►
From YouTube: 2020-09-28 Crossplane Community Meeting
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
All
right
the
recording
has
started-
and
this
is
the
september
28th
2020
cross
playing
community
meeting,
so
we
are
working
towards
the
0.13
release.
So
let's
start
the
conversation
about
that.
While
I
let
more
people
into
meeting.
B
A
Yeah
so
let's
bring
up
the
that's
your.13
board
and
folks
have
been
working
quite
hard
recently
and
getting
a
lot
of
features
in
that
we'll
we'll
talk
through
here,
nick
or
phil
does
one
of
you
kind
of
want
to
give
like
an
overview
of
the
board
before
we
drive
into
you
know
some
of
the
more
specific
features
later
on
here.
C
I
can
give
you
an
overview.
I
I
yes
are
you?
Okay?
If
I
give
you
another
view
of,
what's
still
open
for
0.13,
that's
perfect!
Yes,
absolutely
the
issues
yeah,
so
my
understanding
is
that
the
key
issues
that
we're
still
working
on
as
we
said
that
the
goal
is
to
release
one
v
0.13
on
thursday,
so
we're
very
much
hoping
to
go
into
qa
mode.
Slash
already
have
mostly
got
into
qa
mode
this
week
for
for
just
testing
the
release
and
updating
documentation.
C
The
big
open
things
that
we
have
left
to
get
the
release
ready
are
having
a
cuddle
plugin
that
will
allow
you
to
build
and
push
new
crossplan
packages.
So
if
you
need
to
package
a
provider
or
package
composite
resources,
configuration
for
crossblade,
you
would
use
that
cli
tool,
casey's
working
on
that,
and
has
it
just
about
done.
I
believe
another
thing
that
we
noticed
somewhat
last
minute
is
that
there
are
certain
use
cases
there.
There
is.
C
There
are
some
issues
around
deletion
of
composite
resources
in
obscu
when
there
are
obscure
relationships
between
the
composed
resources
and
how
they
depend
on
each
other.
So
we're
trying
to
get
that
sorted
out,
but
I
are
working
on
that
for
the
release.
C
C
The
s3
bucket,
as
you
can
see,
there
is
still
open,
and
I
would
love
if
christian's
around
at
some
point
to
to
get
a
forecast
as
to
whether
that's
likely
to
be
wrapped
by
thursday
or
not,
and
that's
about
it.
I
think.
A
Yeah
awesome
yeah
one
thing:
I've
noticed
in
the
last
week
or
so.
Let
me
check
the
chat,
make
sure
I'm
not
missing
anything.
Oh
okay.
Yes,
yes,
we
do
have
a
youtube
channel.
We'll
drop
a
link
to
it.
Dave.
A
Drop
a
link
to
the
youtube
chat
across
my
youtube
channel
into
the
zoom
chat.
If
you
don't
mind
yeah,
so
one
thing
I've
noticed
the
past
week.
That's
been
really
awesome.
I
think,
is
that
as
we're
qaying
and
as
we're
developing
things
and
finding
some
some
issues,
we
are
doing
a
really
good
job.
I
think
of
capturing
those
like
the
behavior
that
we're
seeing
you
know
how
we
should
mitigate
it.
Getting
discussion
going
around.
A
You
know
how
to
deal
with
it,
so
I
think
that
there's
been
some
really
good
convergence
and
transparency
visibility
into
what's
going
on
and
like
what
we're
working
on
as
we're
the
problems
we're
finding
as
we're
going
towards
the
release
here.
So
that's
that's
been
really
cool,
like
good
good
job
to
everybody
for
doing
that.
A
And
also
dick
thanks
for
updating
the
kind
sh
script
with
a
better,
a
better
solution.
I
think
that
I
had
locally
put
together.
D
Yeah
and
then
for
0.13,
so
nick
and
ryan,
you
guys
agreed
that
you're
pulling
ohm
0
1.
is
that
right,
because
I
was
just
cut
a
little
while
back
and
we're
just
going
to
pull
that
in.
C
There
is
ohm0111
does
have
a
an
rbac
roll
binding
to
costa
admin
that
I
would
like
to
change,
which
I
presume
will
require
a
new
release
of
ohm,
probably
just
a
patch
release,
because
it's
just
an
our
back
roll
change.
But
if
ryan's
with
that
code,
I'm
uncomfortable
shooting
that
as
well.
B
Yeah
sure
yeah,
we
we
add
some
we
in
the
meantime,
we
added
some
bug
fixes.
We
could
just
that.
We
can
just
take
that
along.
C
B
C
Yeah
what
whatever's
whatever's
in
the
code
base?
Oh
I'm
happy
to
take
you
to
the
code
0.11
or
some
new
bug
fixes
as
well.
D
And
when,
when
you
roughly
think
we'll
be
able
to
do
that
like
before
wednesday
or
because
we're.
D
C
C
Yes,
so
it
would
be
ideal
to
have
the
rm1
released
on
on
wednesday
just
to
give
us
a
bit
of
buffer
there.
I
haven't
had
a
chance
to
go
and
update
that
cluster.
Yet
so
I'm
probably
the
main
blocker
at
this
at
the
moment.
A
Awesome,
that's
that's
really
great
to
have
that
come
together
and
it's
are
there.
Changes
about
like
any
of
the
home
charts
or
anything
for
for
the
ohm
to
be
deployed
is
like
a
long
incompatible
with
0.13,
or
is
that
all
in.
E
And
figured
out
yeah
I'm
going
to
update
to
sub,
chart
it
and
then
there's
a
few
other
home
updates
as
well,
one
of
them
being
installing
the
new
providers
and
that
sort
of
thing
so
there'll
be
a
few
kind
of
like
additions
to
it.
But
they
shouldn't
be
very
major.
A
Awesome,
so
that's
all
that's
part
of
the
plan
here.
It
looks
like
then.
A
Very,
very
great:
okay,.
D
My
question
on
the
0.13
so
with
the
new
home
version
and
things
moving
out
like
who's
going
to
tackle
the
docs
for
the
getting
started.
Is
that
something
ryan
that
you
all
can
help
with
or
and
maybe
we
can
just
discuss
it
in
the
channel
but
it'd
be.
B
Good,
it
depends
on.
I
can
definitely
be
the
like.
The
qa
at
the
end,
make
sure
everything
works,
but
it
kind
of
looks
like
depends
on
how
the
subcharge
is
working.
It's
working
that
may
be
yeah
change
a
little
bit
just
the
install
scripts,
I
think,
would
need
to
be
changed
a
little
bit.
I
assume
after
the
subcharges
installed.
It's
the
rest
will
be
the
same.
C
B
Nothing
nothing
has
like
really
changed
yeah,
we
added
some
features,
but
with
it
it's
not
reflected
in
the
document.
Those
are
more
advanced
features
that
it's
not
doesn't
belong
in
the
like.
First
start,
quick,
start
or
readme.
C
B
Yeah
yeah,
the
retail,
should
all
just
work,
but
there's
no
changes
that
change
that
it's
just
I
don't
know
how
sub
charge
install
command
is
maybe
a
little
bit
different.
There's
the
flag
or
something.
D
C
Want
to
actually
think
about
that,
I
would
be
totally
fine
with
it
being
installed
by
default
in
cross
play
in
0.13.
But
one
thing
we're
going
to
have
to
think
about
is
we're
considering
in
the
next
few
releases
cosplaying
going
to
v1
and
when
cross-plane
goes
to
p1.
It
probably
should
not
install
alpha
functionality
by
default.
C
D
Goes
v1
we
can
figure
out
how
to
do
that
yeah.
Maybe
we
can
talk
about
that
when
we
get
to
the
1.0
on
the
one
beta
1
discussion
a
little
bit
down
below,
so
if
it
is
by
not
by
default,
then
we
probably
want
to
have
the
docs
just
be
like
hey.
If
you,
you
know
want
to
go
through
this
demo
and
then
you
got
to
install
this
sub
chart
and
dana
sounds
like
you're
doing
that
so
yeah.
D
Cool
yeah,
so
maybe
some
tag,
team
qa
after
dan's
got
those
changes
in
would
be
good.
A
Sweet
all
right,
so
we
can
that's
the
overall
picture,
we're
you
know
targeting
getting
the
release
out
in
this.
B
A
E
E
It's
not
actually
why
we
made
the
switch,
but
it
does
make
that
a
little
bit
easier
before
we,
if
you
you
know,
change
an
annotation
or
set
the
fs
group
on
the
deployment
manually,
then
we
would
just
override
it,
which
basically
meant
you
couldn't
use
the
package
manager
to
do
irsa
stuff,
but
now
for
this
change,
which
was
actually
to
fix
an
issue
with
creating
tokens
over
and
over
again
for
the
service
count,
you
actually
should
be
able
to
set
that
fs
group
manually.
E
So
this
actually
seems
like
a
pretty
decent
solution
for
that,
because
all
of
the
deployments
are
going
to
be
made
in
the
name
space
where
crossplan
is
installed,
which
theoretically
is
where
you
know
an
admin
would
have
permissions
anyway.
So
anyway,
that
was
kind
of
a
side
effect
of
this
fix
for
this
issue
that
nick
identified.
E
So
if
anyone
is
using
that
functionality
on
this
call,
it
should
work
kind
of
similar
to
how
you
do
it
right
now,
but
that
document
in
in
provider
aws
is
going
to
be
updated
to
reflect
any
of
the
new
terminology.
A
Well
sounds
good
dan
thanks
for
the
thoroughness
on
that
nick
any
anything
else
to
specific
to
dive
down
to
in,
like
our
back
controller,
our
back
management
stuff,
no,
not
really.
A
Casey,
I
think
you
are
on
the
call
here.
Can
you
give
us
an
update
on
code
generation
provider
acceleration
stuff
and
that
may
be
not
much
update
since
you've
been
focused
on
helping
out
with
0.13
but
I'll?
Let
you
speak,
sir.
F
Yeah,
that's
correct
the
the
things
that
have
happened
since
last
week
on
an
update
where
I
moved
and
was
gone
for
two
weeks
and
then
I
helped
out
with
crank
command
for
the
last
a
couple
weeks.
So
I'm
going
to
be
switching
gears
back
to
it.
This
week,
though,
and
our
first
step
is
taking
the
proof
of
concept
which
previously
was
gcp
im
that
was
hand
coded
and
then
build
up
the
pipelines.
You
could
generate
that
one
resource
type
then
we'll
start
moving
that
out
to
generating
other
resources
from
there.
A
Awesome
definitely
excited
about
progress
on
that
for
sure
cool.
Any
specific
notes
about
you
know:
composition
going
to
b1
beta
1
or
that's
kind
of
covered
in
the
general
0.13
discussion.
C
A
Sounds
good
and
then
we
talked
about
ohm
and
getting
that
released
and
included
as
a
subchart
in
the
main
cross
plane
helm
chart.
So
I
think
we
covered
most
of
that
there.
Unless
ryan,
you
have
any
other
updates
or
things
you'd
like
to
share.
B
Not
really
we
have,
we
are
mostly
doing
bug
fixes
the
home
chart
is
the
own
runtime
is
currently
used
by
internal
services
in
alibaba
by
like
a
production
service,
so
yeah,
we
probably
can't
bump
it
to
be
at
least
give
a
bump
into
beta
in
the
next
release
when
they
they're
fully
online,
and
it
may
even
get
considered
that
b1.
If
there
are
no
other
issues.
C
B
Yeah
that
that's
actually
a
good
question
so
because
it's
used
internally
already
in
your
production
code,
I'm
not
sure
how
to
upgrade
in
terms
of
like
get
bumped
version.
I
I
don't
know
exactly
how
does
that
work
and
that
that
team
is
not?
You
know
it's
not
our
team
so
and
they're
all
obviously,
in
the
production
team
they're
all
busy,
I'm
not
sure
how
easy
it
is
to
bump
the
version.
B
As
I
said,
we
can
even
bump
it
to
v1
if
they
are
running
stable
there
for
a
while
in
the
production
code,
just
not
sure
how
to
bump
the
version.
D
B
We
actually,
we
already
have
the
crd
hooks
actually
to
to
convert
crds,
I'm
just
not
sure
because
they
are
importing
our
code.
So
the
background
is
they
don't?
I
don't
think
they
will
have
time
to
do
a
big
upgrade
just
because
our
version
bumped
that
that's
the
background
right.
So
we,
if
we
bumped
our
version,
I'm
not
sure
how
can
they
still
keep
up
with
use
our
latest
code?
They
need
to
bump
every
every
version
of
their
code.
I
I
don't
have
experience
on
that.
I
don't
know
in.
C
Theory,
if
if
what
you're
saying
is
that
they're
importing
your
api
types
and
your
api
types
of
v1
alpha
1
and
you
wanted
to
go
to
v1,
beta,
1
or
v1,
what
you
could
do
is
announce
a
timely
deprecation
of
v1,
alpha
1
types,
and
then
you
could
basically
update
your
controllers
to
operate
on
v1
beta1
types
that
were
just
exact
copies
of
the
v1
alpha
1
types,
and
I
believe
you
can
actually
tell
your
crds
that
so
so,
normally,
if
your
v1
beta,
1
or
v1
api
types
had
changed,
you
would
have
to.
C
You
know,
use
those
web
hooks
and
teach
teach,
though
and
runtime
how
to
convert
between
the
two.
But
I
believe
that
if
they're
exactly
the
same,
you
can
actually
just
serve
both
versions
and
just
tell
the
api
server.
Basically,
there's
no
conversion
needed.
If
someone
submits
me
a
v
one
alpha
one
resource
just
treat
it
the
same
as
a
v.
One
beta
one
and
it'll
translate
directly.
B
Does
that
does
that
compile
when
the
version,
even
the
type,
is
exactly
the
same?
How
does
the
server.
C
Compile
so
effectively
what
happens?
Is
you
update
all
of
your
code
to
use
the
v1
alpha?
1
types
you
mean
beta
or
still
sorry,
the
v1
beta
v21
b1
beta1
type,
so
you
would
basically
copy
the
entire
directory
of
alpha1
types,
create
them
as
beta1
types
update
all
of
your
controllers
to
operate
on
those
beta1
pipes,
but
still
add
crds
that
have
alpha
1
in
there
and
then
effectively.
C
B
Okay,
there's
still
details.
We
can
leave
that
to
the
end,
but
there
are
definitely
other
details
we
need.
I
just
try
to
see
how
that
works,
because
it's
yeah
the
ideally,
I
think
it's
like
they.
I
want
them
to
upgrade.
Follow
us
right
up,
follow
the
upstream
to
upgrade
to
v1
by
ta1
and
v1
whatever
it.
I
don't
want
them
to
stay
at
d1,
alpha
1.
C
B
C
Yeah,
I
think
what
we
would
do
here
would
be
aiming
to
give
them
their
own
timeline
so
that
we're
not
blocked
so
that
we
we're
not
in
a
place
where
we
can't
add
support
for
beta
to
oam
until
they're
ready
to
upgrade,
because
we
need
we
need
to
have
a
different
release.
Cadence
from
our
customers.
C
For
v013,
I
believe
we
could
probably
do
that
in
whatever
release
we
were
in
like
if,
if
we
really
are
gonna
cut
v,
one
beta
one
that
has
no
changes
from
v,
one
alpha
one.
I
believe
we
don't
need
conversion,
we're
books
for
that.
So
I
believe
at
that
point
we
could
say
if
you
want
f1
still
there,
you
can
still
use
it,
but
there's
now
you
should
now
it's
deprecated.
You
should
move
over
to
v1
beta
1..
A
Oh
yeah,
that
makes
sense
to
me
too
nice,
okay,
so
that's
everything
related
to
the
upgraded
milestone,
release,
etc.
So
I
think
we
can
move
ahead
to
the
community
topics
here.
Dan,
do
you
want
to
give
us
an
update
on
the
tbs
episode.
E
E
Sure
yeah,
so
me
and
marcus
worked
on
provider
ssh,
which
is
what
it
sounds
like.
It's
basically,
a
provider
that's
backed
by
ssh,
so
you
can
do
things
like
run
scripts
and
stuff
like
that.
So
that's
a
pretty
nice
tutorial
if
you
want
to
like
bootstrap
a
new
provider,
because
we
start
from
basically
nothing
and
build
it
up
to
some
minimal
functionality.
E
So
that'll
be
a
good
one
to
reference
back
to.
We
also
used
the
template
in
the
cross
plane
provider
template
repo,
which
is
a
template
repository,
so
you
can
actually
create
repositories
from
it.
We're
looking
at
how
to
maybe
make
that
a
little
bit
easier
by
parameterizing,
some
of
it.
So
you
don't
have
to
do
so
much
find
and
replace,
but
that's
a
good
one
to
look
at
there's
also
some
opportunity
to
create
other
similar
providers
to
provider
ssh.
E
I
know
nick
mentioned
in
cross-plain
slack
last
week,
a
provider
sequel
which
would
do
similar
things
but
obviously
connecting
to
a
sql
database.
So
interesting
pattern
there.
If
you
want
to
check
that
out
and
then
next
week
we
are
going
to
have
the
cto
of
sourcegraph
joining
us
and
we're
going
to
look
at
deploying
source
graph,
which
depends
on
redis
and
postgres
into
kubernetes
cluster
and
supplying
credential
information
after
provisioning,
redis
and
postgres
database
using
crossplane.
E
So
it
should
be
pretty
cool
to
see
and
hopefully
we'll
be
able
to
include
that
in
the
documentation
at
some
point
for
installing
source
graph.
For
folks
who
want
to
install
via
kubernetes,
which
is
their
recommended
method.
A
Nice
dan,
I
like
that
it's
kind
of
like
a
more
solutions,
focused
episode
there.
It's
like
here
to
tell
you
to
deploy
a
particular
you
know
set
of
software
and
then
also
using
crossplane
to
deploy
all
the
managed
services,
and
you
know
everything
that
it
might
need
there.
So
that's.
E
Exciting,
that's
is
that
this
week
or
next
week
that's
gonna
be
next
thursday.
One
important
thing
to
know:
it's
not
gonna,
be
at
our
regular
time.
It's
gonna
be
a
little
bit
later
in
the
day
to
accommodate
his
schedule.
A
Got
it
cool
sounds
good,
so
that's
thursday
october
8th,
though.
E
Yes,
it
is
going
to
be
at.
Let
me
look
at
my
calendar,
real
quick.
I
believe
it's
going
to
be
at
3
30.
Well,
so
1
30
pacific.
Let
me
actually
double
check
that,
though,.
A
Exciting
nice
all
right,
okay,
then
a
quick
update
on
cncf
onboarding
stuff.
So
we
excuse
me,
we
transferred
the
the
crossband
I
o
domain
over
to
the
care
and
oversight
of
the
cncf
and
their
ite
department.
With
the
linux
foundation.
A
On
friday,
we
did
have
a
like
a
domain
related
outage
where
the
releases.crossline.io
and
the
charts.crossplan.io
dns
records
were
not
migrated
over,
so
they
those
were
not
resolving.
Once
the
domain
had
been
transferred
over
to
the
new
dns
provider
for
the
cncf
uses.
A
So
some
of
the
builds
and
stuff
were
failing
on
friday
and
you
wouldn't
be
able
to
install
like
home,
charts
and
stuff
as
well
on
friday,
but
we
got
that
resolved
and
the
root
cause
of
that
was
that
to
migrate
the
dns
records
to
the
new
dns
provider.
We
had
to
extract
export
all
of
them
out
of
our
old
dns
provider,
which
is
aws
route.
53.
A
aws
in
their
caring,
loving
hands
of
vendor
lock-in
do
not
provide
any
mechanism
to
export
all
the
records.
So
you
have
to
cobble
it
together
with
like
the
aws
cli
tools
and
stuff
and
do
like
some.
You
know
some
parsing
and
you
know
walking
and
gripping
and
nonsense
like
this.
So
when
I
did
that
I
missed
my.
My
little
script
missed
some
of
the
records,
so
I
the
zone
file
I
gave
over
to
the
cncfit
people
did
not
have.
It
was
missing
the
the
entries
for
releases
and
charts.
A
Foundation
id
people
added
those
to
the
dns
entry
almost
immediately
too
so
they
have
a.
They
said
they
have
a
very
quick
turnaround
in
sla
for
domain
related
stuff.
So
that
is
nice.
We
should
not
have
any
more
issues
there.
If
anybody
is
having
dns
issues
with
anything
under
the
crosspin.io
domain,
let
me
know-
and
we
should
get
that
fixed
really
quickly.
D
That
outage
there
well
thanks
for
getting
that
sorted
out,
nice
cncf
now
so.
A
Yep
they
they
have
our
dns,
they
own
the
crossman.io
domain,
the
neutral
home,
okay
and
then
kubecon
north
america.
Tomorrow
is
the
date
that
the
responses,
or
the
exception
acceptances
and
rejections
go
out
for
kubecon,
north
america
call
for
proposals
or
papers
or
whatever.
So
folks
should
learn
tomorrow.
A
If
you
were
accepted
to
speak
at
the
kubecon
n
a
so
it's
exciting,
I'm
always
excited
to
hear
the
the
wave
of
responses
come
in
and
see
who
got,
talks
and
whatnot
cool
anything
else
in
the
community
topic
section
here
for
folks.
A
All
right,
we
did
not
have
any
particular
pr's
listed
here
that
folks
wanted
to
go
over
if
there
were,
let
me
know,
but
otherwise
that
is
the
end
of
the
full
audience
discussions
here
and
I
do
have
an
entry
in
the
optional
time
for
deeper
technical
discussions.
So
if
there
are
no
other,
you
know
full
agenda
items
here,
then
we
can
move
on
to
that
and
anybody
who
wants
to
take
off
besides.
A
I
would
like
at
least
like
nick
and
dan
to
stay,
though,
to
answer
some
of
the
questions
here,
but
other
than
that,
though
folks
can
take
off
if,
if
they
want
to.
A
A
Things
are
going
well,
you
know
with
like
running
across
playing
with,
like
the
conduit
is
kind
that
sh
script,
so
I
have
the
latest
like
built
locally
and
I'm
you
know
for
my
provider
aws
though
I'm
manually
installing
the
crds
into
my
cluster
and
like
doing
make
run
so
I'm
running
the
controller
locally
pointed
at
the
con
cluster.
So
I'm
not.
You
know
using
a
package
or
anything
from
there.
A
What
and
then
with
like
the
compositions,
I'm
creating
I'm
not
building
a
package
out
of
them
either
I'm
just
applying
those
directly
to
the
cluster
as
well.
Can
you
give
me
a
quick
update
on
like
are
the
do?
We
have
the
tools
like
the
crank
tools
and
stuff
to
a
position
where
we
can
do
where
should
be
building
and
installing
packages
locally
or
what's
the
what's?
The
latest
like.
F
E
So,
as
far
as
the
cli
goes,
there's
a
pr
open
that
implements
the
building
and
pushing
and
that
sort
of
thing
which
is
the
same
for
dev
or
you
know
whatever,
if
you're
actually
building
the
package
in
the
case
that
you're,
actually
you
know
just
installing
them
directly,
then
I
would
advise,
as
I
said
in
slack
just
using
like
cluster
admin
or
something
like
that
for
just
testing
locally,
because
obviously
we're
not
able
to
like
you
know
just
arbitrarily
give
permissions
without
you
know
you
basically
indicating
that
you
want
them
so
yeah.
E
That's
what
I
would
say
I
mean
I
guess
we
could,
since
the
kind
dot
sh,
I
think,
is
just
used
for
local
dev.
You
know
we
could
in
there
just
give
it
cluster
admin,
give
cross,
try
and
give
cross
plane
cluster
admin.
However,
that
could
be
confusing
for
someone
if
they
are
running
crossplaying
locally,
but
then
they're
actually
trying
to
like
test
their
package
install
workflow.
E
A
A
Yeah
yeah,
that's
that's
why
I
got
feeling.
That
too,
is
that
you
know
like
I
like,
because
some
of
this
is
an
artifact
of
the
tools
are
being
built
right
now,
so
you
know
like
I,
I'm
excited
to
get
to
a
place
where
I
can.
You
know,
build
my
own
package
locally
and
then
just
you
know
install
that
like
using
the
tool
flow
with
with
my
local
cluster,
but
and
so
yeah.
A
So
I
would
want
to
depend
on
like
the
automated,
our
back
stuff
and
the
permissions
getting
you
know
that
I
have
to
find
the
right
permissions
or,
like
that's
all
done
correctly
as
opposed
to
having
it
covered
up
by
you
know
the
cluster
admin
being
bound
to
the
crosstown
service
account
automatically
through
the
kind
script.
So
what
is
what
is
the?
E
So
so
I
think
I'm
a
little
confused
because
for
like,
if
you're,
not
installing
it
as
a
package,
the
local
dev
flow
you
know
should
mirror
how
crossplaying
works.
So
we're
not
gonna.
I
don't
think
we're.
Gonna
have
tooling,
that's
saying,
like
you
know,
apply
these
permissions
when
I
haven't
actually
installed
it
as
a
package.
So
if
you
just
like
manually,
install
crvs,
crossplane
isn't
going
to
be
given.
A
Yeah,
that's
totally
fine
yeah.
I
I
completely
understand
that,
like
I
I'm
more
curious
about
what
like
what
is
left
or
what
what
what
will
be
the
experience
like
to
be
able
to
when
I'm
to
be
able
to
build
my
own
package
like
I
would
like
that's.
The
next
thing
I
want
to
be
able
to
do
is
like
hey
cool.
E
Yeah,
so
if
you're
installing
via
package,
then
it's
done
it's
ready
right.
So
it's
going
to
give
you
the
correct
permissions
for
your
package.
There
are
a
few
minor
updates,
I
think,
maybe
got
merged
this
morning,
but
that
should
be
working.
So
how
do
I
do
that?
Then?
I.
E
Right
so
you
can
manually,
do
it,
but
the
crank
package
build
is
not
has
not
been
merged,
yet
the
pr
is
open,
we're
resolving
some
dependencies
and
that
sort
of
thing,
but
once
it's
ready
just
in
your
directory,
where
you
have
your
compositions,
you
would
just
say
I
guess:
cube
control,
cross,
plane,
build
and
then
cube
control
cross
plane
push.
Essentially,
oh.
A
A
A
E
A
Oh,
no,
it's
not
a
blocker,
it's
not
a
buck.
I'm
just
excited
to
be
able,
like
almost
like
a
qa
effort.
Here
too,
you
know
to
be
like
sweet.
I've
got
a
package
and
I'm
and
I'm
installing
it
and
I'm
exercising
the
new
code
where
we're
building
and
we're
doing
qa
cycles
too.
With
this
work,
I'm
doing
that's
that's
what
I'm
saying.
C
We're
working
on
it
as
fast
as
we
can
and
we'll
totally
let
people
know
when
that
when
that
is
done,
it's
been
much
more
complicated
than
expected
to
plug
up
a
cli
tool
to
docker.
So
we'll
we'll
make
it
very
clear
when
that
tool
is
available,
sweet
yeah,
awesome,
yeah,.
A
C
Then
keep
in
mind
that
it
does
just
make
a
docker
image,
so
it's
totally
possible
to
just
basically
take
the
directory
structure
that
we
talked
about
in
all
of
our
examples,
like
example,
cmp
and
just
docker,
build
them
into
a
dockerfile
and
push
that,
and
you
know
you
can
make
your
own
package.
You
don't
need
this
tooling.
If,
if
you
wanted
to
try
it
out,
if
you
want
to
try
everything
after
the
cli,
tooling,
okay.
A
C
A
Got
it
do
like
a
from
like
the
base
image,
the
package
unpack.
E
Yeah
yeah
and
if
you
look
at
the
provider
gcp
and
then
provider
aws
just
merged,
I
saw
there's
an
example
of
what
the
docker
image
looks
like
and
yeah.
It's
it's
literally
from
package
unpack
master
right
now
I
mean
you
can
pick
whatever
you
want
and
and
then
like
add
the
directories
called
package.
Add
package
dot.
It
can
actually
be.
You
could
have
whatever
nesting
structure
you
want
there
and
yeah.
E
If
you
just
docker,
build
that
and
then
like
load
that
in
your
kind
cluster
for
instance,
and
then
create
a
provider
configuration,
then
that
should
work
great
for
you.