►
From YouTube: Kubernetes SIG CLI 20201104
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Okay,
good
morning,
good
evening,
good
afternoon,
depending
on
where
you
are
today
is
november
4th-
and
this
is
another
of
our
six
cli
bi-weekly
meetings-
my
name
is
mati
and
I'll,
be
your
host
today,
and
the
agenda
is
literally
packed
with
lots
of
good
stuff.
A
So
let
me
quickly
start
with
with
a
couple
announcements:
we
have
a
blocker
bug
happening,
the
exact
same
time
next
wednesday.
On
november
11th,
we
have
a
kubernetes
120
code
freeze,
which
is
happening
the
next
thursday
november
12th.
So
if
you
care
about
your
code
being
reviewed
or
merged
I'll,
make
sure
to
do
it
before
that
november
12th,
the
next
week,
we
will
be
holding
virtual
kubecon
north
america,
that
is
happening
from
tuesday
november
17th
until
20..
I
linked
the
agenda.
A
If
you're
interested
we
as
a
sixth
cli,
we
have
our
own
session,
we've
managed
to
put
four
people
on
that
short
20
20,
something
slot
you'll
hear
from
phil
from
sean
from
eddie
and
and
myself
so
be
sure
to
check
it
out
as
well.
A
A
Please
fill
it
out.
We
would
like
to
hear
from
you
from
people
that
are
using
if
you
have
other
folks,
please
socialize
the
the
survey
link
as
well.
It
will
be
very
useful
for
us
to
gather
some
feedback
about
that
command
so
that
we
can
make
it
even
better
than
it
is
yeah
lee
also
sent
it
out
to
6
cli
mailing
list
and
cube
def,
so
that
should
reach
sufficient
mass.
I
hope
ok,
so
moving
on
to
the
main
topics,
unless
phil,
eddie
or
sean
have
anything
to
add.
B
A
A
Okay,
so
the
main
the
first
topic
is:
keep
cuddle
to
staging.
So
with
great
help
from
jan
recently,
we've
managed
to
move
cube
cattle
to
staging
almost
entirely
there's
a
1pr,
that's
96
190.
That
is
still
weighting
approvals.
I
hope
it
will
be
green.
There
was
one
problem
with
flax,
the
cube
cattle
convert,
plugin
that
we
created,
but
I'm
hoping
that
that
should
be
solved.
A
The
auth
shared
code,
that
was
the
biggest
issue,
was
moved
by
jan
houka
over
to
the
recently
created
repo,
which
is
called
kate's
that
I
owe
competent,
helpers
and
jan
also
right
after
that
one
merge.
He
moved
the
auth
code
to
staging,
so
the
only
one
remaining
fit
is
moving
the
rest
of
keep
cuddle
to
staging
and
converting
the
current
convert
command
to
a
plugin
and-
and
that's
what's
happening
in
the
in
the
in
the
final
pr.
A
What
is
important
to
notice
is
that
for
120
we
will
continue
to
build
cube
cuddle
from
the
main
kubernetes
repo,
and
the
only
reason
for
that
is
that
at
this
point
in
time,
we
don't
want
to
complicate
or
change
the
release
process,
because
that
wasn't
agreed
yet
we're
hoping
to
start
the
discussion
with
the
sick
release
and
sick
testing,
because
there
will
be
a
lot
of.
There
are
still
a
lot
of
questions
open
about
how
to
properly
test
cube
cuddle.
A
A
So
we
will
be
looking
to
answer
those
questions
in
the
in
the
next
releases
and
then
once
we
have
all
those
answers
in
place,
we
will
be
able
to
cut
the
court
finally
and
move
the
entire
development
to
its
own
separate
repository.
A
But
the
fact
that
we
are
already
living
in
staging
ensures
that
no
of
no
dependency
on
the
main
kubernetes
kubernetes
code
will
be
introduced.
So
that's
a
big
one,
and
here
I
would
like
to
thank
every
single
person
that
either
shared
their
ideas,
helped
reviewing
or
made
actual
changes
to
the
code
to
make
this
happen,
and
over
the
past
I
I
think,
almost
three
years
there
was
a
lot
of
people
involved
in
in
that
work.
A
I
could
probably
start
naming
them,
but
that's
there's
so
many
of
them
that
I'm
I'm
worried
that
I
might
forget
about
some
folks.
So,
if
you're
interested,
I
would
probably
encourage
you
to
look
at
the
get
history
and
see
who
was
who
whom
to
blame
for
this.
So
thanks
a
lot
to
everyone
involved
in
that
effort.
A
A
Okay
hearing
none,
so
I
guess
eddie
and
you
will
probably
want
to
share
the
screen.
Let
me
quickly
stop
sharing
and
make
a
co-host.
C
So
this
kind
of
ties
in
to
what
you
just
talked
about
a
tiny
bit.
Do
we
have
plans
to?
I
guess
we
need
to
talk
to
nikita
in
the
github
team,
but
do
we
want
future
cube?
Ctl
issues
opened
on
the
cube,
ctl
repo,
and
is
there
a
way
we
can
kind
of
force
or
automate
that.
A
My
guess
is
that
we
could
probably
improve
the
github
template
for
bugs
to
point
people
if
they
are
opening
cube,
call
specific
bugs
they
should
be
opened
against
kubernetes
cube,
cuddle
repo
instead
of
the
main
kk.
A
That's
one
possibility,
I
don't,
I
can't
think
of
anything
else
yet
so
definitely,
as
you
said,
probably
syncing
with
nikita
and
the
rest
of
the
remind
me.
What's
the
sig
name
that
I'm
thinking-
and
I
can't
yes
controvex
exactly.
D
We
also
need
to
kind
of
update
the
readme
as
well.
A
Yes,
that's
that
that's
definitely
a
good
point.
There's
a
lot
of.
D
D
C
Okay,
so
there
was
mache
hit
me
up
was
the
last
week
of
the
week
before
about.
We
have
several
issues
that
are
opened
around
like
specific
things
right.
So
if
you
take
like
you
control
weight,
for
example,
I
think
there's
like
two
or
three
issues
that
are
very
specifically
tied
to
that.
We
don't
have
a
good
way
right
now
of
tracking
any
of
that,
and
so
we
were
talking
and
then
I
pinged
laurie
and
pulled
lori
in
and
there's
a
conversation
in
the
the
six
year
channel.
C
If
you
want
to
check
that
out,
but
the
idea
seems
to
be,
if
we
start
with
a
project
board
and
then
we
can
maybe
expand
and
explore
using
some
further
labels
like
area
cube
control
weight,
I'm
pretty
sure
these
are
arbitrary
and
we
don't
need
the
pr
anything
to
add
them
in
I'm
pretty
sure
we
can
just
do
a
slash
pro
command
for
them,
but
I
took
a
shot
at
just
throwing
together
a
quick
board
and
digging
into
kind
of
what
the
permissions
are,
and
so
this
is
project
board
47
under
the
kubernetes
org.
C
It's
not
under
any
specific
repo
and
anyone
who's.
An
org
member
should
have
full
right
access
to
this,
so
you
shouldn't
need
to
have
any
special
permissions
to
be
on
any
special
teams,
which
is
different
actually
from
repo
specific
projects.
You
need
to
actually
have
specific
permissions
and
be
on
a
certain
team
to
actually
control
that.
So
this
is
an
org
level
board,
and
this
lets
us
pull
in
issues
from
different
projects
under
the
kubernetes
org
and
the
automation
that's
built
in
here
is
pretty
rudimentary.
C
I
was
hoping
it'd
be
a
little
bit
better.
I
am
going
to
talk
to
nikita
about
what
we
can
do
there,
but
it's
pretty
much.
We
manually
have
to
add
cards
right
now
and
it
looks
like
we
can't
even
add
them
via
the
issue
itself.
This
might
actually
be
a
special
permission
that
needs
to
be
added
somewhere,
but
right
now,
once
this
is
added
to
a
card,
I
can
come
in
here
and
move
this,
but
there's
actually
no
way
to
add
it
to
a
project
from
this
repo
view.
C
So
you
actually
have
to
come
into
the
the
project
here,
search
for
the
card
via
the
issue
number,
and
then
you
can
add
it
to
the
project
here,
and
so
I
get.
This
is
just
like
a
rough
shot.
I
mean
we
can
change
this.
We
don't
have
to
use
this
either
and
we
can
modify
the
columns
but
more
just
like
presenting
a
proof
of
concept,
so
we
can
start
a
discussion
around.
How
do
we
want
to
track
these
things,
and
would
this
be
beneficial?
C
A
So
my
original
idea,
when
I
hit
up
eddie
with
this,
was
that
we
currently
have-
and
that
was
I
think
that
was
somehow
related
to
one
of
our
previous
bug.
Scrubs,
where
again,
we
had
a
specific
issue
about
well
cube.
Cuddle
weight
was
a
good
example,
because
I
think
it
it
was
brought
up
along
those
lines
and
basically
currently
we
don't
have
any
good
way
to
say.
Oh,
these
are
the
open,
cupcake
weight
issues.
A
You
literally
have
to
search
for
all
cube,
cuddle
weight
issues.
We
don't
group
them
in
any
way.
A
I
don't
think
the
current
labeling
system
allows
for
this
kind
of
grouping,
so
I
figure
out
that,
having
not
sure
if
one
or
probably
several
of
project
boards
would
allow
us
to
somehow
organize
our
thoughts
and
issues
opened
and
divide
them
by
categories
or
by.
A
C
E
So
so
one
of
the
ways
that
we
connect
issues
or
pr's
that
are
similar
is
we
just
throw
links
from
one
end
to
the
other,
which
that
this
is
much
cleaner
and
and
better
to
organize
it,
but
yeah
that
that
kind
of
more
awkward
waves,
some
how
we
normally
kept
track
of
things
that
were
were
related.
A
Yeah
eddie
and
I
so
if
you,
if
you
work
on
kubernetes
stuff,
there
are
different
options
and
cross-linking
is
one,
but
that's:
okay,
if
you're
closing,
because
that's
a
duplicate
of
something
else,
if
you're
tracking
several
issues
together
it
it's
getting
more
and
more
complicated
and
having
a
summary,
so
I've
noticed
a
lot.
Oftentimes
people
will
create
umbrella
issues
and
they
will
keep
on
updating
the
descript.
The
initial
description
of
the
issue.
A
But
then
again
over
time,
the
description
might
get
outdated
and
this
again
falls
out
of
any
reasonable
tracking,
and
you
end
up
with
yet
another
issue
about
this
particular
topic
that
people
cross
link
pretty
much
everywhere.
So
that's
why
we
figure
out
that
having
some
kind
of
a
board
to
track.
This
would
be
a
good
approach.
C
C
Okay,
I'll
spend
some
more
time
playing
in
here
and
see.
If
I
can
figure
out
the,
I
wish
that
there
was
a
way
for
us
to
automatically
get
issues
added
that
were
labeled
via
something
it
doesn't
look
like
that's
the
case
right
now,
so
I'm
gonna
go
talk
to
the
the
github
management
team,
see
if
they
have
any
advice.
C
I
think
that's
the
difference
with
project
repos
for
yeah
a
project
under
repos.
You
can
set
that
up
automatically,
but
I
don't
know
I'll
have
to
test
it.
A
And
we
can
probably
go
through
open
issues,
since
we
are
reasonably
up
to
date
with
issues
thanks
to
our
bug,
scrubs,
that
eddie
is
organizing
once
a
month.
We
could-
probably
just
you
know,
revisit
the
ones
that
we
already
visited
and
shuffle
them
into
appropriate
projects
once
we
set
it
up.
C
Okay,
I
will,
I
will
keep
trucking
and
we'll
see
where
we
go
with
this,
I'm
going
to
talk
to
the
github
team.
First
cool
next
topic
was
very
similar,
so
we've
been
using
this
called
triage
party.
C
What
is
it
there?
We
go
so
if
you've
been
to
a
bug
scrub
before
we've
walked
through
this,
it's
used
in
a
couple
other
kubernetes
cigs
right
now,
sig
release
is
actually
using
it
a
ton.
C
They
have
a
really
nice
config
that
we
rolled
out
so
if
you've
been
to
bugsub
you've
seen
this
before
this
basically
is
a
github
issue
and
pull
request
aggregator
that
we
can
write
a
bunch
of
fancy,
yaml
queries
for,
and
everything
in
here
is
configurable
in
terms
of
sections,
I'm
just
using
the
default
config
right
now,
and
so
I
guess
the
the
reason
I
brought
this
up
on
the
agenda
is
we
need
to
figure
out
if
a
this
is
going
to
be
useful
to
anyone
else
outside
of
bug,
scrub
and
b,
what
kind
of
shape
or
architecture
do
we
want
to
make
this
config
look
like
and
there's?
C
Actually
I
had
a
really
nice
dock
from
sig
release
here.
It
is
sig
release
and
in
laurie,
specifically
lori
is
just
an
amazing
pm
in
general,
so
lori
put
together
this
very
awesome
I'll
I'll
share
this
in
the
doc,
but
this
is
their
plan
for
creating
sections
and
prioritizing
things
and
figuring
out
what
goes
where-
and
this
is
just
hosted
under
a
personal
account
right
now
for
me,
but
we
can
actually
move
this
to
the
kubernetes
org
like
sig
release.
C
C
C
F
Hey
are
you
where's?
The
link
to
this
is
that
in
the
the
meeting
notes
right
now
yeah,
let
me
add
that
there.
F
It
looks
really
it
looks,
really
cool,
also
complex,
not
sure,
would.
C
Like
to
try
it,
it
is
awesome,
and
I
so
where
I
keep
getting
stuck,
is
I
keep
opening
the
yaml
config
file,
and
it
looks
something
very
similar
to
this,
where
you
know
it's
just
a
bunch
of
yaml
and
filters
and
queries,
and
so
every
time
I
open
this
file,
I
just
look
at
it
and
close
it
and
that's
why.
I
think
that
lori's
approach
was
a
lot
better
in
that
they
started
with.
You
know
a
high
level
overview
of
what
they
wanted
to
look
like
versus
just
jumping
into
writing.
Yaml.
F
Right
linking
both
of
those
would
be
great,
it'd,
be
good
to
have
a
I'm
assuming
there
might
be
discussion
here
that
compares
this
approach
to
like
using
the
github,
kanban
style
project
organization
tool.
C
Yeah,
so
what
I
like
is
so
sid
really
specifically
uses
both
and
they
use
the
the
board
to
track.
You
know
individual
things
via
kanban,
and
this
is
nice
for,
like
seeing
all
of
your
issues
up
front
with
labels
and
like
getting
them
out
of
filters.
So
the
section
that
we've
been
doing
in
bug
triage
is
like
unprioritized
issues
older
than
seven
days
and
that
we
usually
go
through
here
figure
out.
You
know
this
is
usually
like
the
past
week
or
two
weeks
worth
of
issues,
and
this
lets
us
quickly
like
okay.
C
C
Okay,
I
will.
I
will
bump
that
for
now.
A
Oh
eddie
dragged
me
into
a
meeting
about
the
triage
party.
I
think
that
was
last
week
or
two
weeks
ago
still
the
guys
that
put
together
that
treasure
party
they're,
pretty
cool,
so
the
project
itself
is
super
helpful
jeff.
If
you're
interested
to
have
a
closer
look
next
time
when
we
will
be
having
next
week
next
wednesday,
then
we'll
be
having
our
box
crop.
You
can
join
and
see
for
yourself
for
a
couple
of
minutes.
Yeah.
F
A
I
assumed
there
were
no
further
questions,
so
we
can
jump
over
to
the
next
topics.
Quick
time
check.
We
still
have
plenty
of
time
to
to
continue.
A
The
tyler
introducing
colored
output
feature
in
cube
cuddle
want
to
talk
about
it.
G
Oh
thank
you,
so
I
have
one
thing
that
I
want
to
talk
here
today.
So
the
context
is,
I
already
shared
the
link
in
the
google
docs
that
you
have
issue
comment.
G
G
So
I
found
this
issue
and
I
found
a
lot
of
comments
are
left
on
this
issue
so,
and
I
finally
found
currently
colorizing
output
feature
is
not
supported
by
offshore
cube
control,
so
I've
developed
a
software.
I
call
it
cube.
Color
cube
color
is
a
wrapper
tool
for
cube
control.
It
internally
calls
cube
control
and
get
the
output
and
it
shows
the
output
in
readable
colors.
G
So
I
share
the
two
in
this
issue
and
thanks
to
eddie
sorry,
if
I'm
calling
you
in
the
wrong
way,
but
thanks
so
eddie
suggests
me
to
edit
from
my
comment,
and
he
suggested
me
to
talk
about
getting
cube
color
in
upstream
in
cube
control,
and
he
suggests
me
to
talk
about
this
in
this
six
series
meeting.
G
So
I
am
joining
this
meeting
for
the
first
time
today.
So
what
I
want
to
know
is,
if
cube
control
should
officially
support
this
feature
or
not.
So
I
want
to
get
your
feedback
and
your
opinions
rightly
so
if
this
feature
looks
acceptable
and
if
this
switch
should
be
implemented
in
upstream
cube
control,
so
I
can
start
working
on
some
design
and
start
implementation.
Probably
I
will
start
with
very
minimal
and
a
small
implementation,
because
I
am
actually
still
thinking.
A
A
G
I
started
writing
design,
so
my
current
idea
is
because
I
don't
want
to
make
any
breaking
changes
for
this
feature.
G
So
we
we
don't
need
to
make
any
breaking
changes,
so
we
don't
need
to
make
users
confused
so
yeah.
That
is
what
I'm
thinking
right
now,
but
today
today
I
want
to
get
your
feedback
about
whether
we
should
implement
this
feature
in
cube
control.
Whether
cube
control
should
offshore
support
this
feature
but
yeah
what
what
guy?
What
do
you
guys
think
about
this.
A
Okay,
so
definitely
the
route
of
default
off
would
be
something
that
I
would.
A
A
I'm
asking
that,
because
I've
noticed
that
as
examples
you
have
get
pods,
which
is
using
the
server
side
printing,
which
dictates
us
how
we
should
be
printing
stuff
and
then
so.
Cubicadoget
does
not
have
control
over
what
the
output
of
those
will
look
like,
especially
that
one
of
the
reasons
for
server-side
printing
was
to
support
custom
resource
definitions
of
getting.
A
But
if
you
do
describe
a
any
kind
of
crt,
then
it's
a
problem
because
we
have
only
a
generic
describe,
which
is
not
necessarily
helpful
to
some
users.
So
I'm
curious
based
on
what
attributes
or
what
conditions
you're
picking
specific
colors.
H
G
So
if
we,
if
we
introduce
this
feature
in
cube
control,
we
should
first
decide
server
side
or
client
side
we,
which
should
apply
the
color
to
the
output,
so
yeah
that
that's
what
I'm
thinking
right
now
and
I
didn't
find-
which
is
better
right
now
so.
A
The
overall
goal
that
we
have
in
cube
cuddle
is
to
move
like,
ideally
to
have
everything
done
on
the
server,
so
cube.
Cuddle
should
have
only
minimal
knowledge
of
elements
living
on
the
server.
Ideally,
we
would
have
all
the
information
for
describe
for
get
coming
from
the
server
and
we
would
only
present
the
data
as
given
by
the
server.
That's
why
I'm
curious
how
to
properly
recognize
that
fact
and
build
on
top
of
that?
A
There
will
be
only
a
handful
of
client-side
commands
where
we'll
be
doing
something
special
and
the
only
reason
for
that
is
because
special
usually
means
it.
It
won't
work
with
with
custom
resource
definitions
and,
with
the
current
push
up
stream,
where
a
lot
of
function
that
functionality
is
built
as
an
extension
to
the
core
kubernetes.
We
need
to
ensure
that
the
core,
kubernetes
and
cube
cuddle
functionality
works
properly
with
the
extensions.
B
I
think
we
can
kind
of
break
this
up
into
two
concepts
like
one
is
conceptually
like
age,
for
instance,
will
be
present
like
like
that,
doesn't
differ
for
custom
resources
right
like
ages
should
be
age
or
like
if
it's
not
age,
it's
creation
time
or
I
forget
exactly
what
is
shown
there
same
thing
with
name
and
then
and
then
I
think
this
can
like
fold
into
the
like
standardized
status
conventions
right
so
like
whether
it's
called
name
or
conceptually
or
called
ready
in
status
or
conceptually
maps
to
ready
in
status
like
we
have.
B
We
already
have
a
notion
of
like
using
conditions
and
this
sort
of
thing
to
communicate
back
to
the
client
like
whether
something
is
or
what
state
it's
in,
whether
it
is
completed
it's
roll
out
or
not.
So
we
could.
We
could
just
like
standardize
with
some
of
those
conventions
or
at
least
map
to
there's
something
that
determines
whether,
like
it's
in
a
steady
state,
whatever
that
is
it's
green
and
then
maybe
another
thing
is
the
I'm
not
sure.
B
If
you
already
mentioned
this,
but
like
that,
the
examples
you've
shown
has
a
color
background,
that's
black,
and
so
these
show
up
pretty
well
like.
If
the
terminal
is
white,
do
we
change?
Are
there
different
colors?
Do
they
show
up
just
as
well
and
that
sort
of
stuff,
like
might
not
make
sense
for
the
server
like
just
cue,
control,
send
in
a
header
like
my
terminal
color
is
you
know
you
know
sand
or
something
right,
and
then
the
server
sends
back
like
okay?
That
means
this
thing.
G
So
about
background
color
in
current
cube
color,
it
supports
the
frag
like
dash
dash
light
background,
so
if
it
is
specified
it
will
use
the
colors
preset
for
light
background
environment
yeah.
But
now
I
think
so.
If
we,
if
we
should
not
implement
the
colorizing
feature
in
cube
control,
but
we
should
implement
it
in
server
side
yeah.
I
think
that's
I
I
found
it
very
difficult
in
implement
yeah,
so.
B
I
think
it's
more
the
right
and,
like
you,
could
imagine
imagine
users
wanting
to
swap
out
their
palettes
right.
So
I
think
it's
more
marking
up
specifically
like
these.
This
field
is
colored
this
way
because
it
is
a
field
name
right
or
this
is
like
a
priority,
is
zero.
It's
pink!
Maybe
that's,
because
it's
an
integer,
it
looks
like
color,
integers,
pink
or
something
like
that.
B
D
Not
about
the
current,
let's
say
we
have
the
mapping
and
the
standardizations.
Then
we
can
still
provide
the
colors
in
the
client-sides
in
it.
D
B
C
Yeah
before
we
before
we
keep
bike
shutting
on
technical
details
which
we'll
need
for
eventually,
I
don't
want
to
interrupt
anyone,
but
thank
you
for
coming
dee
tyler.
I
appreciate
you
joining
us
and
sharing.
I
think
we
should
just
like
take
away
step
back
and
like
do.
We
want
color
support
added
here.
Officially,
all
right,
like
this
issue
has
been
open
for
two
years.
We
haven't
closed
it
as
a
no,
yet
detailer
has
a
very
clear
interest
in
building
and
owning
this.
C
So
obviously
you
know
if
I
think
we
just
need
to
make
a
decision
to
start
like
do.
We
want
colors
in
cube
control,
and
then
we
can,
you
know,
figure
out
the
path
forward
to
figuring
out
the
actual
implementation,
everyone's
so
eager
to
talk
about
like
the
technical
details.
All
the
time.
A
I'm
trying
to
find,
if
I
remember
correctly,
there
was
in
the
past
approach
for
colorizing
cupcato
output.
A
A
Slightly
minimal
because
he
started
with
only
coloring
certain
aspects,
but
I
can't
remember
what
was
the
answer
to
that
one
which
would
help
us
drive
the
discussion,
because
right.
B
So
the
arguments
against
right
would
be
technical
debt
right,
and
is
this
a
priority
right
so
someone's
going
to
have
to
review
this
code
and
and
that
sort
of
thing
and
make
sure
it's
tested
and
do
all
the
the
normal
work
that
needs
to
be
done?
I
like
from
a
feature
perspective
I
like
it.
It
looks
pretty
and
if
it's
opt-in
perfect,
I
would
probably
use
it.
G
So
it
sounds
like
we
are
basically
positive
for
colorizing
output,
but
we
need
more
detailed
design
or
we
need
more
discussion
about
how
to
implement
that
right.
F
Hey
one
one
question:
maybe
phil
knows
the
answer
to
this
when
I
do
colorized
I'll
put
say
from
ls
and
I
pipe
it
in
more
or
grep,
I
lose
the
escape
codes
that
do
the
coloring.
So
if
that's
like
a
that's
a
big
downside,
if
coloring
only
works,
if
I
use,
if
I,
if
I
can't
grab
her
more
that
limits,
the
utility
in
some
sense.
G
B
B
Is
it
too
complex
from
the
server
side
and
therefore
we
don't
want
to
introduce
it
like
these
are
things
I
think
we
just
need
to
explore
and
the
design
itself,
so
my
vote
would
be:
let's
move
forward
with
a
cap,
knowing
that,
like
we
need
to
figure
out
the
practical,
like
there's
a
couple
practical
design
pieces
that
we
need
to
figure
out.
We
need
to
make
sure
they
don't
scope,
creep
or
create
a
bunch
of
complexity
and
can
maintain
kind
of
the
line
we've
been
taking
about
support
for
external
types.
C
So
it
sounds
like
this
is
probably
the
best
place
to
start,
and
I
guess
I'm
I'm
hoping
that
you'd
want
to
own
this
and
drive
it
through.
So
we
would
need
to
write
up
a
cap
and
submit
it
and
then
what
phil
said
we
could
dig
into
the
actual
technical
implementation.
G
I
want
to
spend
my
time
for
more
detailed
design,
get
your
feedback
again.
A
A
I
I
look,
but
I
cannot
find
the
past
if
I
find
it
I'll
I'll,
definitely
make
sure
to
link
the
previous
issue
for
coloring.
But
one
of
the
reasons
somebody
wanted
to
have
coloring
is,
for
example,
to
be
able
to
distinguish
easily
in
the
cli
when
a
part
is
not
running,
so
we
would
colorize
the
path
status
or
the
entire
row
up
to
you
based
on
the
availability
of
the
pot.
So
it
was
if
it
was
running
it
would
probably
the
way
it
is
or
green
whatever,
but
it
in
any
other
state.
A
It
would
be
different,
pending,
maybe
yellow
and
anything
else,
from
image
failures
to
failed
or
something
like
that
would
be
red
so
that
when
you're
doing
get
paused,
it
stands
out
that
there
is
a
problem
with
this
particular.
The
only
downside
and
the
problem
with
this
is
that
it
will
have
to
be
resource,
specific
and
yeah.
B
We
have
again,
I
think
we
should
fold
this
into
the
standardized
conditions
right,
because,
because
the
cli
that
we've
already
built
a
number
of
libraries
and
such
that
are
used
to
generalize
from
tli's
perspective,
this
thing
is
in
progress.
This
thing
is
stalled
and
not
going
to
complete
this
thing's,
not
in
practice,
and
I
finished
it
and
so
those
I
think
would
map
like
if
we
can
just
build
on
that
like
that
to
me
kind
of
would
be.
A
So
there
is
a
topic
that
I'm
pushing
one
of
my
team
members
to
pursue
of
standardizing
the
the
conditions
for
workloads.
A
But
since
it's
a
bigger
topic,
we
have
that
in
our
heads.
But
we
haven't
put
together
a
proposal
and
actually
start
implementing,
especially
that
it
will
require
a
significant
effort,
because
we
would
have
to
rewrite
well,
maybe
not
right.
But
we
have
to
change
all
of
the
controllers,
still
maintaining
the
old
conditions
and
introducing
the
new
ones.
In
parallel.
B
I
I
think
we
already
have
a
library
that
does
this
like
sean.
Maybe
do
you
remember?
I
mean
like
I
remember
that
I
know
I
can
see
all
the
utils
are
or
where
it
is
exactly,
but
I
believe
there's
already
a
library
that
looks
at
like
st
has
a
standard
approach
and
then
looks
at
the
built-in
types
for
like
deployments
and
staple
sets
and
then
maps
those
to
like
the
standard
conditions.
A
Okay,
so
but
you're
talking
about
a
library
and
that's
fine,
I'm
on
the
other
hand,
talking
about
doing
the
standardization
in
the
cube
itself,
which
is
which
is
somewhat
similar,
but
it
will
be
just
easier
for
users
to
consume
yeah
libraries
are
fine
as
a
intermediate
state
for
now,
so
I
don't
check
on
that.
One.
E
Just
just
a
quick
note
there
there
are
other
efforts
like
addressing
what
what
phil
said.
There
are
other
efforts
already
underway
if
we're
aware
of
them
to
standardized
conditions,
which
I
think
it's
in
sig
architecture
now,
but
anyway,
just
so.
We
know
that
that's
that
appears
to
be
underway
elsewhere.
B
C
A
C
Yeti
yeah
d
tyler
feel
free
to
ping
me,
I'm
eddie,
zane
in
slack
or
doug
lands,
I'm
sorry
doug.
I
can't
land
graph
feel
free
to
ping
me
or
doug
doug's,
actually
working
on
a
cup
right
now
that
I
think,
is
actually
pretty
important
to
this
one.
C
Last
week
we
had
talked
about
you
know
all
the
different
bits
about
xdgg
configs
and
one
of
the
things
that
was
brought
up
was
the
user
separating
user
configs
for
cubecontrol
out
into
its
own
file
like
acute
control
rc,
and
I
think
that
this
is
a
clear
need
for
something
like
that.
So
you
know
if
folks
want
to
opt
into
the
color
by
default,
they
don't
want
to
have
to
pass
pretty
true
every
time
or
do
an
alias
right
like
they
want
to
set.
C
You
know
pretty
true
into
a
config
file
somewhere,
and
it
would
just
automatically
default
that
so
I
see
you
saw
what
I
said
doug
so
yeah
we
should.
We
should
tie
these
use
cases
together
for
sure.
A
Well,
act
actually
moving
on
to
the
next
topic,
because
we
have
about
10
minutes
left
and
I
want
to
love
to
cover
the
remaining
topics.
A
Back
has
a
pr
review
and
I
think
I
owe
you
doug
review
on
that
one,
I'm
hoping
to
go
through
that
later
today,
tomorrow
at
latest,
unless
you
wanna,
unless
you
wanna
cover
something
else
from
that
one,
I'm
not
sure.
If
that
is
on
the
call
yeah,
I
see
him.
A
Yeah
sure
we'll
definitely
make
sure
to
have
it
reviewed
this
this
week
and
I'm
guessing
the
xdg
base
directory
specification.
That's
for
a
broader
review!
I'm
not
sure
if
I
want
to
talk
a
little
bit
more
about
the
proposal
itself
or
we
will
leave
it
for
first
amount
of
time.
I
think
it's
a
something
bigger
to
discuss.
A
J
A
Phil
you're
you're:
do
you
want
to
talk
about
the
new
dark
side.
B
Yeah
sure
I
just
threw
that
out
of
there,
okay,
so
it's
hosted
it's
hosted
out
of
the
cli
experimental,
is
where
the
docs
live.
This
isn't
the
domain
address,
we're
going
to
use
we're
in
the
process
of
getting
like
a
dns
like
cli.sages.kids.io,
but
and
that
will
allow
us
to
move
the
site
between
repos
or
whatever
we
want,
but
it's
living
here
for
now,
because
it
was
kind
of
the
lowest
bar
for
for
getting
it
deployed
in
a
location.
B
So
that
means
that
anyone
like
over
the
next
week
or
two
we
plan
on
trying
to
get
the
other
sites
pointing
here
we're
gonna,
get
like.
I
said,
a
new
domain
for
this
thing
and
then
try
and
get
the
control
dot
docs
pointing
here
and
then
try
and
figure
out
the
customized
site.
B
I
think
that
the
customized
content
and
control
docs
content
was
both
migrated
into
this
page,
basically
migrated
exactly
as
it
was,
and
then
sam
and
I
spent
time
rewriting
it
and
restructuring
it
and
deduping,
and
that
sort
of
thing,
but
you
shouldn't,
be
losing
a
whole
lot
here,
so
we're
gonna,
try
and
update
those
sites
to
redirect
here
and
then
this
will
be
the
new
site,
and
so
there's
like
a
section
there's
like
a
contributing
section
in
a
back
section.
B
So
that
means
like
anyone
who's
in
the
discussions
in
the
past,
where
we've
been
like,
we
need
it
back.
That
would
be
a
location
there.
That's
a
good
good
location.
A
B
Yep,
so
I
am
usually
can't
make
it
because
it's
pretty
late,
but
it's
siam,
sundar
cave
and
you
can
he's
on
the
kubernetes
slack.
So
if
you
want
to
congratulate
him,
go
send
a
message
and
say
great
job.
A
Yeah
there
is
a
quantity,
is
currently
part
of
the
beta
program
for
shout
outs
on
github.
So
you
can
definitely
use
that
in
some
of
the
repos
and
soon.
K
A
Great
okay,
we
still
have
about
five
minutes
left
before
we
move
on
to
stand-ups.
There's
one
thing
that
I
forgot
to
do,
which
is
introducing
new
folks
on
the
call,
and
I've
noticed
that
we.
F
Yeah
so
matcha,
if
you're
willing
yeah
go
ahead,
can
you
I
got
about
one
two
minutes
of
content
on
that
link?
If
you
click
on
that
guy,
because
I
don't
want
to
present
my
I'm
running
a
intel
pentium
from
1981
here,
okay,
so
the
goal,
our
goal
is
to
release
every
two
weeks,
ideally
on
wednesday,
but
in
practice
it
turns
out
to
be
we
close
out
all
the
pr's
on
wednesday
and
release
on
thursday.
F
So
the
next
release
is
going
to
be
387
on
the
12th.
We're
trying
this
kanban
process
there's
a
link
to
that.
We're
still
learning
how
to
use
it.
The
goal
is
to
get
the
headings
correct
like
what's
under,
consideration
should
be
under
consideration.
F
F
So
what
do
you
think
about
the
following
I'll
show
up
at
the
at
the
60
live
meeting
on
the
bug
scrub
on
the
11th
and
start
a
breakout
session
for
anybody?
F
Who's
interested
in
you
know
detailed
context
on
the
work
and
do
sprint
planning
and
combine
style
planning
on
that
day
and
then
just
do
that
every
two
weeks
see
how
that
thing
goes
and
then
the
highlights
for
the
next
release
is:
we've
got
now
a
built-in
plug-in
for
helm
and
we
will
probably
add
more
primitives
to
the
customization
file
to
let
make
helm
easier
to
integrate,
we're
doing
a
better
job
on
container
port
merging,
there's
multiple
keys
in
the
container
port,
not
just
one
key,
so
we're
closing
a
long-standing
bug,
at
least
from
the
point
of
view
of
customize,
we're
closing
on
sorry,
coop
cuddle,
there's
a
long-standing
bug,
we're
working
on
better
release
notes
and
making
that
automated
and
brian
presley.
F
I
still
owe
you
a
review.
It's
20
20.,
I'm
sorry!
Things
are
not
going
so
great
this
year,
but
that
is
a
high
priority
for
me
to
to
get
that
thing
tested.
I
want
to
test
a
few
things
on
my
own
clone
and
and
make
sure
that
it
makes
sense
to
me.
So
that's
it
so
next,
the
big
news
next
week
I'm
going
to
try
our
breakout
session
for
planning
for
customers.
You
want
to
show
up
that
sounds
good.
Sound
good
mark.
Does
that
make
sense,
yeah
perfect.
F
We
yeah,
if
it's
dedicated
sure
other,
because
I
don't
want
to.
I
also
don't
want
to
usurp
couple
work
either
right.
So
I
was
thinking
the
breakthrough,
but
yeah
either
way
I'll
show
you
if
everybody,
if
everybody
who
shows
up
just
wants
to
do
customize
fine,
then
we'll
just
do
customize.
Otherwise
they
can.
We
could
split
it.
A
Out
yeah,
the
other
option
is,
we
can
do
half
by
half,
half
do
cucaro,
diva,
keep
cuddle,
bug,
scrub
and
the
other
half
we
can
devote
to
customize
just
to
kick
it
off.
F
A
Yeah,
I
think
so
that
that's
a
good
starting
point
cool.
Okay,
all
cats
stand
up
here.
We
have
another
minute
and
I'll
would
love
to
get
the
newcomers
to
introduce
themselves.
I
think
we
lost
one
person,
but
going
from
the
top
the
tyler
introduced
himself.
I
think
mike
camp
is
new.
Is
that
correct?
You
want
to
introduce
yourself.
I
was
here
last
week.
You
are
here:
okay,
cool,
that's
good,
and
if
that's
the
case
muhammad
were
you
here.
Last.
B
D
L
A
Yeah
ricardo
was
helping
me
with
a
couple
of
stuff,
so
hi
ricardo
and
thanks
a
lot
for
your
help.
Hi
thank.
H
A
Please
manage
to
to
to
merge
several
create
commands
and
and
relate
it
cool.
Is
there
anyone
else
that
I
know
I
think
irve
ayani?
She
was
here
and
she
was
speaking
up,
but
I
I
don't
see
her
anymore
and
in
the
participants
list.
If
you're
new,
I
think
ram,
is
new
here
and
if
you
want
to
introduce
yourselves
now,
is
your
chance.
K
A
Okay,
okay,
I
think
we
have
all
covered
so
we're
one
minute
over.
I
don't
want
to
keep
you
any
longer.
Thank
you
very
much.
As
usual,
it
was
nice
seeing
you
all
and
thank
you
very
much
see
you
next
time.
Bye
thanks.