►
From YouTube: Kubernetes SIG CLI 20221019
Description
Kubernetes SIG CLI Bi-Weekly meeting on October 19th, 2022.
Agenda and Notes: https://docs.google.com/document/d/1r0YElcXt6G5mOWxwZiXgGu_X6he3F--wKwg-9UBc29I/edit#bookmark=kix.9k21ak2kszwq
A
Start
recording
and
ensure
that
we
are
recording
Yep
looks
like
we
are
so
hello
and
welcome
to
the
October
19th
edition
of
the
6C
live
bi-weekly
meeting.
My
name
is
Sean
Sullivan
and
I
will
be
your
host
today
and
we're
going
to
get
right
into
the
announcements.
A
So
there
aren't
many
announcements
that
are
at
least
on
the
agenda
number
one
is.
Let
us
please
remember
that
we
have
the
code
freeze
coming
up
in
about
four
weeks
in
early
November,
maybe
it's
three
weeks
three
or
four
weeks,
November
9th
on
a
Wednesday
and
then
also
very
excitingly.
We
are
going
to
have
kubecon
North
America
next
week
in
Detroit
for
the
entire
week
October
24th
through
the
28th
that
was
all
I,
could
think
of
to
add
to
the
announcements.
A
This
is
the
time
of
the
meeting
where
you
we
all
get
to
meet
our
colleagues,
and
so,
if
you're
new
to
the
group
or
if
you
haven't
been
here
for
a
while
and
you'd
like
to
to
meet
your
sexy,
like
colleagues,
please
introduce
yourself
and
if
you
don't
want
to,
then
that's
fine
as
well.
Is
there.
A
And
the
first
topic
is
a
topic
that
that
I'm
gonna
present
and
maybe
I'll
get
a
Little
Help
from
Alex
and
Antoine,
if
they're
here,
but
but
basically,
there's
been
some
movement
on
the
Coupe
cuddle
apply
prune
and
there
are
two
documents
that
I
would
like
to
introduce
to
the
Sig
which
have
to
do
with
a
coupe
cuddle
applied
prune,
and
the
first
is
a
a
document
describing
basically
why
prune
is
what
what
the
problems
are
with
the
current
implementation
of
prune.
A
What
what
is
the
current
implementation
and
what
are
those
limitations
and
I'll
just
go
ahead
and
open
that
document
and
I'm
not
going
to
dig
too
deep
into
this,
but
you
could
see
in
the
table
of
contents.
It
describes
what
prune
is
what
the
implementation
is
and
then
what
are
those
limitations
for
the
current
implementation?
A
And
if,
if
we
decide
as
a
Sig
that
we
would
like
to
fix
prune,
then
this
next
document
is
three
particular
approaches
which
have
been
discussed
at
length
before
in
and
out
of
the
Sig
that
I've
kind
of
put
an
overview
for
for
those
three
approaches,
and
mostly
this
is
these
aren't
meant
to
be
complete.
It's
not
meant
to
be
a
complete
design
doc.
All
it
is,
is
meant
to
give
context
to
the
rest
of
the
Sig
on
this
particular
topic.
A
So
I
didn't
like
I,
said:
I,
don't
want
to
spend
a
whole
lot
of
time
on
this,
but
I
just
want
to
make
sure
that
the
Sig
that
that
you
all
have
the
information
you
need
so
that
we
can
have
discussions
on
on
how
we
move
forward
with
Coupe
cuddle
applied
prune
and
and
with
that
I'm
going
to
ask
if
Katrina
would
like
to
if
she'd
like
to
say
a
few
words,
because
she
has
a
lot
of
context
on
this
issue
and
and
also
a
passion
for
it.
Katrina.
A
Would
you
like
to
say
anything
about
apply
prune.
C
C
It's
also
a
very
dangerous
command,
but
a
very
useful
one,
that
sort
of
allows
folks
to
treat
a
whole
set
of
documents
the
same
way
they
would
treat
Fields
when
they
use
apply
because,
often
when
you
are
creating
a
set
of
resources,
you
are
managing
them
as
as
a
set
and
you're
applying
them
as
a
set
so
being
able
to
say
like
this
is
exactly
what
I
want
to
have
in
my
namespace
or
label
delivered
or
labeled
to
limited
subset
of
that.
C
Namespace
can
be
really
useful
if
you
manage
to
use
pruning
safely,
which
is
hard
so
I
think.
My
opinion
on
this
is
that
we
we
have
had
this
in
there
for
so
long
and
I
believe
that
we've
agreed
to
this
at
a
previous
Stig
meeting
that
we
can't
just
get
rid
of
it
like
treat
it
like
a
normal
Alpha
and
just
delete
it
one
day
to
the
next
or
anything
like
that.
C
It's
it's
been
in
I
forget
since
what
version,
but
an
extremely
an
extremely
early
version.
So
there
are
lots
of
people
at
this
point
who
likely
rely
on
it.
C
In
fact,
we
have,
at
previous
day
meetings
decided
that
we
can't
even
change
make
Behavior
changes
without
a
deprecation
process
that
normally
wouldn't
apply
to
an
alpha
so
yeah.
That's
that's
something
that
to
keep
in
mind
when
we're
discussing
this
topic
that
we
actually
need
to
treat
it
like
it's
more
evolved
than
awful,
normally
is
so
yeah
and
I
maintained
a
tool
for
my
company
that
actually
uses
this
under
the
hood
in
a
constrained
way
to
make
it
safe.
C
So
even
from
that
personal
perspective,
I
do
think
this
is
a
valuable
feature.
That's
worth
evolving,
though
I
totally
agree
about
the
limitations
described.
A
Okay,
great
and
it
sounds
like
there's
going
to
be
a
very
good
contingent
of
the
Sig
CLI
present
at
kubecon
next
week
and
so
I'm,
hoping
that
we
can.
We
can
have
some
substantive
discussions
and
and
make
some
progress
on
that.
What
do
you
think
Katrina?
Do
you
think
that's
appropriate
for
him.
C
Do
we
have
a
particular
session
in
mind
for
that?
Like
did
we
get
time
at
I.
A
C
Well,
if
the
goal
is
to
make
progress
on
a
specific
issue,
then
I
think
yeah.
We
should
solicit
feedback
on
on
who
is
interested
in
participating
and
make
sure
that
they
all
have
the
opportunity
to
be
there,
which
could
also
mean
that
we
can't
do
it
at
capecon,
because
the
folks
who
are
most
passionate
about
this
aren't
there,
for
example,.
B
A
Okay,
so
moving
on
to
the
next
topic,
which
is
which
was
brought
up
at
the
it
was
added
to
the
agenda
after
we
came
across
this
issue
again
at
our
bug,
scrub,
and
so
this
is
that
get
all
is
not
all
and
let
me
bring
up
the
issue
is
the
issue
visible.
A
Okay,
so
this
this
actually
it
was
closed,
as
as
Eddie
pointed
out,
it's
mentioned
in
our
docs
that
you
know
this
particular
item
is,
is
it's
it's
Legacy,
it's
hard-coded
and
you
know
I.
We
brought
it
up,
I.
Think
to
this
sig
to
to
say
you
know:
what
should
we
do
to
make
this
less
confusing
I
know
I
see
Eddie
is
Here
Eddie.
Would
you
like
to
comment
on
on
that.
A
Okay,
Okay
so,
and
we
yeah,
we
already
have
the
a
mention
of
adding
a
warning.
B
A
Okay
was
Sir.
Was
there
anything
else,
Eddie
that
we
wanted
to?
Besides,
adding
the
warning,
do
you
should
we
be
able
to
move
on
to
the
next
topic.
C
Yeah
I
think
it
could
be
noisy,
so
we
need
to
figure
out,
like
you
said
how
that
would
actually
work,
but
yeah
or
even
I
think
we
talked
about
this
last
time
and
I
suggested
like
because
this
is
always
outputting
a
subset.
Even
you
know
it
outputs
a
like
a
section
for
each
resource
type
that
it
actually
found,
but
typically
you
might
not
even
have
all
the
kinds
that
it
looked
for.
So
we
could
have
a
message
that
says
something
like
did
not
find
any.
C
E
A
So
so,
usually
all
of
the
queries
are
by
the
gbk
and
so
all
will
I'm
I,
think
issues,
multiple
queries
with
empty
label
to
list
particular
gbks
and
the
gvks
are
the
ones
that
are
within
the
all
category
or
have
they
all
category?
How
does
that
work?
Brian
I
know
you.
You
recently.
A
D
It's
just
a
category
on
the
resource
that
optional
category
and
some
of
them
happen
to
have
a
category
called
all
and-
and
it's
really
it's
kind
of
really
one
thing
that
occurs
me
while
we're
talking
about
this
is
that
could
call
get,
can
do
two
things.
It
can
get
a
individual
resource
or
category,
and
maybe
maybe
it's
maybe
we
shouldn't
let
them.
D
D
D
E
C
Breaking
change
and
those
are
really
difficult
and
slow
to
make
so
I
think
the
the
warning
might
be
the
more
tractable
approach.
C
A
And
by
making
an
explicit
I
think
it
would
remove
a
lot
of
the
confusion.
As
Brian
said,.
A
C
It
feels
like
a
pretty
big
change
to
alleviate
the
confusion,
specifically
for
one
subset
of
the
usage.
E
The
warning
and
we'll
we
can
talk
more
about
how
to
do
the
warnings
later.
E
C
Certainly
adding
it
to
the
docs,
which
should
be
pretty
straightforward
and
valuable.
Folks
read
the
cube,
control
doc
or
the
kubernetes
docs
more
than
they
read
our
Docs.
A
A
Okay,
if,
if
we're
agreed,
we
could
move
on
to
the
next
topic.
A
So
Marley,
would
you
like
to
introduce
this
next
topic.
F
Yes,
I
have
done
some
work
with
the
Json
path.
Library,
that's
maintained
imagery
for
for
us,
and
it's
come
up
a
couple
of
times
an
issues
just
like
it.
F
Fully
implement
the
spec,
which
is
understandable.
Most
of
the
changes
or
most
of
the
most
of
the
implementation
was
done
like
more
than
five
years
ago,
and
I
was
curious.
If
anybody
cared
enough
to
make
it
worth
like
trying
to
make
it
more
in
line
with
the
Json
path
spec,
because
I
was
considering
tackling
that.
F
But
what
is
the
context
of
this
Json
path
there?
So
we
keep
Kettle
uses
Json
path
for
a
couple
of
things.
F
Probably
the
most
common
thing
is
printers
like
deciding
what
values
get
printed
out
and
things
like
that
whenever
you
get
a
resource
or
something
you
can
like
dig
down
into
like
specifically
just
what
you
want
to
pronounce
I
personally
have
mostly
been
working
on
it
in
the
context
of
the
cube
Kettle
config
set
command
and
making
it
possible
for
people
to
use
Json
path
to
access,
just
like
any
K
in
the
cubeconfig
file.
F
That,
in
particular,
has
led
me
to
make
some
updates
to
the
Json
path,
module
that
we
have,
and
you
know
looking
at
them
like
this
is
gonna
suck
to
do
like
this
is
going
to
be
a
really
hard
problem,
basically
and
so
I
just
before
I,
dug
into
all
of
that
I
wanted
to
see
if
any
I
don't
know,
if
there's
any
political
will
basically
within
the
Sig.
For
that.
E
Which
I
didn't
know
about
until
I
just
searched
for
it
looks
like
it
was
closed.
F
But
yeah
this
was
I
was
gonna
make
a
cup
for
because
this
is
probably
going
to
be
a
pretty
long
running
effort.
Yeah.
E
If
David
Eaves
is
the
one
who
closed
this
David
knows
all
about
the
the
dragons
that
lie
in
that
code.
From
the
last
time
you
worked
on
this,
so
it's
you
could
always
just
reach
out
to
David
and
see
if
you
know
he
would
chat
or
give
any
information
about
that.
E
F
And
I
assume
that
there's
no
because
active
will
against
just
using
an
external
dependency
to
import
like
Json
path
support
because
that's
generally
the
opinion
from
what
I
understand.
F
I
looked
around
at
one
point
and
there
just
aren't
a
whole
lot
of
libraries
that
I
found
that
do
Json
path
implementation,
but
there
are
a
couple
and
I
can
see.
I
will
have
to
like
front
them
all
so
that
this
doesn't
become
like
a
huge
breaking
change.
Obviously,
but
yeah
okay
I'll
probably
start
there,
because
then
I
don't
have
to
do
a
bunch
of
like
weird
string,
parsing
and
algorithm
writing.
B
Is
the
problem
you
know,
you're
also
going
to
hear
how
you
should
use
JQ
if
you
want
to
do
something
specific
I
think
what
I've
heard
a
lot
of
I
remember
talking
about
this
in
the
past,
but
a
very
long
time
ago,
I
think
people
in
general
have
been
of
the
opinion
that
you're
trying
to
do
something
complicated.
You
should
use
JQ,
because
why
would
we
Implement
JQ
in
cook,
pearl.
C
C
Have
empathy
for
people
who
are
you
know
they
think
it
looks
like
Jakey
or
sorry.
It
looks
like
Jason
path.
We
say
it's
Jason
path,
but
no
it's
the
special
subset
of
Jason
path.
They
have
to
then
go
learn
like
that's.
It's
not
very
empathetic
of
us
to
just
say
like
well.
They
have
to
go,
learn
it
because
because
we
don't.
C
F
The
oh
well
I
mean
it's
not
a
matter
of
laziness
like
that's
a
hard
problem
to
solve,
but
yeah
the
the
so
the
last
person
that
brought
up
the
issue
actually,
so
they
mentioned
that
they
just
didn't
want
to
install
JQ
in
their
CI
CD
pipeline,
which
I'm
generally
of
the
opinion
of
to
you,
I,
don't
like
it
solving
extra
crap
in
my
CI
CD
pipeline,
and
so
that's
kind
of
like
where
I'm
thinking
of
it
basically
I.
F
You
know,
I,
don't
know
if
I
can
find
a
way
to
do
this
without
it
being
like
horrible,
then
that
would
be
great
and
it
is
really
frustrating
that
you
can't
just
use
like
the
way
that
this
person
was
talking
about
using
these
things.
I
did
end
up
finding
out
a
way
to
just
do
it
using
bash
variables
and
but
yeah.
It's
not
ideal.
F
Writing
your
own
Json
path.
Implementation
would
be.
F
F
I
I
just
need
to
re-go
through
the
ones
that
I
found
the
first
time.
I
was
looking
for
this
and
just
double
check
and
make
sure
that
I'm,
not
forgetting
ones
that
were
worth
pulling
in,
but
I
originally
I
hadn't
thought
about
it
until
I
started
talking
with
us
now
that
using
external
dependencies
was
probably
okay,
if
it
was
like
actually
worth
using
the
external
dependency,
not
using
it
just
to
use
it.
E
B
A
F
Okay,
yeah.
That
would
be
good
as
for
different
parts
of
cube
Kettle
using
like
different
ish
implementations,
yeah,
that's
the
I,
wouldn't
say
it's
like
different.
F
Exactly
they
all
call
the
Json
path,
library
that
we
have,
but
they
just
do
different
things
with
it
and
so
like,
for
example,
what
I
did
for
config
set
is
not
really
the
same
as
what
gets
used
for
the
printers
consumption
of
it,
because
I'm
just
like
ignoring
most
things
that
it
provides,
because
we
just
don't
need
it
for
config
set,
because
all
you're
doing
is
looking
up
a
value
and
setting
it,
and
maybe
potentially
I
can't
even
imagine
a
use
for
this
exactly
but
like
looking
up
a
bunch
of
different
values
and
setting
them
that
way.
F
Yeah
so
I'll
just
I'll
start
by
looking
to
see,
if
there's
some
external
dependency
that
we
can
pull
in
that's
worth,
pulling
in
and
figuring
out
a
way
to
wrap
it
so
that
it's
not
like
a
breaking
change.
Because
if
it's
going
to
be
a
freaking
change,
it's
going
to
be
impossible
because
it's
used
in
so
many
different
places.
A
E
I
think
we
can
look
at
that.
One
real,
quick
I
got
a
survey
for
folks
to
look
into
and
respond
to
for
localization
something
we've
talked
about
a
lot
and
Brian's
take
on
ownership
of
it.
So
it
looks
like
there
is
a
draft
of
questions
here
and
Brian
is
probably
looking
for
some
feedback
on
that
draft.
A
E
E
E
E
A
C
A
C
A
Okay,
well,
it
looks
like
we're
not
going
to
take
the
full
allotment
of
an
hour
for
this
meeting
before
we
sign
off.
Is
there
anything
anybody
else
would
like
to
add.
A
Okay,
well,
we
appreciate
your
time
and
for
those
of
you
going
to
kubecon
in
person,
we'll
see
you
next
week
and
if
you
aren't
going
in
person,
maybe
you
could
go
virtually
it's
going
to
be
a
really
interesting
conference
and
the
six
CLI
will
be.
We
will
actually
have
a
we'll
be
talking
about
the
updates,
we'll
be
giving
an
overview
of
what's
going
on
in
the
six
CLI
and
while
I'm
at
it
thanks
a
lot
to
Eddie
for
all
of
the
effort.