►
From YouTube: Kubernetes SIG CLI 20220706 - 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
Okay,
I
think
it's
recording
now,
okay,
so
hello,
everybody!
This
is
a
customized
bug
scrub,
so
we
are
going
to
be
going
through
the
customized,
triage
board
and
hopefully
resolve
or
at
least
try
out
some
issues.
So
I
have
it
sorted
here
by
order
of
creation,
so
the
oldest
issues
will
come
up
first,
but
before
I
get
started,
are
there
any
issues
that
anyone
wants
to
bring
up
to
give
particular
attention.
A
Okay,
so
it
sounds
like
the
remote
urls,
don't
work
with
private
repositories,
and
this
person,
for
some
reason,
was
able
to
find
a
hacky
work
around
by
modifying
their
auto
completion
script.
B
Do
you
know
whether
we
ever
properly
worked
with
private
repositories
like?
Is
this
regression
or
effectively
a
missing
feature.
A
I
don't
I
don't
know
for
sure.
If
I
had
to
guess
I
would
say
it's
a
missing
feature.
The
helm
plug-in
at
least
the
way
it's
implemented.
Is
it
doesn't
use
the
local
credentials?
It's
it's
possible
that
the
get
code
doesn't
either
anna.
Do
you
know
you've
been
working
with
that
code.
C
D
It
will
just
pick
that
when
doing
get
get
clone
get
fat
or
whatever
good
operation,
which
case
it
will
just
work
and
in
case
of
a
https,
password
login,
then
I
guess
that
would
depend
on.
I
don't
know
even
how
that
one
works
personally.
A
B
I
guess
it
kind
of
falls
in
the
bucket
of
you
know
how
we
want
remote
handling
to
work.
We
have
some
other
requests
like
for
oci
support
and
other
stuff.
B
The
original
purpose
of
the
remote
feature
was
more
to
enable
debugging
and
we're
hoping
that
folks
would
use
the
upcoming
customized
localize
for
workflows
that
heavily
dependent
on
remotes,
but
it
seems
like
in
reality,
people
do
use
this
in
production
all
the
time,
so
there's
kind
of
a
tension
there
in
terms
of
enhancing
this
or
not.
A
I
do
think
it's
kind
of
an
arbitrary
line
to
draw
at
like
private
repos,
just
because
someone.
B
We
have
another
issue
here
that
I
found
will
paste
in
the
chat
that
somebody
is
complaining
about
that
prompt
at
all,
but
like
customize
in
general,
wants
everything
to
be
explicitly
declared
and
committed.
But
obviously
you
don't
want
to
do
that
with
the
username
and
password.
So
we
need
to
figure
out
some
other
solution.
B
Yeah
compared
to
other
things
that
we've
had
requested,
I
think
this
isn't
really
a
high
priority
for
us.
Maybe
it's
a
case
where,
like
we
could
document
that
it
doesn't
seem
to
be
working
very
well
and
that
we
would
like
to
have
a
better
solution,
and
if
somebody
who
needs
it
like
has
a
proposal,
then
then
we
could,
you
know,
review
the
proposal
and
potentially
accept
the
future.
A
A
A
B
A
A
Okay,
anything
else.
A
Okay,
I
might
just.
B
Re-Title,
like
improve,
I
don't
know
something
about
private
repository.
C
B
Yes,
specifically
about
the
default,
namespace
handling,
there's
actually
a
related
issue
that
was
opened
very
recently
with
some
tests.
Well,
it's
related
in
that,
if
it's
probably
caused
by
the
same
part
of
our
code,
where
we're
doing
something
wrong
in
the
cluster
role,
binding
handling
for
the
name
tracking.
Specifically,
I
see,
I
think
the
difficulty
we
had
with
this
one
is
that
it
seems
very,
very
niche
and
specific
what
they're
trying
to
the
behavior
change
that
they're
trying
to
affect
and
like.
B
I
don't
think
so.
I
probably
just
did
that
after
glancing
at
the
code
and
wanting
to
see
if
it
actually
worked,
you
know
sometimes
we
get
contributions
that
that
the
tests
don't
pass
or
that
you
know
they're
not
actually
really
ready
for
a
look.
That's
probably
all
I
was
doing.
A
B
Yeah
the
problem
is,
and
probably
what
we
said
in
the
last
meeting
is
that
we
would
look
at
it
in
more
detail.
This
there's
like
there's
a
lot
of
specifics
to
look
at
here,
which
is
difficult
to
do
quickly
on
a
call,
I
think,
I'm
having
that
same
problem
now
trying
to
figure
out.
You
know
what
happened
when
they
tried
our
suggested
solution,
what
they
actually
want
and
why
the
transformer
is
doing
something
different.
A
A
A
Because
I
also
still
need
to
write
up
our
home
policy,
so
I
can
look
into
it
and
then
see
if
it's
something
we
care
about
that.
We
want
to
support.
But
I
think
I
want
to
like
go
into
the
code
a
bit
to
see
how
this
is
getting
set,
and
I.
B
A
A
B
Actually,
it's
not
in
the
right
place
in
the
docs,
but
there's
already
a
little
bit
of
ducts
on
the
generator
here.
It's
only
in
the
advanced
built-ins
reference
instead
of
in
the
field
reference
by
the
looks
of
it,
but
it
does
have
the
values
listed
as
a
possible
field.
A
A
So
I
do
have
support
for
this
in,
like
google's
version
of
the
helm
function,
and
I
planned
to
add
that
to
the
the
one
we
have
in
the
six
click
rm
function
registry.
I
think
the
question
is:
is
this
something
we
want
to
add
to
customize.
B
I
guess
we
didn't
write
down
yet
what
the
policy
would
be,
but
this
does
seem
like
a
new
feature.
So
if
the
line
was
that,
like
we
were
just
gonna,
add
support
for
fields
that
were
not
passing
through
yet
and
fix
bugs
in
it,
then
I
think
this
would
fall
outside
that
scope.
A
Yeah
we're
not
trying
to
support
new
features.
I
will
I'll
write
up
our
home
statement
today
and
then
I'll
link
to
that
in
this
issue.
A
A
Yeah,
it
just
updated
the
field
spec.
I
have
no
objections
to
supporting
this.
B
A
Yeah,
it
updated
the
field
spec
for
the
prefix
suffix
transformer,
but
not
the
namespace
transformer.
It
looks
like.
B
That's
a
nice
easy
one!
If
someone
on
the
call
is
looking
for
an
issue
to
pick
up
the
it's
making
the
equivalent
change
to
the
one
in
the
other
pr
and
you
kind
of
get
to
see
how
the
transformers
are
configured.
A
A
B
A
C
A
B
A
B
Yeah
there
is
something
that
they
found.
It
seems
to
have
to
do
with
the
valid
string
detection.
It's
interesting.
A
A
B
A
A
A
Okay,
never
mind.
I
was
thinking
it's
just
for
high
level
fields,
not
subfields,
so
this
is
something
else.
Could
you.
D
B
It
looks
like
they're
they're
wanting,
merge,
behavior
and
that's
not
what
the
edit
command
is
doing.
I
haven't
worked
on
the
edit
command
personally
ever
so
you
could
be
wrong,
but
there's
a
report
in
the
middle
as
well
that
it's
not
even
merging
with
the
same
entry
like
if
you,
if
you
do
repeat
the
command
multiple
times
it
just
depends,
depends
depends.
B
Yeah
and
I
just
reproduced
it
does
still
happen,
the
new
tag
gets
removed
running
it.
Multiple
successive
times
doesn't
actually
append
to
those,
I'm
not
sure
what
they're
talking
the
other
person
in
the
middle
is
talking
about.
B
Yeah,
I
think,
that's
reasonable.
I
guess
the
one
thing
that
could
the
counter
argument
would
be
that
if
people
are
using
this
in
their
ci
and
they're,
expecting
it
to
overwrite,
because
it's
overwritten
forever,
then
it
could
be
unexpected,
but
edit
doesn't
have
the
same
guarantees
as
build,
and
I
think
that
the
merge
behavior
is
much
more
reasonable.
A
A
Sure
so,
if
you
open
this
issue-
and
you
comment
like
slash.
B
A
B
I
just
opened
up
the
edit
code,
sorry
to
go
back
to
the
other
one.
I
noticed
that
there
is
a
undocumented
way
to
preserve
the
tag,
so
I
think
this
might
not
be
an
issue.
After
all,
it
might
be
a
documentation
problem.
B
B
Sure
I
can
do
that
sorry
to
sorry
to
remove
the
contribution
opportunity.
A
There
are
lots
of
other
contributions,
opportunities.
A
B
Shouldn't
it
be
our
fault
like
we
would
be
finding
that
object
before
attempting
to
apply
the
patch
right
and
if
we
don't
find
it,
then
we
wouldn't
even
invoke
the
patching
library.
B
A
This
code
is
hard
to
follow,
but
it
looks
like
it's
trying
to
do
it
on
every
resource
like
it
passes
all
the
nodes
through
the
patch
filter.
A
Detail
and
maybe
doesn't
influence
whether.
B
Yeah,
I
guess,
do
we
want
to
output
a
warning,
or
should
this
actually
be
an
error
and
what
is
the
behavior
of
an
smp,
which
should
probably
be
the
same
like.
A
Yeah,
we
eventually
wanted
to
defecate
these
fields
right,
patches,
json6902
and
patches,
strategic,
merge
and
just
consolidate
on
patches.
So
we
could
see
what
the
behavior
is
on
patches
and
fix
that,
based
on
what
we
decide.
A
For
what
it's
worth
replacements
does
throw
an
error.
If
there's
like
no
matching
targets,
I
think
that's
good.
A
Do
we
have
any
concerns
about
someone
relying
on
this
behavior
and
it
breaking
their
workflow?
If
we
change
it.
B
Maybe
I
it
it
seems
like
a
unintended
behavior,
it's
not
going
to
change
it
in
a
way
that
they
won't
notice,
and
that
will
just
suddenly
be
wrong.
It
would
start
flagging
them
a
problem
with
their
configuration
that
wasn't
being
flagged
before,
because
it's
like
the
behavior
before
was
that
it
it
it's
doing
nothing
right.
It's
no
whopping,
it's
useless.
So,
theoretically,
the
fix
is
just
to
remove
it
from
it
can
fix
it.
Yeah.
A
A
Thanks:
okay,
I
think
this
one
is
a
good
contribution
opportunity
as
well.
If
anyone
wants
to
assign
it
to
themselves
audit
all
the
patches
fields
and
make
sure
they're
doing
the
right
thing.
A
Okay,
cool
thanks
everyone.
If
there
are
any
issues
that
you
would
like
to
work
on,
you
can
always
reach
out
to
katrina,
and
I,
on
slack.