
we,
we
separate
events
which
are
server
things
that
happen,
occurrences
right
that
you
lots
of
stuff
might
happen.
A
You
might
be
subscribing
to
events
as
like,
a
a
cli
application
or
whatever
you
not
even
have
a
visual
component
versus
notifications,
which
are
things
end,
users
see
and
manifest
in
toast,
pop-ups
and
manifest
in
a
notification
center,
which
is
a
a
sidebar
panel,
and
so
the
jupiter
lab
notifications
extension
that
the
cal
poly
team
built
last
year
does
have
a
toast
notifications
and
it
uses
this
library
that
I
linked
above
that
comes
from
jupiter
lab
contrib
and
I
believe,
fred
authored.
A
Is
an
actual
toast
ui
in
that
extension
and
there's
a
toast
library
that
it
uses
that
it's
specifically
intended
for
being
in?
I.
F
Mean
I'm
fine,
I'm
fine
with
just
cleaning
it
up
and
breaking
it
in.
A
J
And
they
do
so.
I
agree
there
is
a
bit
of
separation
of
concerns
there.
I
think
intersection
point
is
the
the
data
and
event
model
that
the
ui
consumes
and
then
there's
a
question
of.
Is
there
a
layer
between
the
posting
of
the
notification
and
the
showing
of
the
notification
that
filters
like
if
the
users
dismiss
those
don't
show
those,
even
though
that
extension's
trying
to
post
it.
F
A
B
A
So
we
had
the
save
conversation
as
well,
but
is
there
some
piece
of
it
that
was
left
unsaid
or
should
we
move
up.
A
Okay,
so
I
opened
an
issue
in
the
team
compass
and
it's
linked
I'll
put
in
the
chat
as
well,
and
it
was
about
how
long
the
voting
period
should
be
for
the
times
we
actually
do
vote.
So
I
think
we've
had
three
votes
so
far,
not
counting
nominations
to
the
jupiter
lab
council
and
the
default
language
of
the
decision-making
guide
and
the
jupiter
jupiter
governance.
A
Repo
is
that
votes
take
seven
days,
and
this
is
a
question
of
asking-
is
seven
days
sufficient
and
I
personally
don't
actually
have
a
strong
inclination
here,
but
I
know
that
there
are
people
who
ask
for
it
to
be
longer
and
there
are
people
who
believe
actually
70s
is
good
because
otherwise
we
might
get
bogged
down.
So
if
you
have
an
opinion-
and
you
want
to
express
it
now-
that's
cool
if
you
want
to
express
it
in
the
issue-
that's
also
cool.
A
If
it
turns
out
that
we
want
to
do
something
different
than
what
is
currently
happening,
then
we
can
see
if
we
have
consensus
on
that
and
just
do
it,
and
if
we
don't,
then
we
can
see
if
we
should
put
it
to
a
vote
now
there
is
one
option:
that's
outside
of
just
jupiter
lab,
which
is
modifying
the
jupiter
wide
decision
making
guide.
A
I
personally
am
kind
of
disinclined
to
do
that,
because
different
projects
are
different
and
they
all
sort
of
either
will
be
satisfied
with
the
default
seven
day
period
or
will
have
to
have
their
own
conversation,
but
because
I
don't
really
have
a
position
to
advocate
for
I'll
give
a
second
for
others
to
chime
in
or
we
can
move
on.
I
I
could
try
to
verbalize
something
I've
been
thinking
about,
so
there's
so
I've
kind
of
mixed
on
this
based
on
a
lot
of
people's
comments
in
that
pr,
I
feel
like
there's,
really
solid
reasons
born
against
it.
I
agree.
I
don't
think
we
should
be
changing
it
jupiter
wide.
I
think,
if
we
do
do
anything,
it
would
be
jupiter
lab.
I
Only
seven
days
should
normally
be
enough,
but
there's
always
going
to
be
circumstances
which
could
change
that
and
I
feel
like
one
of
the
biggest
ones
that
we've
run
into
is
discussion
where
we're
where
it's
brought
to
a
vote.
J
B
J
The
way
the
governance
or
the
decision-making
guide
is
written
is
that
decisions
are
made
through
two
two
phases
where
the
first
page
is
required.
The
second
is
optional,
the
first
phase
of
just
consensus,
and
if
some
project
can
reach
consensus,
no
bulk
are
needed.
A
vote
is
typically
called
when
consensus
blocks.
J
J
The
guy
explicitly
says,
votes
should
not
be
taken.
J
E
Yeah
so,
as
far
as
I
understand
things
in
jupiter
lab
here,
the
times
where
you
get
a
vote,
as
ryan
said,
is
when
consensus
breaks
down
right.
So
if
that
that
happens,
it
means
it's.
Something
contentious
is
going
on
right
this,
and
ideally
this
should
be
rare
occurrences.
E
Well,
hopefully,
there
will
be
rare
occurrences
right,
so
if
we
only
need
to
make
a
decision
like
a
quick
decision
for
you
know,
should
we
bring
in
this
other
repo
into
the
org
right?
That
happens
from
time
to
time
on
the
team
compass
that
those
things
have
happened,
those
would
normally
go
by
concept.
Consensus
right,
you
won't
actually
have
to
have
a
formal
vote
on
it,
so
those
those
will
still
be
able
to
be
short
and
quick
and
whatever-
and
you
won't
have
a
huge
thing
there.
E
E
It's
maybe
those
things
can
happen,
yeah.
If
somebody
calls
a
vote,
nobody
should
call
a
vote
in
20th
of
this
december
right.
F
E
A
Well,
no,
the
rules
are
there
to
cap
how
long
the
vote
takes,
but
the
whole
thing
is
predicated
on
people
acting
in
good
faith
and
not
waiting
until
no
one's
looking
to
sneak
something
in
like
the
the
legitimacy
comes
because
we
trust
each
other,
not
because
we
voted
right
all
right
well,
because
we
only
have
five
minutes
left,
and
there
is
another
thing
I
would
suggest
carrying
on
the
conversation
in
the
thread
who
wanted
to
make
the
report
from
scipy,
and
should
I
stop
the
recording
before
you
do
it
or
after.