►
From YouTube: Carvel Community Meeting - August 12, 2021
Description
Carvel Community Meeting - August 12, 2021
We meet every Thursday at 10:30am PT. We'd love for you to join us live!
Today's meeting covered kapp-controller blog post, Carvel KubeCon keynote announcement, GoDocs, and a question from a community member about kapp-controller, Tanzu package plugin, values.yaml files and annotations (#@ and #!). Full details here: https://hackmd.io/F7g3RT2hR3OcIh-Iznk2hw
A
All
right,
hi,
everyone
welcome
to
this
week's
edition
for
the
carville
community
meeting.
We
were
out
last
week
so
we're
back
this
week
after
a
long
week
with
so
much
needham,
r
and
r
for
the
team
and
we're
excited
to
be
back
with
our
meetings
that
are
held
every
thursday
10
30
a.m.
Pacific
time.
A
So,
if
you're
watching
this
from
home,
we
would
love
to
have
you
join
us
live
if
you
aren't
able
to
join
us,
live
and
have
any
questions
you
want
to
bring
to
the
team
we
can
be
reached
out
on
slack
and
twitter.
You
can
find
that
information
here
on
the
agenda.
A
A
Today's
date
is
august,
12th
2021.,
when
you
attend
the
meetings,
we
ask
that
you
add
your
name
and
any
affiliation
that
you're
associated
with
this
helps
us
keep
track
of
who,
from
outside
of
the
team,
is
joining
these
and
we
can
keep
those
lines
of
communication
open.
A
We
would
love
for
you
to
read
over
this
blog
post
and
provide
any
feedback
you
may
have,
or
if
you
have
any
questions
on
how
to
get
started
with
this,
please
feel
free
to
reach
out
to
us.
On
slack,
we've
also
created
a
lovely
tutorial
on
catacota
that
will
help
you
get
started
and
sort
of
walk
you
through
the
entire
tool.
A
Next
up,
if
you
are
part
of
the
kubernetes
community,
you
saw
that
kubecon
cloudnativecon
north
america's
schedule
was
released
last
week.
A
Part
of
that
schedule
is
carmel
is
on
the
keynote
stage,
we'll
be
on
the
keynote
stage,
friday
october
15th
from
9
40
to
9
45
a.m.
Pacific
time,
there's
a
link
here
for
that
particular
talk,
it'll,
be
given
by
shadow
rupa,
nandi
she's,
the
engineering
director
at
vmware
tanzu,
and
it's
just
a
five
minute
keynote
covering
carvel
and
all
sorts
of
fun
stuff.
With
the
tool
suite,
you
can
register
to
attend,
kubecon,
either
virtually
or
in
person.
A
So,
if
you're
not
able
to
attend,
given
the
times
that
we
are
dealing
with
now,
there
is
an
option
to
watch
this
virtually.
A
All
right,
and
as
always,
if
you
have
anything
that
you
wish
to
discuss
in
this
meeting,
please
feel
free
to
put
that
down
in
this
discussion.
Topic
section
like
we
have
a
few
already
posted,
but
if
there's
anything
that
you
need
help
with,
please
feel
free
to
add
that
there.
B
Yeah
private
registry
authentication
is
very
much
top
of
our
priority
list
right
now
and
we're
actively
working
on
it
timeline
of
august
is,
I
think,
still
still
realistic,
probably
maybe
closer
to
the
end
of
august,
but
but
yeah.
We,
we
did
recently
submit
a
proposal
that
that
covers
this
in
a
bit
more
detail.
B
So
if,
if
there
are
any
questions
about
what
this
is
about
or
how
we're
planning
to
implement
it,
please
check
out
the
proposal.
B
Yeah
so
it's
number
208.
A
Just
add
that
my
tabs
will
be
I'll.
Add
it
to
the
notes
in
the
moment
and
then
anything
regarding
the.
C
Yeah
yeah
go
for
it.
I
could
call
out
two
ytt
stories
that
have
been
done
in
this
last
two
weeks,
just
kind
of
filling
the
gap
of
certain
error
messages
that
had
come
out
of
our
schema
work
that
we're
missing
information.
C
So
we've
done
our
best
to
populate
these
error
messages
with
the
information
that
we
had
available,
the
best
that
we
could
so
hopefully
we
can
get
a
release
out
with
these
changes
before
working
on
the
next
step
in
ytt
or
the
next
goal
in
ytt,
which
is
the
open
api
schema
track.
B
I
think
there
have
been
general
other,
I
don't
know,
either
bug
fixes
or
smaller
enhancements
that
have
been
the
focus,
but
then,
as
garrett
mentioned,
probably
the
next
big
track.
That's
about
to
get
started
is
generating
open
api
schemas
in
ytt,
and
this
is
a
follow-up
from
the
recently
released.
Ytt
data
value
schema
functionality
that
was
released
in
july,
but
this
is
just
taking
it
one
step
further
to
be
able
to
generate
a
more
compatible
form
of
the
schemas
so
that
other
consumers
can
can
easily
integrate
with
it.
A
Right
now,
if
a
person
were
to
go
through
the
in-progress
issues,
would
that
give
them
a
better
picture
of
other
items
that
maybe
the
team
is
working
on
outside
of
the
roadmap.
B
A
D
Yes,
for
the
go
doc,
I
would
like
to
share
my
screen,
so
I
think,
maybe,
a
month
ago
we
kind
of
talked
about
godoc
and
we
said
well,
you
know,
let's,
let's
try
it
in
a
couple
places
first
and
so
and
there's
maybe
some
questions
of
like
well.
D
What's
this
really
gonna
look
like
so,
I
wanna
show
you
as
part
of
the
package
repository
work,
we're
working
on
secret,
gen
controller,
and
this
is
a
nice
opportunity
because
it's
relatively
small
and
it's
not
being
relied
on
in
the
kind
of
critical
path.
Yet
by
anything,
so
we've
we've
been
able
to
go
through
and
and
I'm
you
can
see,
I'm
serving
this
locally,
so
the
command
to
serve.
D
It
is
just
godoc,
you
know
with
a
with
a
port,
and
so
you
know
you
can
kind
of
see
like
look
this.
This
is
really
what
it
looks
like
like
this
is
you
know
documentation
that
I
kind
of
convinced
dimitri
to
write
yesterday,
and
so
it's
you
know
each
each
of
these
exported
functions
has
a
brief.
You
know
description
of
itself
now
or
many
of
them
do,
and
you
know
the
like
this
warm-up
function.
D
Maybe
you
don't
know
what
it
does
and
then
you
can
say:
oh
okay,
it's
it's
hydrating
a
cache,
so
this
anyways,
I
thought
I
would
give
a
quick
like
demonstration
of
what
this
looks
like
in
in
practice,
and
I
wanted
to
ask
folks
on
the
call,
if
we're
ready
to
start
expanding
this,
I
guess
the
other
thing
I
should
say
is
that
the
way
that
we
were
able
to
turn
it
on
is
you
you
put
a
git
sha
into
your
linter
config,
and
you
say
from
this
from
this
sha
after
this
sha,
all
new
diffs
were
actually
gonna
require
require
godoc,
so
the
the
linter
will
require
documentation
for
all
new
changes
and
that
way,
there's
no
big
sweeping
cut.
D
You
have
to
do
you
just
start
documenting,
so
I
wanted
to
take
a
poll
of
like
comfort
level
of
like
starting
to
kind
of
ratchet
this
forward
in,
like
some
of
our
other
repos.
I
think
folks
here
today
is,
I
think
ytt
is
especially
well
represented.
D
So
I
don't
know
you
know
garrett.
John
any
thoughts.
D
That's
a
good
question,
so
I
think
the
the
linter
is
satisfied.
If
you
have
you
know
the
name
of
the
exported
symbol
followed
by
almost
anything-
and
I
hopefully
like
my
my
goal-
is
that
it
should
be
more
informative
than
just
reading
the
the
function
name
right
and
sometimes
the
function
name
is
very
explanatory
and
that
there
really
isn't
that
much
to
say
but
like
sometimes
there's
at
least
a
synonym
that
you
can
that
you
can
put
in.
So
I
think.
D
I'll
I'll
give
one
quick
example.
So
this
this
function
match
secrets
for
import,
and
you
know,
especially
if
your
head
is
in
secret,
gen
controller
land,
you
might
say.
Well,
obviously
I
have
an
import.
D
It's
going
to
return
an
array
of
secrets,
it's
going
to
return
an
array
of
secrets
that
match,
and
maybe,
if
your
head
is
not
in
secret,
gen
controller
land,
it's
going
to
be
helpful
to
see
that
it
filters
secrets
that
are
exported
that
are
in
our
export
cache
and
that
it
filters
them
by
this.
This
matcher
criteria,
so
to
me
like
like
this,
is
a
a
good
example
of
like
a
minimum
value.
Add
does
that
is
that
helpful.
E
D
Any
other
input.
C
D
Well,
cool
cool!
So
I'm
happy
to
pair
briefly
or
point
you
all
to
the
the
relevant
ci
config
and
then
I
think,
that's
kind
of
a
segue
into
a
broader
discussion
like
I
know,
I've
been
working
in
the
last
couple
weeks.
Sorry
nancy!
D
This
is
the
next
item
on
the
agenda:
okay,
to
to
pull
in
like
some
of
the
like
image
packages
using
go
releaser
and
image
package
and
I
think
ytt
are
using
like
or
testify,
and
I
I
was
kind
of
wanted
just
to
take
a
temperature
of
the
group
of
like
how
much
effort
should
we
be
putting
into
all
using
the
same
tools
or
same
libraries
like
this.
That's
that's
my
question.
E
All
right
yeah,
I
have
an
opinion
so
so
not
nominally
if
we
can
get
to
agreement
about
like
what
a
baseline
set
of
tools,
because
we
are
under
an
umbrella
as
a
set
of
projects,
that's
obviously
valuable.
It
means
that
people
can
contribute
to
each
of
the
tools
without
having
to
learn
a
whole
different
set
of
things.
So
that's
sort
of
that
that's
a
kind
of
a
typical
argument
for
standardized,
but
personally,
I'm
always
open
to
hey.
E
Maybe
there's
a
particular
need
on
on
a
project
and
wouldn't
be
pedantic
about
well
yeah.
We
we
use
this
other
framework
or
additional
framework
or
library
whatever,
because
for
cause
personally,
but
it'd
be
great
if
we
can
just
kind
of
like
there's
a
lot
of
the
kind
of
programming
that
we're
doing
that
fits
within
that,
but
that
are
sufficiently
similar.
It'd
be
surprising
that
we'd
have
to
have
completely
different
sets
of
tools,
so
it
feels
like
we
could
probably
lean
into.
E
Maybe
it's
just
a
matter
of
just
putting
attention
to
it
and
then,
rather
than
having
be
a
little
bit
more
intentional,
slightly
less
organic
and
we
probably
get
there
pretty
quickly.
Is
my
guess.
D
And
I
guess
maybe
one
more
follow-up,
because
I
think
I
broadly
agree
with
how
you
put
it
john
and
one
more
follow-up
is
like
I've
just
listed
off
a
handful
of
things
right.
This
around
godok
go
releaser
the
way,
we're
doing
assertions
in
our
unit
test.
E
B
One
thing
that
isn't
necessarily,
I
guess
a
library
but
is
github
actions
and
trying
to
keep
those
consistent,
might
be
useful
and,
like
secret
giant
controller,
comes
to
mind
since
we
we
haven't
spent
as
much
time
there
so
like
that.
That
might
be
one
that
we
want
to
bring
up
to
speed.
D
So
that
yeah,
I
know
like
dimitri-
and
I
just
actually
went
through
some
of
that
yesterday-
we
basically
copied
things
out
of
cap
controller,
but
I
don't
know
honestly
whether
cap
controller
was
up
to
speed
with
some
of
the
older
tools.
A
D
So
I
think
maybe
I
I
can
take
point
on
that
and
just
just
roll
in,
like
here's,
here's
our
standard
kind
of.
If,
if
you
were
to
start,
you
know
one
more
carvel
tool:
here's,
hopefully
what
it
would
be
leaning
on
here's
our
stack
right.
D
Yeah
I
like
that,
a
lot
okay,
so
I
I
can.
I
can
own
the
next
step
there
and
bring
a
proposal
back
around
and
then
okay.
So
unless
there's
any
other
thoughts,
we
can
move
on
thanks
all.
A
Well
thanks
joe
nick
you're
up,
you
got
a
question
about
cap
controller
tons
of
package,
plugin
values
all
that
good
stuff.
F
Yeah,
so
I
I
think
that
this
might
be
a
change
that
came,
I
guess
in
a
more
recent
version
of
cap
controller,
but
so
when
I
have
a
package-
and
I
install
it
into
you-
know
the
cluster
using
using
the
you
know
the
tons
of
package
plugin
everything
works,
just
fine
and
obviously,
inside
of
that
package,
you
know
we
have
a
values.yaml
file,
which
is
where
we
can
specify
some.
F
You
know
you
can
then
specify
a
dash
dash.
You
know
values
file
equals
and
then
give
it
the
the
name
of
your
your
now
local.
You
know
custom
values,
file
for
that
package
and
when
it
goes
to,
you
know,
install
the
package.
If,
if
that
values
file
has
any
lines
in
it
that
contain,
you
know
the
directives
like
or
any
line
starting
with,
you
know
the
hash
at
sign
or
hash
bang.
You
know
for
a
comment.
F
The
the
installation
of
the
package
will
fail.
You
know
basically
saying,
oh,
my
god,
there
are
these
these
lines.
I
can't
deal
with
that.
So
I
guess
it's
like.
What
I
don't
understand
is
is
that,
if,
if
I,
if
I
just
install
the
package
and
that
values
file,
has
those
annotations
in
it,
it
works.
But
if
I
make
an
exact
copy
of
that
file,
I
don't
even
have
to
make
any
changes.
F
G
G
We
didn't
want
the
secret,
contain
the
values
to
have
to
know
what
sort
of
back
end
the
package
is
using
for
templating.
So,
in
order
to
do
that,
we
wanted
to
expose
a
generic
interface
of
just
plain
yaml,
and
then
that
gets
passed
through
to
the
tool
and
then
I
don't
know
john.
If
you
want
to
talk
about
the
ytt
design
decisions
around
disallowing
annotations
in
that
file
with
it.
E
Yeah,
so
the
the
idea
is
yeah.
First
of
all,
I
just
want
to
say
I
can.
E
I
can
relate
to
like
how
that
ought
to
just
ought
to
just
work,
and
it's
frustrating
that
it
doesn't
that
data
dash
values.file
flag
when
we
added
that
we
intended
that
to
be
like
a
new
kind
of
input,
and
it's
it's
it's
just
it's
it's
another
take
on
data
values,
and
it
was
inspired
by
our
need
to
integrate
with
that
integrators
like
cat
controller,
would
want
to
use
and
and
and
part
part
of
the
motivation.
For
saying
this
must
be
a
plain
yaml
file
rather
than
a
ytt
template.
E
Part
of
it
is
exactly
what
eli
said
was
that
we
want
to
make
sure
that,
as
a
package
author,
you
don't
inadvertently
run
into
this
problem
where
you've
exposed
your
community
of
user
who
are
using
your
package
to
like,
have
a
ytt
template
as
your
data
values,
or
it's
actually
it's
technically
a
ytt
overlay
as
your
data
values
and
then
for
whatever
reason
you
want
to
switch
what
tool
you're
using
on
the
back
end
to
render
and
that
you
want
to
switch
to
say,
helm
in
your
next
version.
E
That
would
be
a
breaking
change
for
everybody
using
your
package.
So
in
part
it's
an
attempt
to
do
that.
The
the
other
part,
too,
is
that
it
simplifies
the
whole
providing
data
values.
Experience
minus
this
bit
that
you
described
so,
unfortunately,
you're
hitting,
like
the
other
end
of
a
trade-off
and
the
simplification,
is
this
that,
for
a
lot
of
users
coming
to
ytt
in
general,
it's
a
it.
It's
a
bit
odd
and
confusing
for
the
first
thing
out.
The
gate
when
I'm
providing
configuration
values
is
to
supply
this
file.
E
That
sort
of
looks
like
it's
just
a
bunch
of
key
value
pairs,
but
turns
out
it
does
more
than
that,
like
I
can't
add
two
of
them
together,
like
they
don't
combine
well
at
first
and
then
I
get
these
overlay
errors.
I
don't
even
know
what
what
do
you
mean
overlay
as
a
new
user,
and
so
what
we
wanted
to
do
was
provide
a
much
more
straightforward
interface
for
providing
data
values,
and
so
so
so
the
the
the
the
kind
of
input
that
we're
getting
on
that
flag
is
a
plain
yaml
document.
E
So
I
know
that
that's
not
very
satisfying.
It's
like
it's
okay,
so
it's
not
broken.
This
is
by
design.
I
I
know
it's
awkward.
I
think
our
hope
is,
and
I
and
I
think
I
might
be
kicking
this
back
to
you
eli.
I
hope
it.
I.
E
I
believe
that
our
intent
is
to
be
able
to
give
you
the
ability
to
provide
a
data
values
overlay
feature
so
that
you
can
like,
like
provide
a
data
values,
overlay
file
in
the
way
that
you're
you're
doing,
and
that
would
start
to
feel
more
natural,
because
it
would
actually
interpret
that
as
a
dynamic
file,
a
file
that
has
expressions
inside
of
it,
as
opposed
to
just
a
plain,
yellow
file.
G
Yeah,
no
that's
true.
We
added
an
annotation
to
the
package
install
to
sort
of
swap
the
type
of
input
plot
input
file
from
this
plane
data
values
file
to
a
the
overlay
kind.
So
I
can
link
you
to
some
docs
around
that.
E
F
Yeah,
no,
I
guess
it's
good
to
hear
that.
I
I
guess
I
can
understand
why
it
was
happening.
I
guess
I'd.
My
biggest
thing
was
I
didn't.
Maybe
I
wasn't
paying
attention.
I
didn't
know
that
that
was
that
that
change
had
had
popped
in,
and
so
it
was
it
kind
of
threw
me
for
a
loop.
When
you
know
I
got
these
errors
and
it,
I
guess
the
more
painful
bit.
F
I
guess
is
that
now
for
all
of
the
packages
that
we've
currently
authored,
it's
like,
I
got
to
go
back
and
change
all
the
data
files
and
I
got
to
strip
out
all
the
all
the
annotations,
because
if
in
the
future,
if
somebody
was
doing
like,
I
did
and
just
said
hey,
I
just
want
to
see
this
work.
You
know
and
I'm
just
well,
I'm
just
going
to
take
the
existing
data
file
and
I'm
not
going
to
change
it.
I'm
just
going
to
pass
it
in.
I
would
expect
that
to
work.
F
So
you
know
if
we
want
to
provide
a
a
good,
I
guess
default
format
for
a
file.
You
know
it's
like
it
should
be
something
that
actually
works.
E
E
Could
you
please
not
run
the
whole
process
but
stop
right
after
you've
calculated
all
the
data
values
and
give
me
those
and
that
spits
out
plain
ammo?
I
wonder
if
that
might
be
a
and
a
tactic
that
we
want
to
bring
some
attention
to
for
someone
who's
trying
to
do
exactly
what
you're
doing
it's
like
hey.
I
just
I
just
want
a
starter
set.
Then
I
want
to
modify
I'll
prune
it
down
to
the
stuff
that
I
just
need
to
customize.
G
E
My
problem,
okay,
so
you
know
the
other
thing.
The
other
thing
to
hear
nick
is
just
a
little
bit
of
narration
stepping
taking
one
step
back
like
as
we're
building
up
these
tools
and
we're
adding
in
the
required
features.
So
one
thing
that
we've
just
recently
released
is
the
ability
in
ytt
to
define
a
schema
which
effectively
that's
what
your
data
values
inside
your
package
is
it's
it's.
It's
essentially
setting
the
schema
for
inputs.
It's
like
an
api
into
that
config.
E
E
As
the
author,
you
write
your
schema
file
as
a
end
user,
providing
your
configuration,
you
should
be
able
to
query
to
see
what
data
values
have
been
declared
in
the
schema
so
that
I
can
write
my
custom
configuration
and
then
provide
that
as
a
plain
yaml
data
values
file
or
as
like
eli
described,
there's
this
this
more
advanced
way
of
like
no.
I
want
to
use
an
overlay.
I
don't
want
to
just
provide
values.
I
want
to
do
something.
That's
more!
That's
fancier!
E
So
I
I
feel
like
your
pain's
real,
because
it's
it
I
mean,
I
know
it
is
like
you're
experiencing,
but
but
the
what
I'm
trying
to
say
is
there's
this
speed
bumps
in
this
experience,
because
we're
we
haven't
yet
been
able
to
like
show
clearly
show
here's
how
you
would
do
this.
Here's
this
use
case
you
want
to.
You,
want
to
do
x
and
you're
having
to
sort
of
feel
your
way
through
it
yourself.
E
So
I'm
I'm
sensing
a
gap
here
in
how
we
have
described
how
to
use
these
tools
in
concert
that
if
we
had
that
guide,
you'd
be
following
that
guide
and-
and
this
would
be
less
confusing,
frustrating.
F
Yeah
all
right,
well
I'll,
I'm
gonna
be
working
with
packages.
You
know
over
the
next
week
getting
some
stuff
ready
for
for
my
project,
so
I'll
I'll,
keep
thinking
about
this
and
I
might
hit
you
up
offline.
E
Yeah,
if
the
second
you
run
into
any
anything
like
if
please
swing
by
the
the
slack
channel,
we're
super
happy
to
pair
like
over
zoom
or
whatever
just
kind
of
get
things
started.
You
don't
have
to
figure
this
out
on
your
own,
we're
totally
here
to
help
cool.
Thank
you.
Yeah.
A
Okay,
anything
further
on
that
topic.
A
Okay,
it
looks
like
there
aren't
any
more
discussion
topics
listed
out.
Is
there
anything
else
that
anyone
wishes
to
bring
up
before
we
depart.
A
Nope,
okay,
I
will
thanks
everyone
for
joining
us
today
at
the
carville
community
meeting
again,
if
you're
watching
this
from
home,
we
encourage
you
to
come
and
join
us
live
we
meet
every
thursday
at
10,
30
a.m,
pacific
time-
and
it's
just
an
opportunity
for
you
to
listen
in
on
what
the
team
is
working
on,
provide
any
discussion
topics
you
wish
to
bring
up
with
the
team
or
if
you
need
help
with
something
like
nick
did
today,
you
can
bring
that
up
in
this
meeting
and
have
live
feedback.
A
If
there's
anything
that
you
wish
to
bring
up
and
you're
not
able
to
attend
the
meeting,
you
can
just
do
so
in
slack
or
on
twitter
and
we're
also
available
there
as
well
with
that
we
look
forward
to
engaging
with
you
and
have
an
enjoyable
rest
your
day.
Thank
you.