►
From YouTube: 20190508 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).
B
A
C
D
C
A
C
Every
for
every
four
okay,
second
I'm,
going
to
add
a
notification
having
all
the
jobs,
sending
a
notification
to
des
Cygnes,
a
life
cycle
mailing
list
after
failing
for
24
hours,
seems
fair.
What's
the
current
setting?
Oh,
it
is
scattered.
I
have
to
comentate
today
in
on
an
issue,
but
this
Model
S,
the
average.
C
C
A
C
A
C
So
if
you
want
a
su,
continue
and
I
give
a
quick
overview
of
how
to
debug
the
new
kind
of
test
jobs,
I
think
that
it
would
be
useful
to
everyone.
So
basically,
there
are
a
few
point.
To
keep
in
mind
is
the
first
one
is
that
it
jobs.
Each
job
run
a
sequence
of
tasks
where
a
task
is
a
commander
with
arcs.
C
They
are
defined
in
to
admin
repo,
and
this
is
an
example.
They
are
really
simple.
For
instance,
this
is
the
task
list
for
the
upgrade
job.
The
first
job
is
to
pull
a
base
image.
Then
we
start
adding
the
kubernetes
versions
of
this
image.
So
each
task
is
a
Israeli,
a
common
kinder,
be
the
kinder,
create
cluster
or
kind
of
doing
it.
So,
and
this
is
the
and
using
these
we
can
basically
flat
test
whatever
workflow
we
want
and.
C
You
have
the
useful
vlog
where
everything
is
loaded,
but
if
we
look
at
artists
under
artifacts,
we
have
a
separate
log
for
each
step.
So
if
you
want
to
check
out
how
it
went,
you
don't
have
to
search
a
into
the
log,
but
you
can
look
and
single
step
more
easily,
and
then
you
have
all
the
end-to-end
test,
jobs
and
and
all
the
logs
from
the
from
the
cluster,
and
the
last
point
is
important
to
know
that
it
is
really
is.
C
This
is
now
really
easy
to
reproduce
locally
a
the
test,
because
basically,
what
we
have
to
do
is
to
do
locally
on
your
machine,
a
kind
of
test,
workflow
and-
and
you
call
the
same
word
four
or
five-
that
it
is
used
by
the
micro.
So,
by
the
end
to
end
you
can,
even
if
you
want
to
run,
they
hold
the
entire
job.
Luckily
you
do
it
this
way.
C
C
E
E
The
container
and
time
provision,
but
now
kind,
because
we
are
using
kind
as
a
packet
in
our
tests,
kinds
has
moved
to
container
D,
which
means
that
basically,
this
is
a
request
to
cause
this
issue
that
I
created,
because
we
now
already
have
Krakatau
I
mean
not
yet,
but
pretty
soon,
once
we
update
to
the
latest
kind,
we
can
close
this
issue,
so
this
like
we
when
we
did
the
backlog
triage.
We
add
this
array,
this
agenda
item,
but
I
just
wanted
to
give
the
heads
up
that
we
can
cause
this
prism.
A
I'm
saying:
should
we
I'm
asking
you
do
you
feel
like
we
should
open?
We
should
we
leave
this
open,
verifying
the
other
signals
from
the
other
groups,
but
there's
nothing
like
if
there's
like
a
little
bit
of
concern
here,
if
we're
using
kind
for
everything
on
the
back
end.
Now
we're
no
longer
have
CI
signal
coverage
for
doctor
from
convenience
perspective,
reusing
container
D.
Is
that
correct?
A
E
C
A
D
A
B
I
almost
think
that
this
is
worth
I
know
Luba,
Marin,
Fabrizio
audience
and
the
kind
calls
I
show
up
occasionally
I
almost
feel
like
this
is
worth
filing
an
issue
or
staffing
to
produce
kind
that
can
support
you
said.you
hard.
The
hard
switch
to
be
married
have
been
was
talking
about
it,
but
I
know
I
made
that
I
know
they
done
that,
but
I
kind
of
feel
like
if
kind
is
gonna,
be
this
baseline
test
layer
which
really
wants
to
be
and
I
think.
A
E
A
A
We
there's
goo
inside
of
comedian
that
always
gets
revved
every
time.
Docker
updates
right.
You
always
eat
it,
just
because
of
the
vision,
itis
and
the
supported
versions,
because
Dockers
been
known
to
break
over
time
and
we've
eaten
it.
So
we
always
had
that
code
in
there
for
us
to
make
sure
that
we
had
valid
versions
that
we
knew
we
had
CI
Sigma
for
and
because
we
embedded
those
specific
versions
as
part
of
the
deployment
with
Korea
days
anywhere
for
a
better
than
worse.
A
E
A
Here's
a
here's,
a
simple
workaround
if
we,
instead
of
using
kind
for
all
the
base
setup,
is
fine,
but
we
could
have
a
very
simple
job
that
only
just
install
stocker,
a
single
node
in
it
which
does
conform
us
like
it's
a
single
node
run.
We
don't
have
to
have
multi,
don't
run,
but
it
just
verifies
that
docker
it
didn't
break.
That
is
it.
A
So
we
well,
we
can,
we
can
specify
you
can
specify
the
regex
to
get
around
this
and
there
I
believe
there's
an
option
to
that
exists.
If
we
done
what
we
should
plumb
it
to
get
around
this,
because
there's
no
reason
why
you
can't
invoke
the
intent
test
apparatus
to
get
around
this
particular
problem,
but.
A
H
H
A
I
wouldn't
mind
switching
container
D
to
the
default,
but
I
do
think
the
I
think
we
should
do
what
the
users
ask
right
and
because
they
they
drive
our
success
and
most
of
the
users
still
use
docker,
so
I
think
no
plans
at
the
moment.
Yeah
we
have
the
option.
We
had
a
separate
pages
for
multiple
CR
eyes.
Many
user
can
always
choose
CRA
Ivor's,
you
know
cry
over
sweetie,
nerdy
versus
Tucker
and
that
still
exist.
What
what
I
can
do
is
that.
C
C
H
C
D
H
E
E
E
G
A
The
TLDR
for
the
they
ask
from
api
machinery
eons
ago,
was
to
create
a
separate
staging
repo
for
blue,
strap
tokens
and
their
utilities
because
they
didn't
want
it
and
client
go
even
though
I
kind
of
disagree
with
that
idea,
but
the
that
would
mean
for
lifting
everything
out
of
V
1,
beta,
1
and
referencing.
It
externally.
E
C
E
C
A
A
D
E
A
B
I
I
started
calling
it
structured
output
because
it's
really
kind
of
what
we're
describing,
whether
it's
for
machines
or
somebody
else
I,
don't
care
yeah.
So
I
just
wanted
to
thank
ed
for
doing
so.
Much
work
on
this
and
being
upfront
super
professional
about
when
I
kind
of
showed
up
and
said,
hey,
I'm
thinking
about
submitting
a
design
doc
on
this
I
know.
You've
done
a
lot
of
work,
you
don't
care,
do
you,
and
now
he
was
great.
We
who
had
a
long
chat
yesterday
and
thought
about
a
lot
of
different
things
together.
B
B
This
is
how
I
encountered
this
issue
and
I
understand
that
cube.
Atm
token
list
allows
you
to
print
the
joint
or
create
allows
you
to
print
the
joint
command,
but
that
creates
a
new
token.
So
anyway,
there
are
all
sorts
of
other
reasons
why
this
is
needed.
We
have
software
such
as
terraform
ansible,
whatever
that
might
need
to
parse
it
cluster
life
cycles.
Upgrade
tool
is
going
to
want
to
be
able
to
to
parse
this
output
in
a
deterministic
fashion.
B
So
the
stated
goals
I
tried
to
be
vague
at
first,
but
I
did
get
a
little
bit
more
prescriptive,
but
I
did
include
not
necessarily
limited
to.
Basically,
it
says
that
cube
ATM
should
support
JSON
llamo
go
template
and
text
which
is
I'm
defining
text
as
the
current
behavior
you'll
notice
that
it's
divided
into
buffered
and
unbuffered.
That
doesn't
necessarily
mean
what
you
think
I'm
not
going
to
get
into
the
whole
details
here,
but
there's
a
link
essentially
buffered
could
mean
it
doesn't
necessarily
mean
all
of
the
commands.
B
Output
at
once
cube
token
cube,
ATM
token
list,
for
example,
buffered
may
just
mean
I'm.
Sorry,
unbuffered
may
just
mean
a
token.
At
a
time
it
doesn't
mean
a
byte
at
a
time,
but
if
command
emits
multiple
objects,
it
may
just
mean
an
object
at
a
time,
even
when
it's
unbuffered.
These
are
some
of
the
commands.
That
would
benefit
from
structured
output.
But
again
it's
not
an
exhaustive
list.
So
this
is
where
it
gets
interesting
and
I
really
struggled
with
how
to
structure
this
bluebird
mayer
was
was
of
great
help.
B
So
was
so
it's
Eddie,
so
basically
the
non
goals-
and
this
is
one
difference
from
Ed's
original
design,
doc
I'm
not
proposing
we're,
not
proposing
a
short
variant
for
text
output.
If
a
command
wishes
to
implement
that
as
a
flag,
they
can
simply
do
so
by
using
a
default
code,
template
they
use
for
a
short
output
and
right
now
the
only
command
that
currently
implements.
B
That
is
version,
and
you
all
three
of
these
things
are
the
same,
but
I
listed
them
separately,
because
I
was
getting
some
feedback,
I,
don't
think
it
indicated
I,
don't
think
people
realize
these
were
the
same
or
they
all
linked
to
the
same
cause.
An
on
goal:
we're
not
trying
to
achieve
parity
with
cute
cuddles
output
flat
right
now.
B
We're
really
only
focused
on
Jason
yam
will
go
template
and
text
we're
not
trying
to
prescribe
versioned
output,
although
I
kinda
isn't
I'm
kind
of
in
favor
of
it,
but
it
would
add
a
lot
of
technical
debt
and
increase
the
scope
and,
most
importantly,
because
somebody
asked
about
this
already-
they
said.
Oh,
why
aren't
you
using
the
printers
from
the
CLI
runtime
package?
This
actually
wasn't
called
out
as
an
on
goal
to
about
20
minutes
ago.
B
B
The
CLI
runtime
package
is
what
provides
cube
cuddle
with
its
various
output
formats,
but
every
printer
in
that
package
expects
a
runtime
object
which,
whatever
is
emitted
by
cubed
ATM,
currently
is
not
they're,
not
runtime
objects,
sometimes
they're
just
text,
but
they're,
not
runtime
objects,
not
all
of
them,
the
and
so
to
introduce
run
time
objects
as
the
basis
for
cube.
Atm
output
raises
the
question:
do
we
want
versioned
outlet,
I,
actually
kind
of
think
that
the
entire
basis
of
this
step?
B
The
fact
that
we
want
structured,
deterministic
output
indicates
that
versioned
output
would
be
a
good
idea,
but
I
I
put
that
the
you
know,
both
ed
and
I,
agreed
that
this
might
not
fit
within
the
scope
of
this
cut,
but
be
a
a
second
kept
once
this
is
in
place
and
to
not
scope
it
or
to
put
it
as
a
non
goal,
should
not
indicate
a
judgment
on
versioned
output.
But
just
we
acknowledge
that
that
design
requires
a
much
broader
discussion,
but
I
think
we're
already
made
a
comment
and
basically
translated
to
versioned
output
yeah.
B
But
we
don't
do
it
today
and
I.
Don't
know
how
that
would
necessarily
well
that's
a
much
broader
discussion,
so
the
user
stories
I
just
kind
of
get
into
worse.
There
are
some
examples,
and
these
wouldn't
be
scripts.
Necessarily
it's
just
a
way
to
outline
how
the
output
would
be
used
in
a
friendly
fashion
and
they're
predicated
on
the
existence
of
some
additional
data
that
I
list
up
here.
B
Basically,
like
a
friendly
bootstrap
token,
I
know
there's
already
a
token,
but
this
is
just
to
understand
what
fields
would
be
available
to
a
go
template
and
then
here
are
some
functions
that
we
would
add
to
those
templates.
Node
0
will
give
you
the
address
and
port.
If
you
want
to
do
a
cube,
ATM
join,
say,
sir,
is
the
hash
and
join
is
just
like
go
list.
It
gives
you
a
joint
command,
real,
quick,
so
the
actual
design
details.
B
Oh,
let
me
go
with
this
user
story
real
fast.
You
can
see
the
difference
between
listing
token
IDs
versus
listing
them
as
they're
discovered.
This
see
they
both
use
go
templates,
but
by
adding
this
stream
flag.
Basically,
it
tells
the
command
to
just
print
but
send
the
tokens
to
the
printer
every
time
they
discover
one
a
versus
sending
a
list
of
tokens
and
then
allowing
the
template
to
iterate
over
it.
So
the
design,
it's
very
much
slimmer,
latter,
what
CLI
runtime
does
or
what
you
know.
B
Basically,
whether
it's
an
interface
or
a
single
exported
function
for
the
package
or
the
signature
of
a
printer
should
basically
be
something
like
this
except
the
writer
a
format,
and
this
is
not
likes
printf,
but
rather
Jason
yeah
Mel
texts
go
template
and
data,
which
is
an
interface
I,
know
how
people
feel
about
interfaces,
but
for
the
purpose
of
the
initial
design
it
works.
So
this
is
just
a
very
pseudocode
implementation.
B
Essentially,
if
the
interface
is
a
reader,
an
I/o
dot
reader,
then
we
assume
that
they're
streaming
the
output
and
so
the
like.
It's
just
text
default
behavior
today,
right
you
give
me
a
reader
I'm
just
going
to
copy
from
the
reader
to
the
writer.
Otherwise
we
look
at
the
output
to
decide
what
to
do
with
it,
whether
it's
Jason
and
yam
or
text,
and
if
it's
none
of
those
it
treats
the
format
at
sorry
yeah.
We
look
at
the
format's
that
decide
what
to
do
with
the
data.
B
If
the
format
is
JSON,
we
coded
this
JSON
yamo
as
yeah
mul
text.
We
just
print
it,
and
otherwise
we
treat
it
as
a
go
template
and
we
execute
the
template
against
the
writer.
So
how
will
the
commands
actually
do
the
printing?
You
know
there
are
a
couple
of
possibilities.
You
know
they
could
either
receive
a
printer
object.
They
could
send
the
results
to
some
exported
function.
They
can
extend
instantiate
a
printer
directly.
B
This
is
one
that
I
tried
outlining
more
specifically,
but
it
got
way
too
prescriptive
and
I
didn't
want
people
to
make
observations
based
on
what
was
essentially
just
pseudocode,
but
essentially
a
command
could
either
receive
or
return
a
channel
on
which
a
result
is
placed
and
that
result
could
be
an
I/o
reader
or
that
same
channel
could
be
used
to
block
until
the
actual
buffered
result
is
received.
So
obviously
that
differs
based
on
how
the
printers
are
implemented.
B
So
the
cap
just
includes
a
link
of
related
works,
and
you
know
edie
like
we're
going
to
sort
of
let
this
cap
stand
in
for
his
original
design,
doc
buffered
versus
on
buffered.
We
kind
of
went
over
that,
and
this
is
that
subject
where
I
think
we
already
covered,
but
the
parity
with
cube,
cuddles
output,
flag-
and
these
are
just
some
example-
outputs.
B
So,
for
example,
if
you
did
in
it
oh
JSON,
you
would
see
enough
information
that
would
allow
you
to
construct,
cube,
ATM,
join
commands
and
a
raw
for
commands
that
do
have
a
lot
of
data,
so,
whether
it's
Jason
or
yamo,
or
even
a
go
template
that
we
give.
We
would
allow
them
to
include
the
actual
raw
data
if
that
was
useful
for
them.
B
So
risks
and
mitigation
I
put
Thank
You
Ross
for
indicating
about
the
tab,
writer
I
just
added
as
a
note
or
design,
and
it
should
have
been
a
risk
so
not
including
support
for
version.
The
output
could
possibly
result
in
an
applied
judgment
of
said
version
output,
but
that's
not
the
reason
so
in
the
alpha
to
beta
graduation
criteria.
B
It
lists
that,
if
there
were
discussion
about
whether
versioned
output
should
be
scoped
is
required,
the
second
risk
that
we
noted
was
it
might
appear
that
this
ket
breaks
compatibility
with
existing
output,
such
as
token
list,
which
is
tab,
a
tabular
output.
But
that's
not
the
case,
because
you
could
just
wrap.
You
were
a
writer
in
a
tab
writer
and
then
usually
go
templates
to
print
things,
and
you
know
columns.
So
the
tab
writer
will
know
what
to
do
with
it.
B
So
test
plan
is
nothing
fancy
just
match:
actual
output
against
expected
output
and
then
an
automated
pipeline
that
connects
cube,
ATM
and
NIT
to
a
cube
ATM
joint
through
parsing
that
output
and
constructing
the
join
command
alpha
2
beta
is
pretty
much
what
you
would
expect
topic
aversion
up,
but
these
commands
should
implement
structured
output,
and
this
is
just
kind
of
standard,
stuff,
GA
and
again
very
standard
stuff,
and
so
this
is
my
first
cat.
So
this
is
all
learning
experience
for
me,
I'm
sure,
they're.
You
know
two
dozen
things
that
I
missed.
You
did.
A
I,
like
this
a
lot
I
appreciate
all
the
thought
that
went
into
this
one
question:
I
have
for
the
peanut
gallery
or
deserve
buddy.
Is
that
if
you
screw
up
the
top
you're,
basically
putting
argument
modifiers
for
the
output
and
I,
don't
know
if
there's
yeah
I,
don't
know
if
there's
precedent
for
how
we
graduate
argument
modifiers,
whether
or
not
we
usually
in
the
past,
we
did
experimental,
which
sucks
but
I,
don't
know
it's
that
see
if
there
is
if
there
is
any
other
way
that
we
do
that
for
graduation.
So.
B
Real
quick,
the
reason
this
is
a
cap
at
all
I
again,
first
get
I,
don't
know
if
I
still
have
the
kept
rules
up
here
somewhere,
but
the
last
point
of
a
cap
was
if
it
impacts
the
user
experience
or
they
will
come
to
rely
on
it.
So
I
felt
like
this
is
something
that
on
which
they'll
come
to
rely.
I,
don't
know
in
a
broad
sense
about
argument.
E
A
Technically,
probably
not
because
well
I
want
to
talk
with
people
before
we
did
that
I.
Don't
think
that
follows
I,
believe
the
category
of
the
ability
to
have
an
exception,
because
the
people
would
rely
on
it
immediately.
So
we
wanted
to
make
breaking
changes.
We
don't
reserve
that
right
and
what
I
want
to
show
them
up
go
ahead,
yeah.
What
I
want
to
make
sure
we
do
is
that,
like
first
iteration,
we
will
probably
make
breaking
changes
in
the
future
before
we
want
people
to
rely
on
it.
I
was.
B
Going
to
suggest
that
stream
should
definitely
be
under
experimental,
but
if
you
do
that,
then
that
absolutely
means
that
oh
should
be
as
well.
Although
I
do
believe
that
I,
don't
believe
that
what
Mir
said
and
what
you
said,
Tim
are
mutually
exclusive:
I
think
that
oh
can
be
experimental
because
a
command
is
implemented
it,
but
I
still
think
it
should
also
be
reserved
I,
don't
think
any!
You
know
core
kubernetes
CLI
should
implement,
oh
as
something
other
than
output
at
this
point,
but.
B
Yeah,
it
was
basically
you
know,
I
didn't
know
how
that
cap
should
work
for
this.
This
is
where
I
struggled
I
feel
like
there
is.
A
cap
doesn't
really
seem
to
want
you
to
say
here's
a
phase,
one,
here's
a
face,
I
think
it's
more
like
if
you
have
a
face
that
would
be
a
separate
cap
later,
but
I
basically
wanted
to
say
that
if
this
cap
should
be
adjusted
to
say
that
we're
gonna
version
output
from
the
very
beginning,
then
I
need
to
adjust
the
cap.
So
maybe
it
should
even
been
before.
B
This
is
implementable.
That
discussion
needs
and
that
wasn't
a
graduation
criteria.
I
guess
that's
more
of
an
accepted
as
implementable
criteria,
but
the
topic
of
versioned
output
should
be
thoroughly
and,
like
I,
said,
I.
Think
I'm
in
favor
of
it
I
think
there's
a
unique
situation.
I,
don't
necessarily
think
that
types
defined
in
this
package
or
these
packages
will
be
then
consumed
programmatically.
What's
what's
more
likely?
B
B
Okay,
yeah,
but
I
don't
yeah,
but
that
would
just
be
programmatically.
They
wouldn't
be
showing
cube
code.
Although
I
guess
it's
the
same
problem,
the
same
issue,
you're
right
and.
A
B
B
Although
I
mean
honestly,
the
premier
isn't
that
what
we're
talking
about
doing
anyway
for
this
to
be
useful,
whether
it
has
an
API
version
or
kind
people
still
want
this
to
be
deterministic
they're,
the
one.
The
one
thing
that
I
would
say
about
that
is
as
long
as
we're
not
making
redacted
or
destroy
our
changes
or
renaming
things
then
not
having
the
schema
as
long
as
it's
always
forwards,
compatible
works,
but
I,
don't
necessarily
know.
We
can
know
that
for
sure
and
it
dance
there's.
E
A
A
B
Done
that
pretty
recently,
so
so
in
my
mind,
I
could
probably
you
know
at
least
with
the
boilerplate
in
the
place
in
a
day
or
two
I
mean
that
doesn't
remove
the
amount
of
work
that
has
to
be
done
to
make
sure
the
design
is
correct.
But
if
you're
just
concerned
about
somebody
doing
that
in
front
work,
I
can
do
that.
I.
A
B
C
I
may
comment,
first
of
all,
I
really
like
a
document
and
I
think
that
the
the
first,
the
first
point
that
we
have
to
come
up
with
a
good
solution
is
out
to
get
the
information
flow
from
the
phases
to
toward
where
we
are
sterilizing.
That
is
what
you
call
it.
The
coupling
the
couple
come
under
from
printing,
and
it
is
this.
The
first
part
that
that,
in
my
opinion,
should
be
addressed
said
because
it
has
the
the
huge
impact
on
the
code
base
if
I
think,
for
instance,
to
upgrade.
C
C
My
free
point
is
that
a
good
document-
probably
the
trickiest
part-
is
the
the
part
of
gate
they
for
measure
out
from
the
from
the
various
command
for
the
topic
about
version
API.
In
my
opinion,
we
should
aim
to
have
a
version
eight
API
in
in
the
end,
but
this
should
be
a
different
API
groups
under
could
mean
know.
C
B
A
A
I
So
epic
Western
regarding
the
cap
as
well,
is
it
in
the
scope
of
this
cap
to
also
provide
a
way,
or
at
least
give
a
hint
on
how
errors
are
going
to
be
reported
or
warnings?
Aside
from
the
optimistic
approach
of
assuming
that
the
operation
will
work
and
is
tough,
it's
the
the
parts
that
I
care
about
of
the
output,
so.
B
That's
a
good,
that's
a
good
point.
I
was
operating
under
the
assumption
that
the
successful
output
goes
to
standard
out
and
any
errors
are
going
to
standard
error.
So
if
you're
parsing,
if
your
shell
exacting
this,
you
should
be
paying
attention
to
both.
But
if
it's
I
mean
people
using
cloud
in
it
can
redirect
things
appropriately.
However,
there's
no
reason
that
the
output
objects,
if
we're
using
version
freezing
whatever
the
object
is
right,
couldn't
include
basically,
instead
of
calling
it
raw
just
call
it
standard
out
in
standard
error
and
having
an
exit
status.
B
So
the
basically
the
output
would
include
metadata
that
you
might
see
from
the
process.
You
know
exit
status
standard
out
there
there,
but
then
also
the
additional
data,
but
again
to
me,
the
the
successful
out
result
is
when
you
parse
standard
out.
Otherwise
your
parsing
standard
error,
because
you
got
a
nonzero
exit
code.
That's
how
I
kind
of
looked
at
this!
B
Keep
it
key,
try
to
keep
it
simple
and
that's
generally
how
I
think
that
that
process
works
but
I.
It's
not
a
web
application
where
you're
kind
of
wanting
both
in
one
data
model,
I
think,
but
you
could
certainly
treat
it
that
way.
If
you
want
to
do
I,
don't
know
which
is
you
know
more
right
or
wrong
or
really
there's
one,
that's
more
than
the
other.
A
Yeah
I
think
I
think
we
need
to
percolate
on
I
really
appreciate
the
thoroughness
of
the
proposal
and
looking
everybody
it
together.
I
think
to
think
about.
The
implications
is
going
to
take
a
little
bit,
I,
don't
given
the
time
frame
and
the
fact
that
a
bunch
of
us
are
gonna
be
out
for
pucon,
we'll
be
out
of
pocket
for
a
while
I,
don't
know
whether
or
not
it
would
be
possible
to
get
major
changes
in
place
for
wonderful
good
thing.
A
B
Kenny
already
called
that
out
and
said
this
is
past
the
deadline
for
accepting
new
work,
but
you
could
file
an
exception
and
I
said:
hey
we're
discussing
this
this
morning.
I
don't
know
that
he
put
the
actual
deadline
on
it.
I'm,
looking
at
the
enhancement
exception
it
doesn't,
that
page
doesn't
have
the
date.
It
looks
like
well.
A
A
A
B
The
only
way
I
comfortable,
including
it
and
once
the
team
would
be
without
versioned
output,
with
no
promises
about
how
this
is
moving
forward
and
then
then
come
to
some
agreement,
whether
to
include
or
not,
and
it
seems
like
we're
going
towards
include
version
output
for
1/16,
but
based
on
what
I
just
said,
the
way
I
said
it.
If
we're
not
going
to
include
some
type
of
promise
moving
forward,
it's
really
no
better
than
what
we
have
now,
because
it
doesn't
just
need
to
be
structured.
E
F
B
E
B
I
pre
mean
you
know,
you
gave
me
a
lot
of
good
guidance
there,
even
though
you
said
I
don't
know
if
this
should
be
a
captain.
Look,
you
know
what
no
one
it.
No
one
got
fired
for,
buying
IBM
java
or
writing
a
cat.
So
if
we
want
to
discuss
it,
then
let's
make
it
a
cat
yeah
thanks
and
thanks
again
to
add,
because
you
know
we
had
a
chat
yesterday
and
I
really
appreciate
the
work
he's
done
on
this.
J
Would
say,
I
don't
know,
I
I
would
agree
with
Andrew
with
like
we
should
start
with
with
smaller
scope,
because
it
really
requires
a
lot
more
work.
If
we
like,
we
could
aim
for
it,
but
for
like
near
future,
I
would
say
we
would
be
good
without
it
as
it's
much
better
than
what
we
have
now.
Just
just
because
of
that.
That's
my
opinion.
A
Well,
we
have
t-minus
four
minutes
and
I
think
it
requires
some
in-depth
thought
about
the
implications.
So
why
don't
we
rally
in
the
cap
folks
can
review
it's
greatly
appreciated,
more
more
voices,
more
ideas,
more
thoughts,
the
better,
so
it's
linked
in
the
main
dock
or
it's
linked
from
there's
a
link
in
the
link
from
the
main
dock.
So
maybe
we
should
just
link
it
for
an
action
upon
and
we
can
discuss
again
next
week
that
seemed
reasonable,
more
folks,
I,
don't
know
if
we're
gonna
get
to
all
the
other
stuff.
A
C
It's
here,
let's
try
so
I'm
spending
a
lot
of
time
on
the
certificate
already
you
are,
and
basically
what
what
I
need.
I
want
to
add
a
new
commander
that
allows
the
user
to
show
the
to
be
aware
about
the
certificate
aspiration,
because
we
are
every
day
there
is
a
one
pair
so
that
that
is
it
by
provider
it
by
this
program.
So
I
have
one
proposal,
that
is
to
add
a
new
command
under
alpha
search.
E
Maybe
maybe
we
can
go
with
the
first
command
that
first
helo
cube
ATM
officers
as
the
primary
or
targets
for
this
release
and
don't
touch
the
other
documents
follow.
Yes,.
A
C
C
A
Would
open
an
issue
to
discuss
that
one?
We
are
actually
at
time.
I
want
to
make
sure
we
have
to
stop
because
we
have
like
we
have
collisions
with
the
zoom
account.
So
that
means
they're.
Happy
folks
would
are
waiting
on
us
currently.
So
could
you
open
an
issue
with
that
particular
one
and
we
can
discuss
it?
Yes,
we've
opened
a
few
we
fold.
The
other
point
all.