►
From YouTube: Kubernetes SIG CLI 20220511 - Kustomize Bug Scrub
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
Hello
and
welcome
everyone
to
today's
customized
bug
scrub
today
is
wednesday.
May
11th
before
we
get
started.
Does
anyone
have
an
issue
that
they
would
like
us
to
look
at
first
or
any
point
of
discussion
that
they
would
like
to
bring
up.
A
So
today's
bug
scrub
we're
going
to
use
the
customized
triage
board.
As
usual,
we
have
a
number
of
brand
new
issues
that
we
haven't
had
a
chance
to
look
at.
So
this
time,
I've
sorted
by
creation
date
and
I
think
we'll
go
from
there,
but
if
anyone
has
suggestions
feel
free
to
speak
up
at
any
time,
so
I
guess
we
can
just
get
started
with
the
first
one.
A
The
idea
here
by
the
way
is,
is
to
go
through
and
gain
context
on
the
customized
issues
to
get
some
triage
done
for
the
project
and
also,
if
you
see
something
that
you're
interested
in
working
on,
we
can
really
use
help
with
with
these
issues.
So
please
do
speak
up
we're
happy
to
help
discuss
the
issue
further,
and
I
give
you
pointers
as
to
how
to
get
started
with
the
issues
that
we
go
over
today,
all
right.
So
without
further
ado,
we
are
starting
off
with
a
question.
A
A
B
A
A
So
I'll
save
this
one
to
write
a
more
detailed
response
later,
I'm
not
quite
sure
what
to
say
off
the
cuff
like
there's.
Not
the
concern
is
valid,
like
you
do
need
to
make
a
decision
to
make
sure
that
what
you're
doing
is
secure
and
we
do
want
at
the
same
time
to
be
committing
as
much
as
possible.
But
committing
secrets
in
clear
text
is
not
something
that
you
want
to
do
clearly.
A
So
I
don't
know
the
recommendation
is
almost
under
the
scope
of
customize
like
there
are
a
lot
of
possibilities
outside
it
that
I'm
not
in
a
position
to
make
a
recommendation
about
so
I'll
probably
write
something
to
that
effect.
C
I
read
this
one
this
morning:
they
want
to
apply
labels
to
metadata.labels
and
also
like
pod
templates
from
my
understanding
of
this
issue,
but
not
selectors,
which
was
an
issue
we
had
a
while
ago,
and
I
think
we
had
a
pr
that
fixed
it,
but
I
don't
think
it's
been
released
yet
so
I
can
respond
to
this
issue
with
the
relevant
links
and
let
them
know
it's
coming
soon.
A
That
would
be
great
I'll
assign
you
here,
and
I
agree
that
we
merged
that
very
recently,
so
we
haven't
done
a
release.
A
A
A
A
That's
that's
my
theory.
It
could
also
be
the
strategic
merge.
Patch
implementation,
like
I'm
still
trying
out
we're
gonna
hit
around
this
scenario
here
because
like
when
I'm
guessing,
they
want
to
retain
because
they're
saying
no
bug
is
the
one
that
has
the
null
key
retained,
which
is
coming
from
original,
not
from
the
patch.
So
I'm
guessing.
A
If
they
want
that
retained,
they
want
to
be
like
sending
that
to
the
api
server
so
that
when
the
api
server
requires
the
strategic
merge
patch,
that
key
will
get
deleted
but
like
if
they
ask
us
to
do
a
strategic,
merge
patch
on
the
client
side
as
well,
then
we're
doing
a
strategic,
merge
patch,
so
we'll
remove
the
key
in
theory
so
that
that's
that's
my
second
theory
about
what
what
could
be
going
on.
A
D
A
A
bars
bug,
so,
even
if
this
is
a
bug
we
likely
wouldn't
fix
it
because
of
ours
is
deprecated
just
for
starters.
Here
let
me
paste
that
in
the
chat.
C
I'm
to
find
a
documentation,
but
if
I
remember
correctly,
you
can't
use
a
var
anywhere
right.
I
thought
there
were
only
like
specific
locations.
A
Honestly,
I
don't
have
that
much
experience
with
virus,
I
they
were
already
deprecated
when
I
joined
the
project
and
I
haven't
used
them
personally,
so
I
don't
even
know,
but
let's
see,
if
there's
a
handy
link,
we
could
give
them
about
that.
That
would
be
potentially
good
for
solution.
Not
you've
got
it.
B
Oh
replacements,
which
the
docs
say.
B
C
I
think
they're
in
that
doc
there's
a
link
to
this
file
of
our
reference.go,
which
I
think
is
all
the
places
you
can
put
vars.
C
A
And
so
this
one's
definitely
not
a
bug,
kind
support,
yeah,
just
triage.
E
Yeah,
I
can
just
summarize
briefly,
so
I
worked
on
an
issue
that
was
to
write
more
comprehensive
integration,
tests
for
urls
and
the
resources,
and
so
I
was
reading
the
remotebuild.uh
markdown
under
examples,
and
it
said
that
we
support
resource.
E
We
support
query
string
parameters
along
with
our
urls,
and
so
I
wrote
a
few
tests
for
that
in
the
the
file
is
like
api
krusty
remote
load
underscore
test,
and
what
we
observe
right
now
is
that
three
of
them
throw
errors
and
the
the
I
I
documented
the
the
types
under
the
numbered
list
there.
So
if
the
url
has
no
parameters,
so
if
it
doesn't
have
a
ref,
I
guess
that's
the
most
popular
one
that
fails.
E
A
A
I
mean
that
sounds
like
a
real
problem
that
we
want
to
fix,
so
I'm
inclined
to
accept
this.
Do
we
have
anyone
who's
interested
in
working
on
it.
A
Oh
great
excellent,
can
I
assign
you
on
github,
what's
your
handle.
D
Yeah,
it's
oleg
seo
I'll
try
to
to
take
a
look
about
it.
B
Yeah,
if
they
haven't
participated
in
the
in
the
issue,
yet,
okay.
E
Yes,
this
one's
also
me
I
filed
a
couple,
so
this
is
for
the
same
code
that
I
wrote,
and
so
we
don't.
E
We
don't
specify
in
remote,
build
like
down
that
we
support
ports,
but
we
say
that
we
support
get
clone
urls
and
in
the
on
the
git
clone
page,
they
say
that
they
support
ports
and
when
I
added
tests
for
the
default
ports,
I
guess
the
real
use
case
would
be
custom
ports
right
if
people
want
to
use
different
ports,
but
I
use
the
default
ports
and
they
worked
for
they
worked
locally.
E
When
I
was
trying
on
git
repositories,
I
can't
try
and
sub
directories
because,
like
it
doesn't
natively
support
that,
but
the
tests
fail.
When
I
I
write
our
integration
tests.
A
Interesting,
I
don't
have
any
objection
to
supporting
this
exactly,
but
I
would
be
interested
if
this
is
something
folks
actually
need,
because
we
have
so
many
issues
that
we
need
to
tackle.
And
I'm
wondering
if
this
is.
E
Right
yeah,
like
I
natasha,
also
said
the
same
thing
I
just
filed
it
just
in
case
we
needed
it
as
reference,
but
I
don't
feel
too
strongly.
A
A
E
Unclear
I
mean
something
we
could
also
do
is
in
in
remote
build
just
say
we
don't
support
ports
like
I'm.
Also
down
with
that.
I
just
found
it
a
little
surprising
because
we
we
said
we
did
and
then
there
was
no.
I
guess
qualification.
A
E
A
Right,
okay,
so
maybe
in
the
meantime
that's
a
good
thing
to
do
is
just
like
update
the
docs
to
say
that
they're
not
supported.
A
And
then
I
think
this
is
your
third
one.
Perhaps
you
just
skip
right
to
that.
E
Okay,
yeah,
basically,
all
the
ssh
key.
All
the
ssh
tests
fail.
I
think,
because
the
server
doesn't
have
ssh
keys.
A
C
C
There
are
skips
generally,
but
we
could
add
a.
I
don't
know
a
gate
so
that
we
can
run
them
locally.
A
I
guess
most
of
the
time
we're
running
on
ci
anyway,
because
you
have
so
many
tests
to
take
a
long
time
to
run
locally
but
like
we
do
always
run
the
test
locally
during
release,
so
that
would
still
get
us
sort
of
a
last
ditch
we
broke
something
before
we
release
it
signal.
A
A
Not
a
process
I
like
in
general
to
rely
on
something
like
that
or
a
light
a
last
minute
test
test
run,
but
it's
better
than
nothing.
If
we
can't
you
know,
we
can
have
that
safeguard
in
place
immediately,
instead
of
after
weeks
or
months
that
it
will
take
to
fix
the
ci
server
and
that's
better
than
better
than
nothing.
A
Thanks
for
adding
those
new
tests
by
the
way,
those
are
tests,
we've
never
had
before
for
a
very
underdocumented
feature.
A
A
D
C
Is
there
this
is
like
sort
of
related,
but
is
there
an
easy
way
for
a
user
to
figure
out
what
the
merge
key
is
like
for
me,
I
go
through
the
entire
like
open
api
json
file,
which
is
huge,
but
if
I'm
just
using
customize-
and
I
want
to
write
a
merge
or
like
a
patch
strategic
merge
like
how
do
I
know
what
the
merge
key
is.
B
Yeah,
the
under
containers.
They
start
a
gamble
list
with
env.
B
A
Right
so
or
the
yeah,
that
was
the
merge
key.
What
what
would
we
recommend
people
look
at.
A
Nothing
comes
to
mind.
I
don't
know
if
I've
ever
seen
like
a
good,
a
good
talk
about
what
the
merge
key
is.
It
could
be
in
the
api
docs
in
theory.
Maybe
I'm
really
not
sure.
Usually
it's
name
right,
but
I
think,
what's.
A
C
A
Yeah
it's
in
the
open
api.
I
have
a
link
enough.
A
C
A
Oh,
it's
not
that
one.
Maybe
it
got
closed.
I
I
could
just
when
I
reviewed
a
pr
where
someone
felt
strongly
about
it.
They
updated
the
script.
So
we
were.
A
Right
yeah,
we
couldn't
take,
they
copied
something
from
stack
overflow,
so
we
couldn't
take
it
as
is,
and
it
was
also
over
complicated
compared
to
what
it
needed
to
do.
So
I
made
a
suggestion,
but
the
contributor
disappeared.
I
think
if
somebody
wanted
to
pick
this
up,
we
could
merge
it.
It's
not
super
complicated
to
do,
but
it's
not
high
priority
for
us,
because
we
don't
really
support
the
older
versions.
A
Let
me
just
check
they're
complaining
about
three
eight
five.
I
actually
have
the
list
here
about
what
that
is.
Eight.
C
C
We
don't
have
a
three
eight
five
tag.
The
last
one
is
three
three
one.
Let
me
check
the
releases.
The
release
is
gonna,
be
so
far.
Three
point.
C
C
A
C
Let's
one
sec,
my
tags
are
getting
truncated.
A
Yeah
here
it
is
tag
customize
385
that
actually
should
work.
Oh.
A
Yeah,
doesn't
the
script
look
for
the
version
that
you've,
given,
though
so
it
should
like?
They
should
use
the
version
on
master
of
the
script
and
then
yeah.
If
the
argument
is
given.
A
B
A
A
A
True
go,
yaml
could
be
involved,
but
I
think
regardless
we
should
have
the
ability
to
force
the
quote
style
because
go
yama
lets
us
force.
C
A
A
A
Just
think,
string
and
started
by
double
quotes
yeah.
They
had
this
situation
where
they
were
actually
quoting
some
things
in
the
file
and
not
quoting
other
things,
and
I
remember
this
from
last
time.
A
Yeah,
that's
a
good
one
to
work
on.
It
might
not
be
super
straightforward,
because
working
with
the
yaml
code
is
always
a
bit
of
a
pain,
but
that
would
be
pretty
valuable
if
somebody
watching
the
recording
is
looking
for
a
bug
to
look
into.
C
A
A
I
see
marcus,
you
might
may
look
into
the
quoting
issue
in
the
future.
Thank
you
for,
for
speaking
up.
A
A
The
architecture
on
our
talk
site
needs
work,
it's
tasty.
Let's
get
to
these
things.
Luckily
we're
working
on
that
that's
another
area.
If
anyone
wants
to
get
involved,
we're
working
on
putting
the
content
over
from
the
site
to
the
new
work
in
progress
site
that
we
have
on
our
repo
and
it's
a
fairly
straightforward
matter
of
moving
the
content
piece
by
piece.
We
have
one
person
helping
us
with
that,
but
we
have
a
lot
of
content
to
move.
So
that's
another
opportunity
for
involvement
at
this
time.
A
A
A
I
didn't
so
there's
probably
a
difficult
issue
somewhere.
We
definitely
cannot
do
this
because
I,
I
think
and
natasha
would
know
best,
but
the
home
feature
actually
used
to
have
something
very
similar
to
this,
that
let
you
pass
flags
through,
but
it
was
deemed
a
security
risk
and
we
had
to
remove
it.
So
we
definitely
can't
accept
this
feature
request.
A
I
kind
of
sworn
natasha
had
a
a
good
reply
for
this
one
though
I'm
I'm
going
to
just
take
that,
and
and
like
look
for
natasha's
previous
reply
to
a
very
similar
question,
passed
and
linked
to
the
plan
where
we
have
an
external
transformer,
that's
going
to
aim
to
have
a
field
for
every
single
flag
so
that
we
can
support
all
those
things.
A
This
is
a
kml
issue,
specifically
not
customized,
by
the
looks
of
it.
B
A
That's
weird
yeah.
I
wonder
what
specifically
is
causing
that
if
it's
go
yaml
then,
although
like
it
sounds
like
it
would
be
fine
to
fix,
it
might
be
in
practice
impossible
for
us
right
now.
I'm
also
interested
in
whether
or
not
that
causes
an
actual
problem
for
them
or
if
they
just
don't.
A
A
A
And
then
that
might
give
me
an
idea
of
whether
it's
cinco
yemol,
they
said
they're
going
to
try
to
fix
it
themselves,
which
is
fine,
but
I
should
give
them
a
fair
warning
that
it
won't
be
possible
if,
if
it's
coming
from
go
yama,
we
don't
want
to
do
anything
crazy
to
hack
around
their
behavior.