►
From YouTube: Carvel Community Meeting - May, 10, 2021
Description
Carvel Community Meeting - May, 10, 2021
We meet every Monday at 11:30am PT / 2:30pm ET. We'd love for you to join us!
Agenda notes from this meeting can be found here:
https://hackmd.io/F7g3RT2hR3OcIh-Iznk2hw#May-10-2021-Agenda
A
A
A
For
anything
that
we
do
not
get
to
during
this
meeting,
we
do
have
karbal
office
hours,
which
are
held
every
second
and
fourth
thursday.
So
this
week
we
will
be
meeting
for
office
hours
at
11
30
a.m,
pacific
time,
2,
30,
p.m.
Eastern
time,
please
feel
free
to
join
those
as
well.
If
there's
anything,
you
need
any
sort
of
in-depth
help,
help
on
for
carnival
office
hours
or
just
want
to
attend
to
listen
in.
We
welcome
you
to
join
us.
A
Today's
date
is,
may
10
2021,
please
add
your
name
and
the
organization
you
represent,
or
if
you're,
just
representing
yourself,
we'd
like
to
keep
note
of
who
all
is
attending
these
meetings,
so
add
that
here
to
the
agenda
for
announcements
we
did
have,
we
did
participate
in
kubecon
europe.
Last
week,
virtually
we
had
carville
had
a
a
30
minute
talk
which
we'll
add
to
the
agenda
into
our
youtube
playlist
once
that
is
uploaded
to
the
cncf
kubecon
playlist.
A
A
B
B
It
became
include
non-distributable
layers,
so
if
you
were
using
and
this
flag,
please
update
your
scripts
when
you're
played
to
zero
seven
zero
and
we
also
added
support
for
environment
variables
key
chain.
So
you
can
provide
multiple,
usernames
and
passwords
for
multiple
registries
using
your
environment
variables.
So
those
are
like
the
the
big
chunk
items
here,
so
that
was
that.
A
C
C
E
C
D
D
You
can
do
that
or
data
values.
You
can
do
that
by
adding
your
schema
and
your
additional
templating
that
makes
use
that
that's
in
flight.
The
other
is
a
really
important
feature
around
being
able
to
turn
off
schema
checking
for
certain
portions
of
your
values.
Maybe
there's
some
reason
that
you
just
want
something
to
be
a
pass-through,
and
you
don't
want
any
checking.
That's
also
in
flight
with
with
these
we're
we're
really
looking
toward.
D
The
next
thing
is
starting
to
provide
some
base
documentation
for
the
overall
feature
and
and
starting
to
move
into
a
usability
testing
and
some
hardening.
So
we're
looking
forward
to
that.
That's
kind
of
exciting
that
we're
like
getting
to
the
point
where
we're
coming
in
for
not
quite
at
the
final
approach,
but
we're
getting
there
on
this
really
valuable
and
important
feature.
D
D
B
Not
really
update,
but
I
like
I'm
trying
to
get
for
next
thursday,
see
if
we
can
get
moving
along
our
rename
proposal.
So
I
haven't
decided
what
parts
of
the
proposal
that
we
need
to
look
at
right
now.
I
need
to
get
to
get
a
refresh
on
that,
but
if
there's
anything
on
that
proposal
that
you
have
any
questions,
just
let
me
know
so
that
we
can
pick
them
up
on
next
thursday
so
that
we
can
move
this
forward,
move
this
proposal
forward.
B
C
D
D
D
So
the
the
scope
of
this
is
to
to
craft
the
agenda
for
that.
The
usability
testing
here
and
from
there
we'll
know
like
what
the
exact
set
of
stories
in
terms
of
like
work
items.
We
need
to
actually
conduct
those
interviews,
so
this
this
spawns
out
a
little
bit
more.
So
it's
a
little
bit
of
a
kind
of
like
planning,
slash
exploration,
type
type
story.
If
you
will.
D
On
that
that
last
section
is
just
a
reminder
about
some
helpful
thinking
that
or
something
the
ways
of
thinking
about
that
conversation
that
we
have
when
we're
interviewing
users
that
tend
to
be
more
fruitful.
If
we
ask
people
questions
directly,
we
have
this
natural
tendency
to
wanna
give
each
other
the
answers.
We
think
that
you
wanna
hear
so
providing
a
little
bit
more
of
an
indirect,
providing
the
experience
to
occur
and
then
make
observing
the
the
usage,
as
it
happens,
tends
to
give
us
more
reliable
data.
So
that's
what
that's
about!
D
F
D
D
D
F
D
C
D
It's
going
to
be
different
from
data
value,
the
values
that
we
talked
about
today
and
to
be
clear,
I'll
refer
to
the
kinds
of
files
that
we've
been
referring
to
as
data
values
so
far
refer
to
those
as
data
value,
overlays.
That's
what
we've
been
typically
using
to
date
in
ytt,
to
specify
data
values,
they're,
actually
overlays,
and
in
contrast,
these
new
types
of
files
are
going
to
be
plain
yaml,
so
they
won't
have
any
annotations
in
them.
They'll
just
be
straight.
D
Yaml
they'll
combine
simply
so
that
means
that,
instead
of
having
rules
about
the
user
needing
to
disambiguate,
oh
did
you
mean
to
add
a
new
array
item
here
or
oh
this,
the
type
that's
being
set
used.
The
target
used
to
be
an
array.
Now
it's
a
map
like
you're
trying
to
put
a
map
on
there
are
you
sure,
like
we'll,
actually
just
largely
blindly
follow
whatever
the
new
document
will
do
so
arrays
the
values
for
arrays
will
get
overwritten
if
it's
a
completely
different
type.
D
The
new
type
that
shows
up
is
the
one
that
wins
sort
of
last
one
in
wins
kind
of
thing,
very
simple
combination
rules,
so
that's
sort
of
the
the
overall
what
it
it
sits
in
this
family
of
flags
there's
actually
a
set
of
flags
that
are
like
this.
The
data
values
flags-
you
might
have
recognized
these
as
the
ones
where,
when
you're
pulling
data
values
from
either
the
command
line
or
the
environment,
environment
variables
operating
system,
environment
variables,
those
are
the
the
prefix
for
all
of
those.
D
D
D
So
this
flag
provides
those
integrators
with
the
ability
to
extend
this
kind
of
input
to
their
end
users,
and
so
in
that
way,
like
the
their
end,
users
don't
have
to
learn
about
ytt
itself.
So
it
simplifies
the
story
that
an
integrator
needs
to
tell
to
their
end
users.
So
that's
one
motivation,
the
other-
and
this
is
this-
is
something
that's
sort
of
a
later
in
coming
is
focused
at
those
who
are
using
ytt
directly.
D
So
today,
one
of
the
biggest
sources
of
grief
for
most
of
our
new
users
is
in
using
data
values.
They
can
kind
of
get
out
the
gate,
but
the
moment
they
try
and,
for
example,
split
a
data
values
overlay
into
two
files.
It
starts
to
become
apparent.
They
aren't
just
dealing
with
files
that
combine
easily
they
get
these
error
messages
that
talk
about
a
mechanism.
They
didn't
even
know,
existed,
perhaps
overlaying,
and
so
it's
a
rough
experience.
D
D
So
if
we
pull
up
the
yeah,
so
we
see
that
the
backlog's
in
in
two
pieces
at
the
moment,
we've
got
a
couple
of
stories
that
we
can
point
today
and
then
we
have
a
set
that
I've
identified,
there's
even
more
they're,
going
to
be
identified,
pulled
out
of
the
proposal
and
and
crafted
and
we'll
be
able
to
point
this
in
we'll
tackle
these.
These
two
today,
like
this
one
first.
D
C
D
D
We
won't
be
concerned
and
for
now
the
behavior
will
be
just
as
it
gets
parsed.
Those
are
comments
and
they'll
get
ignored.
There's
a
subsequent
story,
further
down
the
backlog.
To
help
provide
better
experience
like
saying
hey
you,
you
included
a
ytt
annotation.
Here,
that's
not
allowed
that's
later.
D
G
G
D
D
B
D
B
B
Personally,
I
prefer
to
have
a
note
here
saying
that
you
can
only
have
this
flag
once
for
this
story.
Okay
and
then
we
can
handle
it
on
the
next
story.
If
you
want
to
add
more
because
in
the
end,
it's
just
like
a
cobra
function
that
we're
going
to
call
that's
a
little
bit
different,
so
I
think
we
can
handle
that
later.
B
C
B
B
G
G
F
G
D
D
B
B
D
G
G
D
Yeah
yeah
do
notice.
There
is
no
feature
flag
for
this,
so
schemas
would
be
turned
off,
maybe
and
if
they
are
turned
on
they're
they're
ready
for
prime
time
like
if
we've
removed
the
schema
experimental
flag
so
which
I
would
I
would
personally
expect
them
to.
There
isn't
any
reason
why
I
don't
think
that
they
would
combine
well,
but
it's
not
required
even
necessarily
think
about
that.
For
this
story.
D
D
G
D
D
D
G
I
would
I
would
have
put
a
two
just
for
the
code
changes
and
a
one
for
documentation.
I
think
adding
new
things
to
I
mean
I've
just
found
from
my
experience.
It's
just
really
hard
to
just
you
know,
make
code
changes
in
ytt,
it's
a
little,
usually
protected
and
learnings.
That
need
to
be.
They
need
to
happen
so
like,
while
it
might
seem
like
a
one,
I'm
just
adding
another
one
based
on
experience
and
then
another
point
for
documentation.
G
F
C
D
Right,
that's,
that's
a
that's!
A
sort
of
a
weird
thing
right
just
doesn't
show
up
all
that.
Often
let
me
respond
to
that
real
quick.
The
good
news
is
from
ytt's
perspective.
It's
not
it's
just
another
file.
So
by
the
time
the
pro
the
hour
pro
the
ytt
process
is
running.
It's
handed
a
file
descriptor,
just
like
any
other
file.
Descriptor
and
it
looks
like
a
file
it.
It
behaves
all
like
a
file
just
all
it.
It
is
a
file
actually
under
the
cover.
So
the
good
news
is,
it
turns
out.
D
It
has
zero
impact.
It's
a
it's!
A
opaque
interface,
there,
a
convenience
tool
for
a
shell
user
to
be
able
to
articulate
the
contents
of
what
should
be
in
the
shape
of
a
file
and
it
sort
of
coerces
it
into
that
shape.
So,
by
the
time
we
see
it,
that's
it
it's
notable
here
because,
as
you
can
see
by
the
example,
that's
actually
rather
readable,
and
so
somebody
somebody
could
technically
sort
of
use
this
in
this
way.
So.
B
Or
questions
or
not,
really,
I
think
this
is
like
a.
I
do
not
count
documentation
as
complexity,
but
I
think
feels
like
that.
The
process
is
going
to
take
like
it
feels
like
it
is
going
to
be
a
simple
thing,
but
you
never
know
with
ytt,
because
in
the
end
you
should
just
need
to
read
the
file
and
you
provide
this
information
like
the
way
we're
doing
is
we're
trying
to
find
a
an
annotation
to
say
that
the
file
is
a
values
right
now
we
don't
need
to
do
that.
B
D
All
right,
I'm
hearing
a
rough
consensus
about
two
great
chatter:
okay,
so,
let's,
let's
pause
for
just
a
second
to
to
kind
of
take
that
fly
over
together.
So
the
first
story
that
we
talked
about
here
is
just
accepting
one
file.
Just
like
we
talked
about.
You
got
that
baked
in
your
head,
and
here
we
go.
So
what
about
multiple
files
boom
next
story-
and
this
one
is
gonna-
include
either
you're
you're.
D
Invoking
the
data
dash
values,
dash
file,
flag,
multiple
times
and
or
a
given
file
has
multiple
yaml
documents
inside
it
all
of
them
plain,
yellow.
So
that's
the
scope
of
that
story.
So
that's
now
expanding
out
the
set
of
inputs
and
the
third
one
is
how
how
do
these?
How
does
this
feature
interact
with
data
values
overlays?
So
what
we
normally
think
of
as
data
values,
but
we're
going
to
start
trying
to
use
this
terminology,
the
data
values
overlays?
How
do
they
interact
and
the
tldr
there
is
these
act.
D
D
We
don't
want
someone
to
think
that
they're
providing
startler
code,
like
some,
if
statement
or
for
loop,
that
would
blow
up
blow
out
into
sets
of
values
that
should
be
data
values
and
it
just
get
ceremoniously
ignored.
Because
it's
a
comment,
so
we
want
to
be
really
clear
that
this
we
will
not
be
evaluating
anything
inside
this
file,
we're
just
parsing
it
the
best
way
to
do.
That
is
just
to
say
we
don't
accept
any
templating,
and
so
that's
what
this
story
is
about
there.
D
D
D
So
when
you
do
a
library
get
and
then
go
to
evaluate
the
the
data
values
that
showed
up
either
from
a
data
values,
overlay,
there's
a
library,
slash,
ref
annotation,
you
could
say
here
configure
this
dependency
of
the
root
library.
It's
like!
Oh
okay,
I'll,
not
put
those
toward
the
root
but
I'll
target
those
for
that
dependency
and
configure
that
dependency
gives
the
end
user
the
ability
to
reach
in
and
drop
data
values
there.
B
B
D
D
D
B
A
C
D
D
D
D
Okay,
yeah
so
good
this.
This
is
important
kind
of
subtopic,
that's
sort
of
implied
in
the
spec,
but
not
explicit.
So
let's
make
it
explicit.
So
one
of
the
mo
one
of
the
important
things
remember
when
I
talked
about
how
data
values
files
are
going
to
combine
more
simply
so
how?
How
will
that
work
so
in
in
the
case
of
arrays,
we're
actually
going
to
be
replacing
the
values
as
a
spect
here,
and
so
you
might
wonder.
Well
how
might
we
achieve
that?
D
So
this
is
one
of
those
moments
where
I
suspect
I
mean
it's,
it's
it's
going
to
be
up
to
those
who
are
picking
up
the
story
stories
and
working
on
them,
but
I
strongly
suspect
that,
given
we
have
a
mechanism
that
for
combining
structures
that
is
overlays
that
we
might
want
to
use
them
as
an
implementation
detail
for
this,
that
is
we
might.
We
might
want
to
continue
to
reuse
that
mechanism.
D
But
if
that's
the
case,
then
like
joel
you're,
pointing
out
it's
like
well
wait
a
second
that
doesn't
do
replace.
So
that's
a
good
question.
There's
some
there's
some
details
in
here
to
like
out
about
will
perhaps
perhaps
either
we
need
to
find
a
way
of
translating
these
files
into
templates
and
evaluating
them.
D
C
D
D
Somewhere
in
between
there
there
we
are
doing
these
kinds
of
operations
right
now
today.
So
it's
not
like
it's
from
scratch,
but
like
if
you
were
to
just
try
and
overlay
these
two
things
without
doing
anything
special
right
now
it
wouldn't
exactly
work
like
if
you,
if
you
said
okay,
this
this
is
overlaying.
On
top
like
take
this
new
turn
turn
the
data
values
file
that
got
provided
and
then
say
here,
you're
an
overlay
like
just
drop
on
this.
It's
gonna.
G
Needs
work,
it's
a
couple
of
different
implementation
approaches
and
then
there's
like
examples
that
I
can
think
of
that.
Don't
fit
completely
these
three
dots
and
I'm
just
going
to
keep
them
as
we'll
find
we'll
find
our
rules
for
these
examples,
maybe
through
the
proposal
by
just
asking
the
team
and
I'll
keep
in
mind
unknown
complexity.
D
I
was
like
large
three
small
five
in
my
in
my
book
like
feels
like
there's
a
lot
of
stuff
that
exists,
but
like
pulling
all
that
together
and
like
maybe
a
little
bit
of
expertise,
there's
some
unknown
about
like
overall
approach
and
so
inherent
in
that,
I
think,
is
kind
of
sussing
out
like
which
approach
might
be
better.
That's
some
work
in
addition
to
the
documentation
and
yeah
stuff.
C
A
A
Okay,
great,
thank
you
so
much
all
right.
Well,
thank
you.
Everyone
for
joining
today's
community
meeting
for
carvel.
If
you're
unable
to
attend
our
monday
community
meetings,
we
do
meet
every
monday
at
11
30
a.m.
Pacific
time
we
also
have
our
carville
office
hours,
which
we
meet
every
second
and
fourth
thursday
of
the
month
at
11.
30
am
pacific
time.
2
30
p.m.
A
Eastern
time,
and
we
will
be
meeting
this
week
on
may
13th
and,
like
I
just
mentioned
one
of
the
items
that
we
weren't
able
to
get
to
today
will
be
moved
to
that
office
hours
for
further
discussion,
and
if
you
have
anything
that
you
need
help
with
from
the
carnival
maintainers
or
have
any
questions
regarding
the
tool
suite.
Please
join
us
for
this
office
hours
or
the
community
meeting
or
even
join
us
in
slack
on
the
kubernetes
public
workspace
in
the
carville
slack
channel.