►
From YouTube: Kubernetes SIG CLI 20210728
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
B
We
actually
have
two
new
members
on
the
kui
side:
cora,
who
and
maria-
maybe
they
can
say,
hi
tell
them
what
you're
doing.
D
A
A
E
Sure
we
got
the
customized
bug
scrub
officially
scheduled
now,
so
the
first
one
will
be
taking
place
in
what
was
previously
our
off
week
from
the
sig
meetings,
which
is
next
wednesday,
so
next
wednesday
august
4th
at
the
same
time,
9
a.m,
pacific
or
equivalent
local
time.
Please
come
join
us
for
a
customized
bug,
scrub.
A
Hey
awesome
stoked
that
that's
the
thing
okay,
so,
let's
start
with,
we
have
two
topics
we'll
hopefully
get
through
them
really
quickly.
So
the
first
one
sean
you
want
to
explain
this.
F
Sure
so
the
google
security
team
flagged
a
security
vulnerability
in
the
docker
file
in
the
coop
cuddle
book
and
having
talked
to
phil,
who
is
the
driving
force
behind
the
coop
cuddle
book?
It
sounds
like
we
don't
really
need
we're
not
running
this
and
shouldn't
be
running
this
so
just
to
quickly
get
this
out
of
circulation.
F
I've
tempor
I've
created
this
pr
to
temporarily
comment
it
all
out
and
put
a
big
comment
saying
why?
So
I,
instead
of
just
removing
the
file
I
and
we
may
end
up
removing
you
know
all
of
the
book,
but
before
that
I
wanted
to.
You
know
just
comment
this
out
and
put
and
put
a
major
comment
to
point
out
why
so
yeah
I'd
love
to
hear
feedback.
F
If
this
is
the
wrong
direction,
if
you
guys,
if
there's
a
different
direction
and
yeah
again,
my
understanding
from
phil
was
that
we
are
not
running
this
anywhere.
The
the
coop
cuddle
book,
a
lot
of
work.
Oh.
G
Migrate,
this
content
to
the
merged,
customized,
kill
it
all
with
docs,
and
so
I
think
we
can.
We
can
probably
just
get
rid
of
all
this
content.
Bring
down
the
site,
like
we've,
already
moved
the
content,
and
this
should
probably
go
away.
So,
starting
with
just
commenting,
this
out
seems
like
a
great
first
step
and
then
we
should
probably
go
back
and
delete
the
content.
F
Okay,
cool
and
so
just
for
quick
context.
The
security
vulnerability
is
the
curl
piping
to
a
shell,
which
you
see
on
line
18
there
or
somewhere
around
thereabouts,
which
is
the
the
way.
So
if
these,
my
understanding
is
if
the
site
that
we're
curling
from
is
compromised,
then
it
can
get
propagated
to
other
places
that
are
downloading
from
that.
F
F
A
F
Yeah
and
and
so
right
now
we're
in
the
middle
of
a
code
freeze,
and
so
I
think,
we're
going
to
have
fun
trying
to
get
this
particular
thing
merged
since
the
release
is
coming
out,
but
I
I
think
you
know
tell
me
if
I'm
wrong.
I
think
that
this
sounds
like
it
has
a
pretty
good
reason
to
to
get
merged.
F
A
Yeah,
I
would,
I
would
toss
it
in
the
milestone,
probably
check
with
whoever
the
release
team
lead
is
for
this
one.
I
think
it's
sarita,
but
I
don't
see
any
objection
to
that.
H
Yeah
definitely
sync
with
them:
it
shouldn't
affect
anywhere
or
anything
in
the
release
process,
and
that
would
be
good
starting
good,
starting
point
for
making
sure
that
this
is
fixed
for
122
and
for
further
version.
We
will
just
get
rid
of
it
entirely.
A
I
wonder
if
this
I'm
just
thinking
beyond
6dli
here
I
wonder
if
this
is
used
anywhere
else
throughout
the
docs
or
the
project.
Who
should
we
like
tag
to
do
an
audit
on
that.
H
I
would
probably
sync
with
the
docs
team:
I
think
we
moved
the
cube
cuddle
officially
to
the
docs
team,
or
we
were
too.
I
can't
remember
where
that
stands.
That's
definitely
would
be
something
that
I
would
double
check
before
the
final
remove
of
everything.
F
A
Cool,
thank
you,
google
security
team
for
looking
out.
Okay.
Last
topic.
We
have
this
one
just
kind
of
came
up
last
night.
The
tldr
is
that
it
is
not
possible
to
set
a
request
timeout
for
a
discovery,
client
right
now,
if
you're
using
the
discovery,
client,
constructors,
that's
because
go
defaults,
values
to
a
zero
value,
and
we
have,
if
you
look
right
there
on
line
465,
it's
checking.
If
the
timeout
is
zero.
A
This
is
the
user
supplied
timeout,
so
user
can
pass
in
zero,
which
is
what
clientgo
says
to
do
to
disable
timeouts.
So
it's
checking
if
it's
zero
and
then,
if
it
is,
it's
overriding
it
and
setting
it
to
the
default
of
32.
Second
timeout,
so
just
kind
of
like
a
little
go
thing
that
bites
us.
Sometimes.
A
I
think
we'll
probably
need
to
talk
to
sig
api
machinery
on
this,
because
this
is
part
of
client
go
but
yeah
you
can't
you
can't.
You
can
set
a
request
time
out
of
zero
for
the
rest
of
cube
control.
It's
just
the
discovery,
client,
so
the
first
part
of
the
client
that
runs
and
kind
of
caches,
all
the
apis
that
you
are
unable
to
set
a
timeout
of
zero
to.
You
can
set
it
to
negative
one,
but
we
instruct
people
to
set
it
to.
H
The
discover
operation
is
very
intense
because
it
literally
scrapes
top
level
api
and
apis
and
then
walks
through
every
single
api
endpoint
that
is
registered
in
cube
api
server.
All
the
extension
servers
all
the
crds
everything
else
that
is
registered
in
the
cube
api
server,
so
I'm
not
sure
if
that
somehow
wasn't
a
protection
to
not
to
overwhelm.
On
the
other
hand,
we
also
have.
H
A
If
you
flip
to
the
conversation
and
then
go
down
to
the
very
bottom
comment,
the
documentation
for
api
machine
client
go
config
scroll
to
the
little
code
snippet
at
the
bottom,
at
the
bottom
of
the
comment,
right
above
your
name
no
down
down
down
up
down.
That's
the
top
sorry
yeah
right
there
at
the
bottom
of
my
comment.
There
you
go
yeah
scroll
that
over.
So
this
is
a
client
go
config,
and
so
it
says
yeah.
It
says
the
maximum
length
of
time
delay,
blah
blah
blah
a
value
of
zero
means,
no
timeout.
A
H
A
That's
all
I
had
for
this
one
any
other
topics
or
we
can
move
on
to
stand-ups.
A
Cool
anyone
stand
ups,
customized
kui.
B
Not
too
much
on
the
kui
end,
I
think
once
and
maybe
next
time
cora
will
be
ready
to
present
some
of
the
findings.
Initial
findings
from
the
study.
E
Jeff,
I
don't
know
if
you
have
anything
to
add,
but
we
don't-
we
haven't
done
anything
terribly
exciting
lately,
but
some
really
important
work
sort
of
following
up
on
some
particular
technical
debt
that
we've
been
wanting
to
clean
up.
That's
in
our
roadmap
natasha's
doing
a
bunch
of
work
on
that,
which
is
great.
I
think
we
already
mentioned
before
that
we
did
get
our
major
bug
fix
where
we
had
to
fender
go
yaml.
E
Yeah,
it's
forked
internally
to
kml,
so
only
kml
can
use
it.
We
don't
want
to
maintain
it
in
the
long
run
and
the
maintainer
of
glue
gamble
is
still
promising
to
make
the
changes
upstream
that
we
need.
We
still
have
our
prs
open,
so
we're
just
kind
of
waiting
on
that,
but
the
idea
is
to
get
rid
of
the
fork
by
like
before
the
next
release
for
sure.
A
Cool,
I
just
saw
we
have
a
ricardo
carter.
Are
you
here
yeah?
So
we
had
a
topic
of
yours
that
we
punted
it's
at
the
top
should
all
commands
have
label
support.
I
think
this
is
an
old
one.
J
H
H
We're
in
the
agenda
that's
and
where
we
discuss
this,
I
somehow
miss
this.
Maybe.
J
My
only
my
only
concern
about
this
one
is
that
they
are
using
like
k
flag
instead
of
l,
because
l
conflicts
with
the
label
selector
for
files.