►
From YouTube: SIG Architecture 20181108
Description
C
A
Well,
as
everybody
smiles
to
that
comment,
yeah,
so
as
I
wrote
in
the
agenda
at
some
point,
let
me
see
if
I
can
find
the
agenda
again.
You
know
I
think
there
are
two
at
least
two
questions
that
need
to
be
answered
there.
Actually
more
than
two
questions
that
came
up
when
I
was
looking
at
this,
but
one
is:
do
we
think
Windows
meets
the
bar
when
the
support
needs
to
bar
for
GA
and
1:13?
Does
enough
functionality
work?
Is
it
do
we
have
good
enough
tests?
A
Do
we
have
good
enough
Docs
to
explain
to
users
what
works?
What
doesn't
what
the
other
gotchas
are
not
aware
of
any
windows,
specific
features
per
se,
just
stuff
that
works
differently,
whether
it's
DNS
search,
query
paths
or
volume
works
of
volumes,
and
things
like
that
I
think
it
would
be
good
to
get
a
handle
on
that
topic
as
Polly
as
possible
and
then
with
respect
to
usability.
Actually
there's
there's
a
question.
I
didn't
really
know
down
in
the
agenda,
which
is
what
do
we
want
the
user
experience
to
be
right.
A
D
A
What
I
was
saying
yeah,
it
is
gonna,
be
an
optional
feature.
No
optional
features
are
covered
by
conformance
today,
or
so
you
know.
If
we
expect
a
small
number
of
providers
to
support
this
feature,
I,
don't
think
it's
critical.
You
carve
out
a
profile
for
it
right
now,
I'm
thinking,
we
should
sort
out
what
we
want
with
profiles.
A
C
A
E
A
Think
it
is,
and
as
I
said
before,
no
other
optional
feature
today
has
is
part
of
the
conformance
test
week,
or
at
least
that's
the
intent
and
as
an
optional
feature,
it
would
be
unique.
It
has
a
fairly
large
surface
area
compared
to
most
other
optional
features,
and
we
expect
there
to
be
demands
for
the
teacher,
but
no
providers
that
I'm
aware
of
have
Windows
or
I
could
be
wrong.
Does
that
have
window
support.
E
So
if
we
look
at
at
the
tooling,
that's
used
to
set
up
the
clusters,
yet
we've
got
automation
working
on
both
both
on
Azure
and
on
GCE.
No
commercial
supported
provider
has
announced
support
for
this,
yet
because
they're,
looking
on
the
signal
from
the
kubernetes
project,
which
is
why
getting
this
this
initial
set
marks
table
is
so
important,
because
that
paves
the
way
for
all
the
different
providers
to
package
this
up
in
support
of
long
term.
G
Don't
want
to
rush
these
decisions
and
I
agree
with
most
of
what
Brian
had
said.
I
would
really
like
to
have
an
actual
requirements
list
of
what
it
means
for
Windows
support
like
an
actual
like
hard
list
that
we
adhere
to
and
do
our
due
diligence
on
this
at
least
it's
very
clear
and
explicit
there's
a
lot
of
times.
We
have
done
that
and
we
haven't
done
our
due
diligence.
G
H
And
yeah
yeah
I
mean
just
to
sort
of
play
off
that
we
have
we
sort
of
have
a
list
in
the
list
of
ete
tests
right
that
doesn't
cover
every
feature
that
committees
has,
but
it
covers
a
lot
of
them
and
running
down
the
list
that
Patrick
and
Tina
put
it
together.
There's
clearly
some
on
there
that
are
phrased
in
a
way
that
makes
them
Linux
specific
but
I'm
not
sure
they
have
to
be.
E
Yeah
so
I've
got
a
I've
got
a
partial
list
started
and
I've
got
that
other
issue
linked.
We
see
if
it's
in
that
meeting
notes
there,
but
my
plan
was
to
go
back
back
through
and
update
those
in
the
reference
starting
to
generate
API
reference
over
the
next
couple
of
weeks
before
code
freeze,
and
so
that
that
was
one
of
the
tasks
that
I
had
planned
to
complete
there
and
so
right
now,
I've
got
an
outline
of
what
was
there,
but
I
still
need
to
get
the
PRS
in
place.
H
B
So
I
kind
of
like
the
decoupling
idea
for
sure,
but
then
are
we
saying
that
for
113,
if
in
a
cluster,
there
are
windows
nodes,
the
cluster
is
still
going
to
be
conformant
because
it
passes
the
existing
list
of
tests.
Is
that
the
approach
that
we're
going
to
take
for
113
you
mean
it?
Has
linux
nodes?
Yeah
I
mean
it
has
a
mixture
of
Linux
and
Windows
nodes.
It's
still
going
to
be
confirmed
and
we
are
not
trying
to
say
which
tests
run
where,
but
overall,
all
the
test
run.
A
A
known
issue
with
conformance
currently
and
we
need
to
decide
whether
than
what
we
might
do
about
that,
but
that's
true
for
other
kinds
of
functionality
as
well.
So
that's
what
partly
touches
on
what
I
was
saying
before
about
the
user
experience
and
how
the
test
to
run
and
do
we
expect
the
miss
you
tainted
or
labeled
or
how
we
expect
that
to
work.
I.
Think
we'll
right.
B
I
mean
this
is
just
a
1:13
statement.
Right,
say
we
we
do
decoupling
right
and
we
say
that,
okay
and
based
on
that
we'll
do
whether
it
means
whether
we
can
mark
a
ga
or
not.
So
then,
the
follow-up
to
that
is.
When
we
make
that
decision
and
say:
okay,
fine,
we
are
ready
to
go
ga
and
1:13,
and
then
we
have
to
then
make
a
statement
about
in
your
cluster.
In
addition
to
in
Linux
nodes,
if
you
have
Windows
nodes
is
still
going
to
be
confirming.
H
The
way
I
interpret
what
Bryan,
saying
and
I
think
I
agree
with.
Is
we
don't
take
a
position
on
Windows
nodes
as
with
expecting
employments
for
now
for
now
yeah,
and
when,
when
that
topic
comes
up,
we
will
have
to
talk
about
what
does
that
mean
for
people
who
are
already
shipping
Windows?
Who
might
not
be
meeting
the
current
definition
of
performance
I?
Think
we
absolutely
need
to
get
to
the
place
of
conformance,
but
I,
don't
know
we
need
to
block
it
right
now.
The
alternative
is
perpetual
beta
until
we
get
this
conforming
stuff.
F
Yeah
I
just
had
a
very
quick
question
about
what
exactly
the
message
is
that
the
Microsoft
folks
want
to
send
out
with
respect
to
conformance
so
I
mean
one
option
which
I
think
is
the
one
on
the
table
is
that
you
know
we
build
a
cluster
that
has
Linux
nodes.
We
run
all
the
conformance
tests
they
all
pass,
but
actually
they
don't
fast
using
Windows
containers,
but
then
it's
very
difficult
to
make
a
statement
that
you
know
Windows
is
kubernetes
conform
at
all
anything
like
that.
F
On
the
flip
side,
you
know
when,
when
the
conformance
program
was
first
created,
they
were
clearly
lots
of
tests
that
that
should
have
been
required
for
to
declare
a
cluster
conformant,
but
but
we
just
didn't,
have
the
test
yet
or
they
didn't
run
properly
on
enough
distributions
that
we
could
use
them
as
part
of
the
conformance
test,
in
which
case
we
just
temporarily
said.
Okay,
you
don't
have
to
pass
those
yet
we're
gonna
fix
the
tests
in
the
future
so
that
they,
you
know
better
and
and
I.
Guess,
that's
another
option
here
and
yeah.
E
So
basically,
the
signals
that
I'm
looking
at
here
is
like
when
I
look
at
what
kubernetes
calls
something
that's
beta
or
stable.
A
stable
API
is
one
that
we're
not
going
to.
You
know,
pull
the
rug
out
from
under
a
customer
and
say
you
can't
use
this
anymore.
What
we're
trying
to
say
is
the
windows.
Support
is
here
to
stay.
Here's
the
initial
support
or
supported
features
that
are
there
that
are
going
to
work
on
Windows
so
that
my
customers
can
start
deploying
with
confidence.
G
I
just
wanted
to
make
the
statement
that
seed
cluster
lifecycle
has
officially
had
the
position
for
quite
some
time
now
that
we
do
not
support
heterogeneous
deployment
environments
right.
The
reason
being
is
because
there
hasn't
been
an
official
policy
for
how
we
want
to
manage
a
mixed
architecture
style
of
deployment
right
paignton,
it's
fine!
If
we,
if
cigar,
puts
the
position
that
you
know
instant
elevation
is
what
we
will
use
for
heterogeneous
environments
and
then
we
will
go
forth
and
make
sure
that
that
is
tested
and
has
coverage.
A
So
I
have
a
specific
proposal,
which
is
the
people
who
want
to
continue
this
discussion.
We
move
the
discussion
about
what
is
the
bar
for
Windows
GA
to
kubernetes
in
Windows,
and
we
moved
the
question
about
conformance
to
the
conformance
working
group
mailing
list
and
continue
the
discussion
there.
But
currently
I'm
didn't
sound
like
there
was
strong
pushback
to
the
proposal
to
decouple
them
so
I'll
announce
that
to
both
lists
I'm.
H
A
C
F
Yeah
I
I
would
rather
not
merge
it
and
not
acknowledge
that
there
are
specific,
pretty
important
unresolved
issues
or
we
need
to
decide
they're
resolved,
because
you
know
the
chairs
have
decided
that
they
resolved
and
that's
the
end
of
the
story,
but
it
seems
we're
sort
of
in
this.
No
man's
land
in
between
these
states
so
I
think
if.
A
I
didn't
have
time
before
the
meeting
I
checked
yesterday,
but
the
additional
comments
were
posted
after
I
checked
the
two
main
issues:
I
think
our
corporate
diversity
amongst
chairs
and
sig
leadership
generally
for
whatever
sig
leadership
means
and
the
second
one
is
who's.
The
decider
on
technical
escalations
are
those
the
two
main
concerns.
Are
there
other
additional
concerns.
F
F
A
Dims,
and
so
we
had
the
earlier
email
thread,
that
you're
on
Clinton
about
flushing
out
the
sub
projects,
which
was
in
you
know
the
previous
drafts
of
the
yeah
of
the
Charter
and
we're
refined
and
simplified
based
on
feedback
jim's
PR.
That
information
in
it
is
now
in
6:00
a.m.
oh
and
in
the
readme.
A
As
far
as
tech
leads
of
all
of
the
sub
projects,
the
projects
that
have
been
operating
on
kind
of
for
a
longer
period
of
time
do
you
have
explicit
leads
for
the
API
governance.
It's
the
API
approvers
for
conformance
right
now.
The
two
people
who
are
the
approvers
are
me
and
Clayton,
and
you
know,
we've
talked
about.
We
actually
are
actively
working
to
expand
that
list,
as
people
try
participate
in
conformance
and
review
and
come
up
to
speed,
and
you
know
I'll
still
call
out
that
if
people
do
want
to
be
involved,
you
know
we.
A
That
is
the
number
of
times
people
have
said.
Well,
we
should
make
this
to
some
project.
I
mean
we
have
some
sub
projects.
You
know
if
you
want
to
be
involved
with
API
reviews
or
conformance
test
reviews.
You
know
just
reach
out
to
me
Terrance,
so
the
people
who
are
involved
with
that
and
it's
I
think
or
just
email-
the
the
cigar
mailing
list.
A
If
you
have
any
doubts
about
who
the
right
contacts
are
or
reach
out
to
me
as
the
chairs
of
the
sink
and
will
point
you
in
the
direction,
the
some
sub-projects
code
organisation
sub-projects,
is
really
unstaffed.
So
that
is
one
where
we
do
need
to
get
some
people
to
step
up
and
drive
that,
and
then
cap
has
had
very
you
know,
changing
people
driving
in
and
being
involved
with
that
over
time.
So
I
don't
know
if
the
list
of
deciders
is
up
to
date
for
that
firms
of
PR.
A
F
To
catch
a
short
because
they
never
got
a
lot
of
others
talking
the
agenda,
that's
all
fine!
The
the
crux
of
the
matter
here
is
just:
should
we
merge
a
charter
that
doesn't
deal
with
all
the
issues
that
have
been
brought
up
in
the
three
drafts
before
it,
or
should
we
explicitly
list
them
as
unresolved
issues
that
we
need
to
resolve?
You
know
sooner
rather
than
later,
yeah.
A
So
I'm
fine
with
either
of
those
decisions,
actually
the
push
to
get
all
some
flavor
of
charters.
That's
relatively
consistent
with
the
rest
of
the
project
is
sort
of
the
initiative
of
the
steering
committee.
If
we
want
to
put
it
to
do
in
there,
I
am
fine
with
that
yeah
I'm,
fine
with
that
too.
Okay,
okay,
but
again
just
make
it
a
call.
You
know
if
someone
wants
to,
we
could
use
another
chair.
A
K
A
K
A
K
E
K
C
K
I
didn't
read
that
myself,
but
okay,
if
that's
fine,
that's
fine,
that's
it
and
then
the
the
other
thing
that
I
was
just
going
to
say
real,
quick
here
before
passing
off
the
dims
is
that
you
know
we
want
a
diversity
of
chairs
on
this
and
I'm
fully
bought
into
that,
but
that
actually
requires
the
chair.
K
Job
is
a
lot
of
work
and
I've
been
chairing
SIG's
for
a
long
time
a
couple
of
years
now
and
if
you're
gonna
do
a
good
job,
it
takes
a
lot
of
work
and
I
know
that
our
chairs
are
a
bit
overwhelmed
and
they
need
help,
and
so,
if
we
want
to
get
to
a
diversity
of
it,
it's
not
just
a
title.
We
you
know
that
can
be
bestowed
on
folks
least
that's
my
two
cents.
K
It's
actually
folks,
you
show
up
and
chop
wood
and
carry
water
on
a
whole
lot
of
not
all
that
exciting
stuff,
and
so,
if
we're
gonna
get
to
that
diversity,
I
mean
we
actually
have
to
have
people
who
will
do
that
work
in
order
to
get
to
that
diversity
right,
and
so
this
is
my
challenge
to
anybody
who's.
Here,
who
wants
that
diversity
to
find
bodies
to
show
up
and
do
that?
Work
for
for
Jace's
sake
and
for
Brian's
sake?
K
B
Thanks,
that's
right,
I
agree
with
what
you've
said,
so
I
would
definitely
want
to
merge
the
Charter
as
is,
and
I
can't
take
the
action
item
to
read.
We
listen
to
what
Bryant
said
before
and
what
Quinton
raised
before
and
make
sure
that
additional
language
can
be
added
to
the
Charter
later
once
this
Rama
mergers.
F
C
For
the
two
concerns
I
summarized:
okay
also
make
a
surf
spot
posle
the
hot
take
on
this
I
would
like
to
propose
that
we
add
Matt
for
you
as
a
chair
because,
frankly
of
the
people
who
have
been
saying
they
do
things,
the
one
person
that
has
stepped
up
and
done
hours
and
hours
of
work
on
the
snow
is
Matt
and
it
also
represents
company
diversity.
Is
there
anybody
who
would
disagree
with
that
says?
Matt.
C
K
C
F
Great
idea
can
I
suggest
that
we
perhaps
give
it
maybe
until
after
maybe
two
weeks
it
would
be
great
to
find
out.
Who
else
may
be
interested
and
I
have
no
reservations
about
math
just
to
be
clear
but
I.
Think
if
we're
gonna
do
this,
we
may
as
well.
You
know
solicit
whatever
other
candidates,
maybe
might
be
digging
around
didn't
comes
to
mind,
and
there
are
others
as
well
and
see.
You
know,
put
a
list
together
and
you
know
see
what
it
looks
like
and
who
is
and.
C
So
Quentin
I
I
guess
what
I?
What
I'm
saying
is
it
we've
put
this
call
out
now
for
almost
a
year
and
one
person
has
stepped
up
materially
to
take
over
to
the
chaired
duties
from
myself
and
Brian
like
this?
Is
plenty
of
people
have
had
opportunity
to
do
this,
and
really
only
one
person
is
heated
to
call
and
that's.
I
Fine
but
I
do
agree
for
a
point
of
process.
It
would
be
worthwhile
to
have
this
happen
in
some
kind
of
written
medium
rather
than
boom.
Let's
just
suddenly
decide
it
right
now.
I
already
had
problems
with
another
sake,
doing
this
sort
of
thing
where
it
was
a
similar
situation
where
it
was
just
suggested,
and
everybody
said
yes
on
the
video
and
then
we
kind
of
just
needed
to
simmer
down
and
wait
a
week
to
make
sure
that
everybody
had
a
chance
to
totally.
C
I
J
A
How
much
you
know
I'll
just
+1
to
what
Jay
said,
which
is
you
know
anybody
who
thinks
they
might
want
to
do
the
job
reach
out
to
us,
and
we
are
happy
to
help
suggest
useful
things
you
might
do
so
that
we
can.
You
know
so
that
you
would
be
an
obvious
candidate
for
the
teacher
right
I
shouldn't
be
about.
It
should
be
yeah
the
person's
already
doing
a
job
I'm
more
than
happy
to
just
tell
the
rest
of
the
work
items
onto
their
shoulders
way.
A
K
C
C
I
A
A
But
importantly,
one
thing
I
wanted
to
do
is
look
for
things
that
need
more
clinical
oversights,
and
another
thing
I
wanted
to
do
is
to
sanity
check
whether
the
things
being
serviced
through
the
113.
These
processes
matched
the
things
that
we're
getting
surfaced
through
caps
and
API
reviews
and
whatnot,
and
whether
there
is
a
disconnect
in
because
of
the
windows
shoo
I've
been
spending
time
on
that,
so
I
didn't
have
time
to
go.
Do
the
comparison.
I
Mean
I'll
just
add,
I
feel
like
when
we
ran
I,
wanted
us
to
run
through
issues
prior
to
feature,
freeze
and
then
I
wanted
us
to
run
through
them
again
prior
to
code
freeze,
to
give
us
two
decision
points
where
we
could
kick
something
out
if
it
seemed
like
it
was
risky
or
unstable,
I
feel
like
the
only
ones
that
we
weren't
so
that
generated
a
lot
of
Windows
discussion.
I
feel
like
we
have
a
plan
for
that.
We're
happy.
I
D
Current
expectation
is
that
it'll
be
the
first
thing,
those
merged
when
the
branch
opens
back
up
so
net
nam,
113,
that's
the
I'd
say
like
80%
probability
of
missing
13
13.
At
this
point
we
are
gonna,
have
a
PR
out
for
review,
hopefully
tomorrow
or
Monday,
and
maybe
it'll-
be
something
reasonable
to
carry
or
to
merge,
but
maybe
not
we'll
see,
but
we're
not
expecting
to
make
it.
At
this
point
we
have
a
there's
a
long
discussion
of
this
Insignia
machinery.
Yesterday,
ok.
I
D
J
D
I
I
The
couplet
device,
plug-in
registration,
I
believe,
is
something
that's
going
to
be
blogged
about
during
the
113.
Like
that's
going
to
be
one
of
the
four
major
bullet
points
or
three
major
bullet
points
we
talk
about
when
we
talk
about
kubernetes,
so
I
want
to
make
sure
we
feel
good
about
that
is
in
cubit
men
going
to
GA
is
the
other
one.
I
don't
know
that.
There's
any
thing
that
sega
architecture
really
has
to
say
or
do
about
that.
I
Nothing
else
really
jumped
out
at
us
last
time
from
what
I
recall
the
only
action
item,
I
really
really
remember.
Recording
in
the
spreadsheet
here
was
that
Liggett,
Clayton
and
Tim
all
Claire
we're
gonna
go.
Take
a
look
at
the
run
as
group
feature
for
containers
in
a
pod
and
I.
Don't
know
where
we
are
that.
L
I
E
A
I
I
think
it
kind
of
goes
down
to
the
way
to
help
do
that
is
to
mandate
that
everything
is
the
cap
instead
of
the
feature,
because
caps
have
Sagarika
texture
a
little
more
baked
into
the
process
and
a
little
more
technical
oversight
or
we're
trying
to
make
that
true,
but
I,
understand
I
understand
we
are
trying
to
make
that
true.
I'm
telling
you
I
think
that's
the
Leifer
by
which
you
want
to
do
this
and
make
auditing
of
all
this
stuff
a
lot
easier.
C
A
No,
that's
fine
I
added
these
a
couple
of
weeks
ago,
yeah,
so
one
responsibility,
I,
think
of
stick
architecture
is
for
overseeing
the
technical
aspects,
including
the
health
across
the
entire
project.
So
I
started
looking
at
metrics
along
those
lines
as
related
to
reliability.
Discussion
earlier
I
actually
started
a
thread
with
some
contributor
experience
books
about
some
metrics,
getting
some
additional
metrics,
but
I
don't
have
that
data
yet
so,
actually,
I'm
not
ready
to
talk
about
this.
So
I
opted
eacher.
A
C
K
K
A
J
K
C
B
Yeah
I
set
up
some
this
thing
and
then
Tim
can
act
as
the
counter
file.
So
what
what
was
happening
in
a
few
PRS
and
issues
was
we
were
finding
that
we
had
some
steel
dependencies
and
in
a
sense
that,
for
example,
let's
take
one
issue
with
the
docker
dependency
right
when
jess
was
trying
to
implement
something,
she
figured
out
that
our
docker
dependency
was
like
a
year
and
a
half
old.
So
then
she
had
to
move
the
daca
dependency
to
the
latest
first
and
then
she
had
she
could
work
on
the
proc
mount
stuff.
B
The
similar
stuff
had
happened
in
a
couple
of
more
review,
so
I
was
trying
to
see
if
we
could
come
up
with
a
mechanism
with
where
we
take
an
inventory
of.
Where
are
we
and
how
far
are
we
behind
master
of
the
different
repositories
that
we
use
and
then
kind
of
like
go?
Looked
at
the
dependency,
the
differences
and
see
you
know
if
you're
missing
stuff
that
we
could
be
using,
so
that
was
one
way
of
looking
at
it.
B
The
other
way
of
looking
at
it
was
there
are
people
who
are
using
our
dependencies
and
building
on
top
of
that,
and
they
are
finding
that,
since
our
dependencies
are
stay,
they
were
not
able
to
do
the
things
that
they
would
like
to
do
unless
they
update
our
dependencies
first
and
then
they
use
the
main
KK
repository.
So
they
were
like
four
or
five
different
peers
and
issues.
This
kind
of
issues
came
up,
so
I
went
through
the
process
of
trying
to
figure
out
and
logging
the
issue:
six,
nine,
seven,
seven
six.
B
They
were
you
know,
the
list
is
very
long,
but
the
number
of
peers
were
was
not
that
many.
We
probably
ended
up
with
like
six
or
eight
good
PR
s
there
so
and
then,
when
we
were
trying
to
do
that,
and
we
have
to
do
this
for
every
release.
So
this
six,
nine
seven
six
seven
seven
six
was
for
one
thirteen,
because
then,
when
one
fourteen
opens
up,
we
go
through
the
script
that
I
have
to
generate
one
more
list
of
dependencies
and
so
on.
H
From
from
my
perspective,
what
I
see
is
a
stream
of
PRS
that
are
updating
dependencies
without
any
justification
for
why
and
when
I
asked
why
the
only
answers
I
got
was
well
we're
behind
like
clearly
that
not
sufficient
I've
gotten
some
of
these
too,
but
I've
seen
you
ask
why
I've
seen
Justin
SBS?
Why
I've
seen
other
people
ask?
Why
and
the
answer
in
some
cases
is
you
know
this
new
release
has
features
that
we
need
or
fixes
bugs
that
we
need
cool
I'm,
taking
I'm
happy
to
take
those
as
answers
I'm
happy
to
take.
H
H
I
don't
buy
the
we
need
to
keep
up
new
contributors
busy
argument
and
and
if
the
risk
is
more
than
just
like
bugs
breaking
like
there's
security
implications
here
every
time
one
of
these
PRS
come
through,
I
have
to
do
a
scan
of
it.
There's
only
like
five
people
who
can
approve,
go
Depp's
right
now,
right,
I
have
to
do
a
scan
of
it
to
make
sure
there's
no
nefarious
in
it
functions
to
make
sure
that
there's
no
obvious
things,
but
I
don't
have
time
to
do
a
full
code
review
yeah.
A
So
if
we
had
a
staffed
code
organization
effort,
this
is
one
of
the
things
I
would
expect
them
to
be
responsible,
for
is
to
come
up
with
the
policies
and
the
requirements
and
the
tests
and
the
review
process,
and
maybe
an
automation
to
do
this
and
the
criteria
for
when
we
update.
But
you
know,
venturing
and
all
that
stuff
related
to
it
and
staging
and
stuff
related
to
it
and
how
we
pull
in
or
the
repos
and
all
that
stuff.
A
D
C
J
You're
me
you're,
muted,
David
I
just
want
to
make
sure
that
we
can
quote
Tim
on
his
if
it
is
causing
issues
in
the
ecosystem.
We
find
that
to
be
a
reasonable
reason
to
bump
as
someone
who
pulls
in
a
lot
of
different
cube
projects
and
then
a
lot
of
other
projects
on
top
when
levels
conflict
I
try
to
choose
one
that
works,
but
if
there
isn't
one
being
able
to
womp
being
what
about
the
level
in
cube
is
really
helpful.
I'm.
H
J
M
I'm
sorry
I
started
on
I
bought
to
automate
some
of
the
PR
process
here.
So
it
was
like
watches
an
upstream
repo
and
like
creates
a
PR
and
then
updates
the
PR
with
the
change
list.
Sorted
like
I,
said,
maybe
a
little
easier.
It
doesn't
help
with
the
actual
review
problem,
but
we
can
at
least
streamline
this
a
little
and
make
it
like
more
of
a
like.
Maybe
build
a
process
around
it
and
having
it
EE
test
sort
of
I.
H
Mean
I
would
love
to
have
more
tests
around
these
dependencies,
and
that
would
make
me
feel
a
lot
more
confident
in
importing
some
of
these
things.
I
would
actually
like
to
if
we
want
to
keep
new
two
contributors,
busy
I'd
love
to
call
our
dependencies
and
just
start
re-implementing.
Some
of
the
tinier
libraries
that
we
implement
by
bears
yeah.
H
E
B
H
B
H
L
A
H
I
D
I
C
A
D
H
A
Which
is
a
problem
that
we
want
to
cap
process
to
solve
that
exists
with
the
design
proposal
process
as
well
as
the
you
know,
the
discussion
with
sort
of
Peter
out
and
people
just
go
ahead
and
then
the
proposal
will
never
get
merged.
So
we
do
need
process
wise
to
push
harder
on
getting
deciders
approvers
designated
early
and
earlier
in
the
process.
B
B
We've
been
shipping
with
it
for
a
while
now,
but
then
we
kind
of
like
got
to
know
about
this
and
the
previous
issue
that
was
open
against
Hackett
II
from
2016
he's
still
lingering
and
we
haven't
been
able
to
wrap
it
up
so
either
we
are
able
to
get
the
hey
Kitty
upstream
folks
to
change
the
license
or
we
have
to
drop
Gluster
entry
plugin.
I
don't
know
if
we
have
to
do
that.