►
From YouTube: 20180313 kubeadm office hours
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
B
So
we
have
a
tracking
issue
in
the
Canadian
repo
about
the
docks
and
we
already
have
four
four
piers
to
merge.
This
cycle
and
I
got
pinked
on
private
by
Jim
and
Jill
about
our
status.
He
told
told
me
that
we
are
behind
because
apparently
they
wanted
to
start
reviewing
this
week,
but
we
still
don't
have
the
PR
sub
and
later
today,
I'm
going
to
push
my
PR
Maddox
Pierre
is
already
up
there.
B
Our
PR
is
from
a
Rohit,
it's
also
up
and
Iago
promised
to
send
the
PR
later
today.
So
we
are
a
bit
behind
on
schedule,
but
then
again
we
don't
really
have
that
much
new
stuff.
So
we
basically
need
to
collect
some
generic
comments.
What
we
want
to
change
from
our
cycle
and
apply
the
changes
like
later
this
week
and
next
week,
so
yeah
that
that's
all
dogs,
hopefully
we
can
have
everything
up
today,.
A
I'm
still
doing
my
sick
coaster,
life
cycle
pruning
that's
going
to
be
ongoing
for
a
while,
so
literally,
I
have
to
talk
with
stakeholders
before
we're
in
nuclear
docs
and
I
want
people
to
get
really
mad
at
me.
So
it'll
just
take
a
while,
but
hopefully
will
prune
down
that
setup.
Page
and
I
also
did
I'd
also
like
to
make
sure
that
we
distinguish
between
commercial
and
Community
Supported
stuff,
so
I
needed
that
too,
as
well.
B
A
I've
been
slowly
nuking
that
stuff
and
I'm
gonna
work
through
it
overtime
and
also
what
I
want
to
do,
especially
as
we
get
clustered
API
providers
spun
online
to
have
that
also
listed
as
community
supported
things.
You
know
are
tools
they
can
use
right.
I
want
to
I
want
to
sort
of
give
in
on
the
community
facing
Doc's
I
want
to
give
preference
to
the
community
basic
projects
there
Olinda
maintained
by
the
community.
This
is
a
Trinities
community,
niihka
system,
so
it's
gonna
take
a
while,
but
you
know
it's
it's
a
political
minefield
so.
B
A
There
are
ways
we
can:
we
can
get
the
labeling
automatically
and
that
that
forced
labeling
low
then
at
least
give
us
some
way
of
signaling,
at
least
to
the
docs
people,
that
you
know
that
they
should
have
just
blindly
approved.
It
requires
approval,
so
you
can
do
two
parts.
You
can
put
the
metadata
in
the
metadata
section
as
well
as
the
group
and
the
owners
file.
So
they
can
add
that
to
my
to-do
list,.
B
A
B
B
Also
in
terms
of
that,
this
folder,
the
mini
cube
dogs,
I,
spoke
to
I.
Think
it's
Thomas
from
Google
yeah,
they
mini
cube.
The
dogs
are
so
outdated
that
at
this
point
we
are
getting
like
a
spam
of
issues
in
the
website
about
their
dogs,
and
they
need
to
take
action
on
this
as
soon
as
possible.
I
can
force
a
cockatoo
yeah,
otherwise
we
should
basically
find
out
someone
that
wants
to
take
this
and
nuke
the
dogs
for
them,
but
ideally
they
should
have
consent
with
whoever
is
nuking
stuff.
A
B
D
Really
now
I
give
you
an
update
so
for
people
with
who
doesn't
know
what
kind
there
is
a
time
there
is
a
the
selector
arrives,
a
replacement
for
the
kebabian
playground.
It
is
a
tool
that
uses
K
kind
as
a
library
and
he
it
was
designed
to
make
the
developer
experience
for
coop
and
me
contribute
or
much
more
easy.
So
it
will
be
much
more
easy
to
set
up
and
it's
a
cluster
locally
to
test,
upgrade
to
test
the
join
control,
plane
and
so
on.
D
In
the
last
few
days,
we
managed
to
to
get
the
few
change
requiring
in-kind
method,
so
I
think
that
in
one
or
today,
I
will
push
kinda
in
the
cubed
Mini
repo
with
some
documents,
and
if
people
can
use
it
for
testing
they
114
release.
This
will
be
great
because,
as
everyone
know,
we
don't
have
singing
singing
l4h
a
and
also
for
the
utility
common
that
are
included
mean.
So
this
is
the
status
I
open.
It
I
can't
push
something
later
today,.
A
B
B
A
The
biggest
impairment
currently
with
getting
the
upgrade
skew
testing
in
kind
is
having
phases
support
they
talked
with.
If
you
have
phases
support,
you
can
break
down
the
control
of
how
you
do
the
deployment,
and
then
you
can
have
the
upgrade
as
part
of
that
deployment
scenario,
the
it
as
edit
as
a
separate
step,
but
the
that's
gonna
make
its
way
into
their
next
release,
which
is
supposed
to
be
this
week.
A
B
So
Ben
is
currently
refactoring
the
the
kind
GVK
config
basically
but
I'm,
not
seeing
support
for
executing
actions
or
skipping
actions
and
actions
are
basically
the
current
way
of
doing
phases.
Like
it's
the
similar
mechanic.
We
don't
have
a
way
to
skip
the
creation
of
a
kubernetes
cluster,
using
cube
ADM
from
the
COI
or
from
the
config.
A
Well,
there's
that
issue
that
ratio
open
did.
We
talked
about
it
too,
and
the
one
that
I
opened
so
I'll
follow
up
with
a
follow
up
with
deeper
state
understanding
next
week
with
Ben.
But
he
was
on
the
same
page
with
regards
to
making
sure
that,
before
the
next
release
that
making
sure
that
some
form
of
phases
existed
in
order
for
us
to
skip
the
actual
full
installation
process,
yeah.
A
That's
there
on
a
monthly
cadence,
so
that
will
still
be
before
you
know.
The
115
cycle
is
so.
If
we
need
to,
we
have
two
strategies.
One
strategy
is
to
push
on
kind.
The
other
strategy
is
to
support
kinder,
intestine,
prep
and
I.
Think
we
will
will
choose
that.
I
will
be
in
a
couple
of
weeks
timeframe
about
what
strategy
we
want
to
approach
and
then
long
term
still
get
to
the
point.
Where
kind
performs
the
operations
we
need
in
a
meaningful
way,
yeah
I.
D
D
D
Unfortunately,
the
the
project
is
still
really
dependent
to
the
to
the
time
that
Ben
Meir,
just
the
biggest
changes,
so
I
think
that
we
can
manage
any
114.
There
are
some
detail
data
that
I
want
to
discuss
with
loop,
Amir
or
or
who
is
interested
in
out.
We
want
to
implement
this
test,
for
instance,
for
the
upgrade
the
existing
Android
test,
basically
spin-ups
a
cluster
run
and
to
end
and
then
make
us
upgrade
and
run
again
and
to
end.
D
Do
we
want
to
support
this
workflow,
or
do
we
want
spin
up
a
cluster
make
upgrade
and
and
2's
that
big
and
to
end,
because
the
the
the
end-to-end
on
the
simple
cluster
basically
is
or
will
be
covered
by
something
else?
So
the
generalizing
is
that
we
have
to
define
what
is
the
requirement
in
order
to
understand
if
the
phases
as
they
are
designed
are
enough,
or
we
should
change,
also
something
in
the
qubit
test.
B
Yes,
so,
in
my
opinion,
we
shouldn't
test
before
upgrading.
We
should
call
it
test
after
because,
like
you
said,
we
already
have
coverage
for
the
cluster
before
a
blade
in
the
regular
tests
and
I
have
some
ideas
about
how
we
do
the
upgrades,
for
instance,
so
we
can
discuss.
Do
we
have
a
do
do?
Do
we
even
need
a
separate
image
load
image,
or
do
we
like
place
the
binaries
inside
the
existing
load,
image
and
I?
I'm?
B
Probably
gonna
go
with
the
second
option
here,
but
we
can
discuss
this
later
in
terms
of
the
road
map.
So
if
we
don't
have
a
solution
until
like
one
month
after
one
14
is
released-
or
at
least,
if
you
don't
have
a
plan,
I
suggest
that
we
start
looking
into
whatever
options
we
have
and
kind
of
can
still
be
used
as
a
back-up
plan
with
a
bash
script.
B
B
A
B
E
A
E
B
A
For
what
it's
worth
I
think
this
is
a
huge
step
forwards
for
the
project
to
be
able
to
do
separate,
CI
integration
that
doesn't
actually
force
you
to
stand
up
and
be
dependent
upon
some
type
of
provider.
So
it's
it's
good
to
see
the
progress.
I
know,
there's
simply
your
bumps
in
the
road,
but
we'll
get
there.
B
B
And
if
somebody
diverges
from
that,
we
claim
that
we
can
claim
that
this
is
not
a
supported
cluster,
because
in
this
case
we're
talking
about
not
like
somebody
replacing
proxy
with
an
ingress.
We
are
talking
about
like
core
API
stuff,
that
we
should
even
evaluate
if
it
passes
conformance
when
we
start
actually
separating
these
things
in
time.
So
this
is
my
concern.
A
So
I
have
very
strong
opinions
that
I
that
align
with
Lucas's
on
this
particular
topic.
There's
this
weird
line
in
the
sand
where
developer
productivity
for
creating
api's
is
a
very
useful
thing
versus,
like
the
user,
doesn't
care
they
just
want
to
function
in
cluster
right
and
that
that
tension
between
developer
productivity
and
flexibility
versus
sort
of
user
experience
is
a
real
tension.
I
do
think
we
should
obfuscate
it
from
the
user,
so
they
don't
see
it.
It
should
be
totally
opaque
to
them
such
that
curb
ATM.
A
So
if
you
look
at
whatever
happens
for
and
what
happened
last
cycle
with
the
testing
support
for
CSI
modifications,
which
were
a
course
er
dehydration,
in
order
for
them
to
enable
that
testing
and
to
get
that
signal
as
part
of
the
CI,
they
had
this
weird
add-on
management,
which
is
underneath
slash
cluster.
That
was
enabled
only
GCE
deployments
in
order
to
get
that
CI
single
out
that
that
I
think
is
antithetical
to
the
community
and
it's
a
total
hack
job.
A
So
I
think
it's
gonna
take
a
long
time,
I'm
totally
in
agreement
that
we
needs
to
manifest
need
to
be
baked
into
comedian
for
better
for
worse
for
some
period
of
time,
and
but
we
also
need
to
figure
out
a
way
to
keep
tracking
these
things
as
they
start
to
federate
in
other
repositories.
So,
okay,
everybody
in
the
loop
here
and
I,
don't
think
anyone
has
any
intention
of
breaking
Canadian
because
pretty
much
the
whole
community
now
depends
upon
this
central
nugget.
B
A
So
I
see
it
as
a
way
there
are,
or
there
are
benefits
for
things
like
I,
want
to
create
an
appliance,
that's
like
kubernetes,
but
has
very
trim
down
behavior,
and
if
you
you
care
about
stuff
like
that,
you
could
create
an
appliance
which
is
like
you
know,
kubernetes
in
my
I,
don't
know
blender
right,
my
my
kubernetes
blender
only
runs
pods,
but
it
doesn't
have
any
of
the
other
stuff
around
it.
That
type
of
behavior
might
be
useful
for
people
creating
appliances,
and
that's
that's
the
end
user
consumer
function
Ella.
A
A
Yeah,
there's
there's
a
there's
a
couple
different
layers
to
customize,
but
I
really
want
to
just
put
that
whole
section
of
goop
into
the
quote:
unquote
add-on
manager
and
we
just
reference
to
add-on
managers
or
library
in
the
grand
scheme
of
things
in
the
fullness
of
time.
We
don't
even
care.
We
just
reference
this
library
and
we
have
options
that
filter
through
Covidien.
A
D
In
my
opinion,
and
what
what
is
happening
so
in
increment
meaning
what
we
had
the
document
mean
generates
basically
some
static,
pods
and
create
some
config
map.
Some
are
back
rules
and
that
we
already
have
a
request
for
people
to
being
able
to
customer
eyes
is
the
Sphinx,
and
this
is
where
the
discussion
about
customized
started.
Okay,
now
the
set
of
cubed
minted
that
cramming
should
Eastern.
D
Maybe
that
will
increase,
because
we
have
to
install
certain
course
30
okay,
end,
but
from
my
side
and
from
the
other
side
might
be
that
some
XML
that
today
are
managed
by
could
admin
will
shift
to
the
add-on
site
Ladon
manager?
This
is
how
I
see
the
problem,
but
the
fact
that
the
adult
manager
will
install
the
Cooper
proxy
does
not
leave
them.
Does
not
review.
Remove
from
cubed
mean
the
defect.
The
Cooper
mean,
creates
some
resources,
so
it
is.
It
is
this
Stila
when
defected
that
what
we
do
with
our
sources?
D
A
This
is
getting
super
meta.
I
think
we
need
to
think
about
it
and
sort
of
get
to
the
point
where
things
are
actually
real
and
then
we
can
actually
think
about
breaking
down
the
problem
space
into
a
set
of
composable
units
and
having
a
flow
for
this
I.
Think
at
this
point,
the
CRD
discussion
is
totally
not
even
baked.
A
It's
like
bread,
pudding
I
would
call
it
bread
pudding,
it's
like
ingredients
right.
You
get
basically
eggs
on
your
counter.
You
want
to
make
a
cake,
so
it's
it's
nowhere
near
ready
for
prime
time,
but
I
think
we
should
still
continue
the
conversation
about
what
in
the
in
the
fullness
of
time,
what
we
would
like
to
see,
but
I
can't
without
some
actual
tangible
layering
of
how
you
how
we
envision
the
pieces
fitting
together.
D
A
A
It's
assumes
a
large
part
of
the
API.
If
you
try
to
bender
in
customize,
you
realize
that
it's
one
of
those
depth
graphs
of
do
and
we
want
to
make
sure
that
we're
careful
there
so
that
way,
it's
a
minimal,
reusable
component
that
does
just
what
we
need
to
do.
It
doesn't
actually
try
to
subsume
the
entire
communities
API
again,
because
all
we
really
want
is
partial
specialization
of
manifests
because
custom.
D
A
D
There
are
I
started
to
look
at
these.
There
are,
in
my
opinion,
two
way
to
do
this.
One
one
way
is
to
basically
keep
the
structure.
The
phases
that
we
have
now
and
for
each
phase
apply
customize
at
the
end.
I
make
an
example:
when
I
created
the
Cuban
mean
config
I
create,
could
I
mean
config,
some
Arabic
rules,
and
so
on.
This
is
a
phase
I
can
put
at
the
end
inside
his
face.
The
customized
thing
or
I
can
really
split.
The
generation
of
manifests
the
preparation
of
manifest
to
the
apply
to
the
cluster.
A
Sounds
very
proposal
II.
Well
we
when
we
work
on
proposals-
and
we
can
start
planning
next
week-
yeah
because
there's
a
couple
different
ways.
We
could
have
it
ho
about
this
one,
but
I
think
it's
highly
useful
to
the
users
right,
I,
think,
users.
We
get
asked
all
the
time
for
people
wanting
some
weird
personal
specialization
plumbs
through
who
talked
about
them
and
we
have
to
like
there's,
there's
gymnastics
with
the
config
in
order
to
make
that
work,
and
we
really
just
want
to
make
it
easy
for
them.
So
yeah
we.
A
A
A
So
that's
a
that's
a
that's
a
large
topic
and
we
should
definitely
think
about
how
we
want
to
address
that.
A
long
haul,
okay
good,
but
with
regards
to
your
CR
D
point
I,
think
we're
all
in
violent
agreement
and
that
well,
we're
gonna
bake
manifests
in
for
some
period,
indeterminate
period
of
time
until
add-on
manager
is
more
mature
and
then
we
have.
We
have
some
thoughts
underneath
their
belt
with
regards
to
how
to
do
this,
because
it's
way
to
extract
at.
A
B
E
A
No
there's
no
decision
at
all.
We
hated
all
the
options.
You
know
so.
Tim
Tims
fit
bald
and
created
several
different
things
and
we
all
kind
of
went
back
and
forth
and
I'm
not
talking
about
myself
in
the
third
person,
that's
to
popular
document,
and
then
we
all
spit
bald
and
we
hated
all
the
options.
So
I
don't
think,
there's
a
way
that
we
all
like
yeah.
It's
again
this
this
topic.
A
It's
it's
very
frustrating
to
me
because
it
was
kind
of
thrust
upon
cluster
and
life
cycle
from
API
machinery,
and
they
didn't
really
think
about
how
it
would
be
consumed
from
the
end-user
side
of
things.
It
was
more
of
an
enablement
feature
for
how
to
get
api's
to,
but
the
the
downstream
management
of
it
and
life
cycle
of
it
needs
a
lot
of
thought,
because
it
is
weird
and
thorny.
A
A
E
A
A
B
B
One
of
the
reasons
is
the
cluster
API
is
probably
don't
depend
on
that.
You
know
which
control
plane
those
joining
the
cluster
and
also
kind
paint
Venice
trying
to
do
this
in
kind
memories
to
paralyze
the
joining
of
control
plane
modes.
But
we
also
told
that,
because
we
don't
have
the
lot
for
the
planning
say
possibly
something
about
them.
I
just.
A
A
B
A
Testing
and
the
testing
is
the
big
one
and
the
other
bullet
points
that
are
the
other
issues
that
are
currently
open.
With
regards
to
HJ
I
need
to
double-check
the
user
experience
walkthrough,
but
the
in
the
ideal
world,
it's
just
like
the
it's
just
the
control
plane
join
just
one
command.
Does
it
all
so
magical
and
hand-wavy
kind
of
like
the
doctor
join?
That
is
that
is
the,
but
but
having
the
full
CI
signal
and
test
apparatus
all
built
up
behind
it
as
well.
D
A
B
A
Now,
we've
we've
always
traditionally
in
cubanÃa,
have
been
very
conservative
and
it's
been
to
our
benefit,
not
detriment,
although
some
some
commercialization
aspects
have
been
to
our
detriment,
but
for
the
for
the
project
and
community,
it
has
been
to
our
benefit,
so
I
think
we
should
continue
the
same
path
forwards
because
it
is.
It
has
proven
to
be
a
good
guide
for
how
we
approach
things,
and
it
also
helps
the
end-user
experience.
A
A
Let's
you
might
want
to
create
an
umbrella
so
like
typically
the
way
we
work
these
things
for
like
high-level
deliverables,
macro
deliverables.
We
create
umbrella
issues
and
those
umbrella
issues,
break
down
and
then
have
references
to
all
the
other
sub
elements
that
are
required
in
order
for
us
to
say
it's
been
done
done,
but
I
would
wait
until
we
actually
have
a
discussion
and
then
we
can
think
about
the
pieces
that
will
be
made.
C
A
If
you
have
ideas
of
what
you
want,
I
highly
recommend
well,
why
don't
I
update
the
notes
here
and
add
a
section
for
planning
and
then
you
know,
feel
free
to
have
topics
listed
below
put
your
name
next
to
it
and
then
we'll
what
kind
of
shuffle
sort
as
a
as
a
sub
project
and
talk
about
what
the
priorities
are.
Ok,.
B
Any
other
last
questions
comments.
Complaints,
concerns,
I
had
a
quick
one.
This
is
more
of
a
notification
for
next
cycle
because
since
I've
been
here,
the
release
team
always
pings
things
on
our
tracking
issues
in
the
communities
before
it
was
the
communities
community.
People
now
is
comedy
transmitted
and
we
have
tracking
issues
there
that
I
with
the
post
original
post.
That
I
cannot
edit
and
it's
out
of
date
and
it's
wrong.
So,
to
my
knowledge,
only
the
milestone
maintainers
can
do
it
and
I.
B
You
know:
I
posted
a
big
proposal
to
the
release,
the
toxic
release
about
this
problem,
and
also
we
have
a
general
generic
problem
of
how
we
track
the
PRS
and
issues
in,
for
instance,
in
hubei
DM.
What
we?
What
issues
do
we
have
for
this
cycle
and
also
what
PRS
and
what
is
the
testing?
Where
is
the
dogs?
This
is
a
super
chaotically
tracked
by
the
release.
Team
currently
and
I
basically
wanted
to
simplify
this,
but
this
might
take
a
while,
and
so
I
wanted
to
ask
like
how
who-who
can
edit
these
tracking
issues.
B
E
A
D
This
is
changing
because
when
we
started
with
kept-kept,
our
design
proposal
meant
to
track
the
the
function
developed
across
different
miners,
and
now
we
are
changing
to
having
kept
for
miners
having
that
I
kept
with
tracking
a
belief,
a
fracking
issue,
so
I'm
confused
about
the
role
of
the
cap,
the
commercial
deployment
of
feature
or
what
we
do
for
a
feature
in
in
our
cycle.
Yep.
A
A
B
So,
currently
the
kept
Amanda
cynic
when
you
create
the
original
proposal
for
the
feature
you
like,
dump
the
whole
information
in
the
Cape
and
like
we
are
doing
in
cube
a
DM
Fitness,
we
have
to
amend
the
Cape
each
cycle
with
the
addition
of
a
new
sub
feature
or
the
expansion
of
basically
what
we
are
working
on
related
to
this
feature
in
the
cycle.
We
have
to
add
comments.
Yes,
okay,.
A
I
think
that's
the
ideal
scenario
and
then
we
actually
get
that
all
merged
in
before
the
cycle,
but
right
now
it's
a
it's
a
bifurcated,
we're
Rube
Goldberg
machine
where
it
has
a
couple
different
parts
to
it.
There's
not
just
updating
the
cap
right
and
I
think
that's
the
problem
as
well.
There's
like
we
have
tracking
issues.
We
have
caps
that
are
not
actually
fully
defined
informed.
So
now
we
also
have
these
other
people
who
are
looking
at
the
tracking
issues
they're.
Looking
at
our
milestones
and
again
it's
not.
This
is
a.
A
B
What
we
can
do
in
the
next
cycle,
at
least,
is
to
try
to
update
our
tracking
issues
for
AJ
and
things
like
that
in
the
case.
Enhancements
repo
and
have
them
up
to
date
and
also
I,
can
propose
myself
as
a
mouse.
Tomatina
I
mean
I'm,
not
insisting,
but
if
I,
because
I
I,
like
the
release
team
ping,
zhis
and
I
have
to
pick
Fabrizio
as
a
mouse
omitted
on
our
team
or
Lucas,
where.
B
That's
also
problem
for
me.
Eventually,
I'm
going
to
burn
out,
you
know
and
I
might
not
be
always
online.
B
A
So
yeah
I
feel
free
to
PR
to
become
a
milestone.
Material
happily
approve
their.
You
shouldn't
have
to
go
through
rigmarole.
There
you've
been
very
active
throughout
the
project,
so
it's
done
a
big
problem,
but
I
do
think
the
mid-point
needs
addressing
in
that
we
have.
We
have
a
weird
non
documented
process
for
tracking
and
releasing
things
out
in
the
wild
and
that
needs
to
change,
especially
for
for
115.
So.
A
It's
one
of
those
things
like
stepping
on
it
like
who
wants
to
jump
on
the
hand
grenade
and
try
to
make
the
community
or
develop
a
standard
process
for
doing
things.
You
know
everyone's
like
no
thanks,
we'll
do
a
little
thing
over
here
and
then
now
we
have
to
duct
tape
these
two
things
together.
So
it's
not
it's
not
an
easy
thing
to
do.
Yeah
I
do
think
that
they're
making
progress,
but
it
it
will
it's
it's
very
disruptive
to
make
a
project
of
a
thousand
people
who
are
working
on
things.
B
Apparently,
there's
some
sort
of
a
tool
that
is
currently
being
developed
in
private
and
I'm,
not
sure
why
in
private
yeah,
I
accidently
think
I
need
to
go
fix
this
to
is
supposed
to
handle
caps
and
stuff
like
that,
and
and
also
the
tracking
of
the
connection
between
capes
and
parking
issues.
I'm
not
sure
how
it's
going
to
work.
B
A
Yeah
there
should
be
only
one
I
think
we
need
a
Highlander
either
you
get
caps
or
you
do
tracking
issues
you
don't
do
both
and
if
we're
gonna
use
caps
as
the
single
source
to
share
with
them.
We
should
try
to
push
that
too
long
into
the
open.
So
again,
well
well
heard
I,
hear
ya,
unclear,
I
know
it's
a
problem.
I'm
gonna
go
and
we
should
also
edge
them
else.
Imitators
and
I'll
go
and
talk
with
cig
arch
as
well
as
Joe
on
this,
because
he
also
believes
that
this
is
a
big
problem.
So.
F
F
B
So
if
we
break
somehow
the
old
branch,
it's
only
is
going
to
be
useful
only
them.
Otherwise,
if
you
merge
it
in
master,
because
it's
a
unit
test
stands
integration
test,
we
can
still
have
it
in
the
mastery
book
and
it's
useful
in
the
mastery
pod.
But
for
the
old
branch
is
not
a
problem
but
still
I,
see
I,
see
the
value.
I
see
the
value
but
yeah.
F
B
A
The
mouse
we
manually
test
them
so
that,
but
it's
not,
there
is
no
automation
around
some
of
the
stuff
and
at
the
end
of
every
cycle,
Mechanical
Turk
it
out
the
door
across
the
group
and
I
do
agree
that
the
whole
purpose
of
building,
because
because
comedian
is
this
on
this
intersection
of
integration,
with
multiple
components
that
we
started.
Building
the
intend
tests
for
carbenium,
but
we
never
actually
hooked
them
all
up
and
kind
of
mom
running
and
had
them
as
a
separate
suite
fully
to
elder
work.