►
From YouTube: Kubernetes SIG Apps 20181112
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
Welcome
to
the
kubernetes
sig
apps
for
Monday
November,
12
2018,
all
right
into
chat
here,
I'll
go
ahead
and
share
the
agenda.
My
name
is
Matt.
I
will
be
your
chair
today
we
don't
have
any
announcements.
One
of
the
things
we
do
have
very
first
up
is
a
demo
and
Dan
I,
see
I.
Think
that's
you
that's
on
you.
There
I'm,
so
we
normally
start
off
with
a
demo
here
and
so
the
the
floor
is
yours.
If
you'd
like
to
demo
for
us.
B
Perfect
so
hi,
my
name
is
Dan
I'm,
doing
a
tool
called
tilt,
and
it's
what's
running
in
my
terminal
right
now
now,
and
it
has
a
bunch
of
names
here.
These
are
different
things
running
in
kubernetes.
We
call
them
services,
but
they
include
deployments
and
they're
running
in
docker
for
Mac
on
my
local
machine,
and
the
idea
here
is
that,
as
you
develop
code
that
runs
in
kubernetes
in
production
tilt
gives
you
a
local
instance,
and
so
I
can
show
you
here.
B
I
can
select
through
the
different
ones
and
I'm
going
to
open
the
front
end
which
you
by
hitting
B,
which
is
here
on
localhost,
and
you
can
see
that
I
have.
This
is
like
our
demo
app
it's
one
of
those
things
that
was
popular
in
the
mid-2000s
a
personalized
home
page,
and
so
each
of
these
different
widgets
is
served
by
a
different
micro
service,
just
to
illustrate
that
we're
about
micro,
service
development-
and
so
you
can
see
here
that
I
have
a
service
for
dog
O's
and
a
service
snack
selector.
B
But
the
thing
that
we
think
is
even
more
useful
is
that
it's
easy
to
mess
up
and
when
you're
doing
micro,
service
development
there's
lots
of
different
ways
that
things
can
go
wrong.
So
here
I'm
going
to
add
this
err
that
should
cause
a
crash
loop
and
so
when
I
save
it
and
edit
you
can
see
the
oh
hey,
I
thought
this
is
going
to
cause
an
error
run
time,
but
actually
I
forgot
to
import
the
code,
and
so
we
see
the
build
err.
B
If
I
go
and
I
fix
that
build
error,
then
hopefully
this
time
the
build
will
finish,
but
it
will
go
into
crash
loop
and
you
can
see
that
we
can
see
the
output,
and
so
this
is
one
of
those
things
where
oftentimes
I've
had
to
play.
20
questions
with
kuba
cuddle
to
figure
out
what's
breaking
and
so
tilt
takes
all
the
airs
and
puts
them
into
one
pane
of
glass
in
the
in
the
same
way.
A
B
You
say:
hey
I
have
a
make
my
snack
service,
it
has
some
animal
and
it
builds
with
this
docker
file
and
I
want
to
port
forward
it
to
9002,
and
so
this
is
something
where
it
doesn't
fit
in
the
kubernetes
e
mo,
because
it's
not
about
what's
happening
in
the
cluster.
It's
about
what's
happening
on
your
local
machine
and
you
can
set
all
these
up
in
this
simple
declarative,
ish
way.
B
But,
for
instance,
we
don't
use
customize
or
case
on
that
we
use
m4,
which
is
this
old
UNIX
templating
engine
temp,
and
we
didn't
have
to
like,
extend
the
tool
to
do
this.
This
done
in
our
configuration
by
just
you
can
run
local
commands
as
part
of
the
flow,
and
so
we're
able
to
take
workflows
that
don't
conform
exactly
to
how
the
tool
thinks
of
it.
Because
you
have
this
program
ability
to
so
soft,
just
hard-coded
configuration,
but
actually
something
you
can
play
with
and
edit.
B
Yeah
and
so
I'm
going
to
go,
fix
this
air
just
because
it's
kind
of
there
and
and
distracting
me
and
so
now,
I
can
see
when
it's
ready
when
it's
all
green,
I
restart
it.
And
here
we
go
so
if
you
have
code
that
has
micro
services
and
you're
editing,
multiple
of
them,
Tilt
gives
you
a
personal
instance
end-to-end
it
give.
It
keeps
that
up-to-date
without
you
having
to
think
about
it.
B
If
you
can
go
to
window
top
engineering,
that's
our
homepage.
You
can
see
it
has
a
little
demo
mode
and
there's
a
video
and
then
we're
down
on
we're
open
source
on
github
and
it's
easy
to
set
up.
If
you
have
your
animal
finals
and
your
darker
files,
it
should
be
about
five
or
10
minutes.
To
start
seeing
it
working
and
we'd
love
to
help
set
you
up
we're
on
Twitter
and
as
twitter.com
slash,
windmill,
Inge
and
I'm
on
Twitter,
as
Twitter
as
at
D,
Bentley
and
I.
C
B
B
B
Yeah,
there's
there's
code
in
here
that
I
didn't
write
were
a
six-person
startup
based
in
New
York
and
so
there's
definitely
code
to
figure
out
from
your
mo
how
to
change
it.
To
tell
it.
You
know
when
to
pull
or
when
not
to
pull
based
on
what
kind
of
kubernetes
it
is
so
out
of
the
box,
there's
logic
to
make
it
work
well,
and
we
can
also
I
didn't
talk
much,
but
we
can
also
well.
We
have
a
lot
of
logic
that
can
actually
run
the
builds
in
the
cluster
on
the
node,
where
it's
running.
B
So,
if
you
configure
it
with
give
us
a
little
more
detail
when
you're
editing
code,
for
instance
our
front-end
binary
links
in
all
the
Cooper
based
client
libraries,
which
means
doing
a
docker
build
from
scratch.
The
go
build
takes
a
minute,
but
we
actually
use
the
previous
cache
so
to
tilt,
keeps
that
around
and
injects
it
when
you
save.
So
you
have
iterative
builds
like
you
have
outside
docker,
so
that
there's
no
need
you
to
not
have
a
reason
to
stop
doing
this.
We
see
so
many
people
come
up
with
escape
hatches.
B
Gonna
have
to
try
this
out
great
yeah
and
for
anyone
else,
yeah
please,
we
can,
if
you,
if
you
reach
out
to
me,
we'd,
love
to
find
time
to
you
know
pair
with
you
to
get
it
set
up,
because
this
is
what
we're
up
to
or
we
can.
You
know,
set
up
as
a
shared
slack
channel.
So,
if
you're
interested
in
this
and
having
onboarding
problems
where
we're
on
call
for.
C
A
A
Monday
I
want
to
say
it's
the
26th,
it's
the
Monday
after
Thanksgiving
in
the
United
States.
The
question
is:
should
we
cancel
that
and
we're
asking
because
that
means
a
short
week
will
have
a
meeting
on
Monday.
Then
we'll
have
Tuesday
Wednesday
and
then
many
folks
are
going
to
have
the
all
the
days
from
there
on
out
off.
Until
we
come
around
to
the
next
say,
gaps
meeting
and
will
we
have
content,
are
folks
interested
or
should
we
cancel
that?
Do
we
have
any
opinions?
A
A
A
B
A
So
Kenan
Adnan
I
hear
you're
both
here,
I'm
not
gonna,
be
around
in
doing
anything
in
that
time
period,
which
probably
means
the
service
definitions
and
applications.
Er
D
and
some
of
the
other
stuff
won't
get
will
either
of
you
being
around
doing
much
with
you're,
both
muted,
by
the
way.
E
A
All
right,
I'm,
actually
gonna
bump
up.
We
have
a
developer
tools.
Discussion
typically
here,
but
I
wanted
to
actually
discuss
the
cap
beforehand.
So
we
have
a
proposal
out
for
port
heiden
I
named
it
I
gave
it
a
very
business.
He
named
portable
service
definitions,
you
sure
my
screen
and
it's
an
idea
that
we've
been
working
around.
C
A
Perfect,
it's
called
portable
service
definitions
right
now,
it's
being
written
in
over
in
hack
MD
before
it
goes
up
for
a
pull
request.
But
if
this
sounds
good
to
folks,
then
we'll
look
at
doing
it
either
later
today
or
tomorrow,
turning
it
into
a
pull
request
and
the
goal
is
okay,
so
simple
goal
I
want
to
go,
deploy
something
into
a
juror
and
gke
and
AWS
and
we'll
say
docker
for
Mac,
and
it's
going
to
be
my
little
application.
For
example,
just
say
WordPress
and
it's
gonna
use
my
sequel
as
a
database
right
and
I.
A
Don't
want
to
maintain
the
database
myself.
I
want
to
use
a
SAS,
and
so
I
want
to
just
say:
give
it
to
me
and
I
want
to
be
able
to
ask
for
it
and
all
the
environments
and
then
get
back
the
credentials
in
a
consistent
manner
everywhere.
So
then
everything
just
works
right.
I
get
the
full
round-trip
I
get
a
secret
I
get
the
credentials.
Everything
starts
up.
My
application
is
able
to
get
it,
doesn't
matter
what
in
I'm
trying
to
go
into
and
and
that's
kind
of
the
motivation
behind
it?
A
How
do
we
get
application
portability
when
we
have
to
deal
with
services
right
now?
Kubernetes
gives
us
application
portability
for
things
we
have
to
deploy
into
kubernetes,
but
then
there
are
lots
of
common
services
and
databases
are
a
really
simple
one
DNS
and
there
there
are
a
number
of
others
that
can
be
labeled
in
the
same
light
that
are
common
services.
Now
it
doesn't
fit
everything,
so
the
goals
here
right
are
to
do
the
portability
to
provide
a
common
way
to
request
a
service
obtain
credentials,
detect
which
services
are
even
available
to
you.
A
To
work
with
the
major
public
clouds
now
an
on
goal,
you
can't
provide
an
out-of-the-box
solution
for
every
bit
spoke
service
and
there
are
more
and
more
and
more
of
them
these
days,
it's
probably
only
gonna
work
for
lots
of
common
services,
but
when
folks
have
radically
different,
api's
and
feature
details
and
their
stuff,
there's
no
way
to
provide
commonality
across
that,
because
it
just
isn't
commonality
and
of
course
we
don't
want
to
replace
the
kubernetes
service
catalog.
This
may
work
with
it.
It
all
depends
on
where
things
are
going
to
go.
A
It's
focused
very
much
on
the
application
operator
experience
of
being
able
to
deploy
into
multiple
environments,
which
is
why
we're
interested
in
this
by
the
way
at
the
beginning
of
it,
you
can
notice
that
this
is
all
the
metadata
on
it.
We
did
put
Service
Catalog
down
as
a
participating
Sagan
we're
going
to
ask
the
chairs
over
there
to
be
reviewers,
because
we
do
want
to
work
well
with
this,
because
we're
stepping
into
a
similar
space
I
did
outline
to
user
stories.
A
One
is
you
can
query
and
see
what
services
are
available
using
a
kubernetes
native
API,
something
like
coop
control,
get
CR
DS
right
and
you
can
see
what
they
are
and
then
another
one
is.
You
can
declaratively
request
an
instance
of
a
service
with
a
custom
resource,
and
then
you
can
get
that
secret
back
in
a
common
way
across
everything,
and
so
the
implementation
details
here,
along
with
a
little
bit
of
plan,
are
to
start
with
something
simple
you'll
see
my
sequel
and
Postgres.
A
The
plan
here
is,
and
the
last
one
is
to
work
out.
Details
is
to
start
with
one
service
and
then
to
work
with
and
try
and
get
it
to
work
with,
say,
kubernetes
locally.
Maybe
it's
a
controller
that
deploys
home
or
just
employees,
resources
just
as
a
proof
of
concept,
and
then
you
have
something
work
with
Sasa's
across
the
three
main
public
clouds,
the
through
by
adoption.
Others
are
included,
but
I
just
happen
to
know
that
we
can.
A
That's
kind
of
the
concept
behind
it.
It's
it's
all
about
closing
the
whole
loop
across
the
board,
and
this
is
gonna
be
hard.
So
the
risks
that
I've
noted
and
if
folks
have
other
risks
are
the
access
controls
and
relationships
are
unaccounted
services.
How
do
we
do
this
with
tenancy
and
things
like
that,
and
then,
of
course,
the
details
when
you
get
into
some
of
these
about
requesting
size
and
features
on
them?
How
do
we
handle
those
kinds
of
things
and
how
do
we
common
eyes
that
kind
of
that
kind
of
detail?
A
Those
are
kind
of
two
of
the
risks.
I
don't
have
mitigations
for
those.
Yet
for
the
graduation
criteria,
I
said
a
certain
number
of
organizations
have
to
have
adopted
this.
We
have
to
have
at
least
at
least
three
services
that
are
created
and
in
use,
and
of
course,
we've
got
to
document
both
using
this
to
go.
Create
controllers,
because
controller
is
part
of
this
as
well.
So
one
of
the
details
that
I
guess
I
didn't
share
yet
was
you
get
the
custom
resources,
then
you
have
controllers
and
those
controllers
can
be
different
in
every
environment.
A
As
long
as
your
resource
is
the
same
in
your
seeker,
coming
back
is
the
same.
How
its
implemented
on
the
backside
can
be
different,
so
you
could
have
an
AWS
controller
and
is
your
control
or
a
Google
Cloud
Controller
and
IBM
cloud
controller,
an
Oracle
cloud
controller
controller
for
docker
for
Mac
and
mini
cube.
You
could
have.
A
We
could
have
a
variety
of
controllers,
but
as
long
as
they
implement
the
same
interface,
that's
fine
and
that's
a
way
to
get
across
environments
right,
and
so
we
want
to
have
documentation
for
how
to
actually
use
this
to
create
controllers
and
what
you've
got
to
do
and
then,
of
course,
servers
consumer
documentation
on.
How
do
you
use
this?
If
you're
gonna
go
create
things,
and
so
that's
the
I?
A
The
idea,
the
alternative
here
is
listed
out,
is
actually
modifying
the
Service
Catalog
to
have
more
of
this
functionality,
which,
as
I
understand
it
is
gonna,
be
really
really
hard
and
so
having
a
separate
component
is
probably
the
easier
way
to
go.
In
addition
to
that,
the
Service
Catalog
is
tightly
coupled
to
the
open
service
broker
and
well
that
will
work
great
in
some
environments
that
have
very
mature
open
service
brokers.
Some
things
like
in
docker
for
Mac.
C
I
was
calling
the
proof
of
concept
cloud
Kynes,
just
because
what
I
was
really
focusing
on
was
being
able
to
have
a
resource
that
was
specific
to
the
type
of
sass
that
you
were
creating
here.
So
if
I
looked
at,
for
example
on
my
sequel,
that's
kind
of
like
the
canonical
example
here.
I
want
to
be
able
to
do
this
and
say
kind.
C
And
so
what
this
looks
like
for
the
prototype
is
I
have
a
main
controller
that
understands
how
to
watch
for
a
dynamic
set
of
kinds.
It
doesn't
use
like
a
tight
client
set,
like
you're,
usually
not
used
to
seeing
it
takes
advantage
of
the
new
dynamic
and
formers
and
dynamic
client
sets
that
are
in
I,
think
scope,
controller,
run
time
and
obtain
kubernetes,
and
so
what
I
can
do
is
if
I
do
keep
CTO
get
other
writers
ignore
all
my
names.
C
It
says
that
the
service
catalog
provider
understands
my
sequel
it.
This
is
just
for
the
demo,
but
so
service
catalog
can
actually
handle
pretty
much
anything.
So
this
could
be
dynamically
updated
by
service
catalog
itself
and
list
out
everything
it
has
in
its
catalog,
and
then
it
also
defines
a
little
web
hook
and
the
main
controller
cob
Kines
watches
for
say
my
sequel
and
when
it
gets
a
new,
my
sequel
resource
created
of
CRD.
It
goes.
Oh,
hey,
I'm
gonna
kick
off
this
web
hook
and
tell
it
you
need
to
do
do
your
thing.
C
C
C
That
said
that
knew
on
a
handle,
my
sequel
in
this
case
it
was
service
catalog,
but
it
could
have
been
anything
they
don't
have
to
be
tied
to
any
particular
spec
or
anything
like
that
and
in
the
webhook
payload.
It
just
says:
hey
here's
this
CRD,
you
say
you
know
on
handle
it
go.
Do
your
thing,
and
so
this
adapter
that
I
had
here
for
Service
Catalog,
essentially
took
that
CRD
and
took
the
payload
out
of
the
spec
like
these
types
of
information
and
turn
that
into
a
service
catalog
service
instance.
C
So
I
have
a
couple:
I
have
a
mini
broker
running
on
this
machine,
so
I
do
Mindy
broker,
allows
you
to
use
a
service
catalog
on
a
mini,
cube
cluster
and
have
it
essentially
provision
home
charts
for
you
so
that
it's
basically
the
OSB
spec
implemented
on
a
local
cluster
for
you.
So
if
I
do
a
QC
TL
get
pods
I
should
have
my
sequel
database
here,
yeah
that
I
can
use.
C
C
So
our
sake
is
interested
in
doing
a
lot
of
really
good
UX
changes,
but
unfortunately,
what
we
have
is
like
tied
completely
to
the
open
service
brokers
back
and
the
ecosystem
didn't
really
adopt
it.
To
be
honest,
the
major
cloud
providers
have
brokers.
Amazon
has
already
decided,
though,
to
invest
more
in
operators
and
there's
other
things
out
there
that
are
cool,
that
people
could
be
making
like
Etsy
d
operator,
for
example,
that
you'd
want
to
be
able
to
use
and
don't
want
to
have
to
force
everything
into
aspect.
C
That
essentially
wasn't
in
my
opinion,
widely
adopted,
though
SB.
So
this
is
a
way
around
that
and
this
just
kind
of
shows
the
UX
of
what
we
could
do
and
it's
cool,
because
we
actually
naming
world
to
get
people
agree
on
what
these
comments
back
should
look
like
and
how
the
secret
should
come
out
because,
like
for
cube
apps,
for
example,
this
has
always
been
a
huge
problem
that
every
provider
gives
your
username
password
and
connection
info
in
different
ways,
and
things
like
that.
A
A
Maybe
sometimes
you'll
want
to
do
that,
and
sometimes
you
won't
some
providers
may
or
may
not
that's
one
of
the
details.
We'll
definitely
need
to
work
out
and
sometimes
you're
not
going
to
be
able
to
get
all
of
the
credentials
as
passed
in,
for
example,
if
you're
gonna
say
give
me
MongoDB
or
you're
gonna
say
give
me
the
my
sequel
connection
string.
A
You
might
be
able
to
pass
in
a
username
and
password,
but
you
may
not
be
able
to
pass
in
the
DNS
name
for
the
location
of
it
right,
so
something
you're
gonna
need
to
get
back
from
them,
because
that's
just
the
way
the
SAS
providers
work.
You
can't
always
pull
that
off
by
saying.
Give
me
this
name
that
I
want
is
she's,
not
gonna
work
now
for
the
database
name
like
she
showed
off,
you
totally
can,
and
you
probably
can
for
some
of
the
things
like
usernames
and
passwords,
but
there's
some
things
in
there.
B
E
Also,
there's
a
difference
between
coordinates
and
credentials,
so
maybe
credentials
would
be
something
that
you
might
want
to
provide
as
part
of
the
declaration,
but
I
would
think
for
most
things
are
to
get
turned
up
automatically
for
you.
The
coordinates
of
the
service
that
you're
going
to
get
back
like
the
my
sequel,
connect
string,
or
maybe
just
an
IP
address,
or
a
DNS
name,
depending
on
the
type
of
thing
that
you're
turning
up
I,
think
that
might
be
something
that
would
be
injected
by
the
cloud
provider.
A
C
One
question
there:
there
are
some
new
axe
things
that
I've
been
trying
to
put
in
Service
Catalog
that
I'm
giving
up
on
and
I'd
love
to
get
something
here,
and
so
I
wanted
to
get
your
opinion
on
this
sure.
So,
for
example,
the
my
sequel
cases
is
pretty
straightforward
when
you
think
about
like
deployment
on
mini
cube,
but
oftentimes
they're,
like
optional
or
even
required,
current
configuration
parameters
that
individual
cloud
providers
need
and
it's
cool
if
the
controller
forsake,
gke
the
controller
for
AWS
or
whatever
defaults
and
things
in
a
useful
way.
C
But
one
of
the
things
I've
been
working
for
with
Service
Catalog
is
allowing
the
cluster
operator
to
say
here.
The
I
want
to
use
on
this
cluster
or
in
this
namespace,
because
maybe
they
don't
like
the
defaults
for
whatever.
Maybe
they
want
to
add
extra
white
listing
of
IPs
or
or
change
how
they
do
backups.
C
You
know
all
sorts
of
like
things
that
you
can
do
and
so
I
thought
it'd
be
neat
if,
in
addition
to
the
the
CRD
being
able
to
say,
here's
all
the
extra
little
toggles
I
need,
but
also
have
a
way
to
not
say
that
and
then
have
the
cluster
operator.
Whoever
the
sysadmin
is
to
that
cluster
kind
of
step
in
and
provide
defaults
that
make
sense
for
whatever
environment
they're
running
and
maybe
like
a
shared
development
environment
where
everyone's
needs
to
use
tiny
databases
or
things
like
that.
C
You
can
say
like
if
you
just
tell
me
my
sequel,
a
lot
of
matically
fill
in
all
the
extra
things
you
need
and
you'll
get
something
why
you
could
configure
say
the
gku
controller
or
something
to
to
be
a
little
more
customized.
Whatever
you're
trying
to
do
and
I
wrote
that
would
belong
in
this
cap
or
it
would
have
to
be
a
separate
thing,
because
it's
near
and
during
my
heart
and
feels
all
very
tied
in
together.
Well.
A
E
There
are
a
couple
of
things
here,
though,
super
there's
a
lot
of
places
in
kubernetes
that
you
can
have
multiple
and
not
multiple,
but
you
can
have
diverse
implementations.
Storage
is
probably
a
good
example
like
dynamic
provisioners
work,
definitely
differently
for
extra
for
every
cloud
and
there's
kind
of
a
base
set
of
parameter
that
you
would
probably
want
to
have
for
MySQL,
for
instance,
right
like
just
general
things
that
are
probably
configurable
across
the
board.
E
One
of
the
issues
is
that
for
RDS
versus,
like
let's
say
cloud
SQL,
that
the
set
of
things
that
you
actually
are
allowed
to
configure
might
differ.
So
one
question
is:
do
you
just
have
the
minimum
subset?
Basically,
the
lowest
common
denominator.
That's
available
everywhere
is
explicit
configuration
parameters.
Then
they're
going
to
be
things
that
are
cloud
specific
that
you
want
to
configure
and
with
most
places
in
the
system
where
you're
touching
an
opaque
cloud
provider
and
you're
still
providing
some
overriding
configuration.
E
It's
basically
like
a
map
string
string,
but
you're
allowed
to
pass
in
arbitrary
parameters
for
your
object.
That
might
be
specific
to
the
underlying
implementation.
So
that
would
be
one
way
to
do
it,
but
I
think
that
would
be
a
first
step
and
that
that's
just
figuring
out.
How
do
you
get
parameters
for
each
environment
into
the
controller
in
a
uniform
way,
and
then
what
you
were
bringing
up
is
defaulting
those
things
to
reasonable
values.
E
A
So
what
I
was
actually
and
I
think
can
I
agree
with
this.
The
reason
that
I
actually
put
in
here
that
when
we
start
off,
we
start
off
with
something
known
like
my
sequel,
and
we
have
to
do
it
across
the
three
major
cloud
platforms
plus
doing
it
in
kubernetes.
That
would
force
us
to
deal
with
the
wide
variety
of
those
configurable
parts
and
try
to
come
up
with
a
sane
way
to
do
something
that
we
all
know
pretty
well
and
come
up
with
some
good
patterns
and
thought
processes
on
it.
A
C
So
please
submit
this
as
a
cap.
Do
you
recommend
that
I
suggest
that
as
a
comment
on
this
or
do
you
think
it
should
just
be
like?
Let's
keep
in
mind
for
two
months
from
now
to
circle
back
on
that
I
just
wanna
make
sure
that
we
don't
like
push
that
off
and
then
we
realize
that,
because
we
need
open,
beginning
it's
harder
to
force
people
to
all
play
so
things
yeah.
A
We
would
start
working
on
it.
It
would
not
be
implementable
yet
because
we
still
have
to
go
figure
out
all
those
little
details
by
doing
a
bunch
of
the
prototyping
work
and
stuff
around
this
eventually
we'd
figure
those
out
and
it
would
move
to
implementable,
in
which
case,
hopefully,
we've
worked
out
enough
and
prototyped
enough.
We
could
go
to
implement
it,
but
this
can
evolve
over
time.
Merging
does
not
mean
it's
done.
Merging
means
it's
kind
of
been
accepted.
It's
searchable,
it's
it's
state,
and
so
there
are
things
like
provisional
if
lamentable
implemented.
A
D
A
There's
a
few
different
things
we
can
do,
but
I
can
I
can
start
carrying
this
to
the
next
step.
Carolyn.
If
you
actually
want
to
write
this
as
another
paragraph
on
this
and
then
ping
me
when
it's
done,
you
can
put
your
thoughts
in
the
defaults,
then
I'm
happy
to
put
it
in
yeah.