►
From YouTube: Kubernetes WG Apply 20180925
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
A
So
as
a
consequence,
we
may
have
some
changes
to
make
on
the
task
list.
Oh
yeah,
that's
gotta,
be
intense,
could
be
intense,
so
the
one
thing
that
we
didn't
get
to
in
our
design
meetings
is
given
these
changes.
What
this
list
of
things
that
we
need
to
do
to
exit
the
feature,
branch
and
call
something
in
alpha?
Yes,
so
hey
guys,
sorry.
D
A
A
A
A
The
the
previous
design
is
to
basically
do
it
the
way
it
was
intended
to
be
done
in
the
first
place
where
it's
like
to
perform
the
apply
operation.
You
subtract
the
last
applied
state
from
the
new
applied
state
and
then
merge
the
result
back
onto
the
live
object
and
the
new
design
that
we're
proposing
instead
keeps
explicit
field
ownership.
A
If,
instead
of
storing
a
last
applied
state
for
each
workflow,
we're
going
to
store
a
field
set
for
each
workflow,
which
is
the
set
of
fields
that
that
workflow
owns,
we
can
use
this.
So
the
the
point
of
this
is
to
give
conflicts
to
the
people
who
need
to
get
conflicts,
and
we
can
use
this
data
structure
to
give
to
detect
when
somebody's
changing
a
field
that
somebody
else
thinks
they
own.
Okay.
So
you.
B
A
Will
store
which
version
the
fields,
artist,
fields,
change
from
version
to
version,
and
one
of
the
things
blocking
this
design
originally
was
that
there's
no
way
to
automatically
upgrade
like
like
convert
a
field
set
between
versions,
so
that
was
blocking
this
design
in
the
first
place.
But
we
believed
that
the
the
plan
I
came
to
solve
the
plan.
I
came
up
with
to
solve
that
problem,
for
the
old
design
also
works
for
this
design.
A
A
A
A
C
A
A
C
A
B
B
Changes
are
so
much
quicker
and
yeah
then
I
started
the
API
server
did
I
took
the
the
swagger
Jason
and
tested
the
end
points
for
her
in
creation,
deletion
updates,
like
post
poped
patch
to
eat,
yeah
and
set
the
feature
gate
for
dry,
run
or
didn't
set
it
and
looked
a
few
errors.
Are
there,
so
it's
I
can
point
on
it.
Try
run
if
a
feature
gate
is
off,
I
am
I,
don't
know.
If
that's
what
the
test
was
all
about,
because
it
was
quite
simple
Indians
and
the
feature
guide
works
as
it
should.
B
B
B
C
B
A
B
B
A
So
we
learn,
we
will
still
be
doing
something
instead
of
saving
the
last
applied,
we're
now
going
to
be
saving
the
fieldset
that
we
compute
from
the
last
applied,
but
we
don't
have
the
code
quite
yet
to
do
that,
we
don't
have
to
foment
yeah.
We
don't
have
before
it.
You
have
something
yeah,
yeah
I've
started
something
and
yeah
we
can.
You
can
all.
B
B
A
A
A
A
A
You
could
still
have
like
somebody
send
actually
actually
so
so
for
the
for
the
beta,
we're
just
gonna
or
for
the
Alpha.
We're
just
gonna
have
one
one
workflow
ID,
so
you
can't
even
have
this
problem
in
that
case,
so
I'm
just
going
to
move
this
down
to
the
top,
because
if
you
can
control
the
workflow
ID,
then
you
can
send
like
a
kajillion
workflows
that
all
claim
the
same
field.
Yeah
will
cause
problems
with
that
that
all
set
the
same
field
to
the
same
value
yeah.
So
that's
annoying
so
yeah
we
could
yeah.
A
D
A
D
D
A
A
D
D
A
Yeah
to
like
test
in
the
new,
so
we
were
right
now.
It's
set
up
assuming
that
it's
that
whatever
we're
testing
is
a
replacement
for
the
three-way
merge
operation,
but
we
were
thinking
about
it
and-
and
we
were
thinking
instead
of
I,
said
I'm
having
that
format,
maybe
the
way
we
wanted
the
way.
The
way
we
want
to
do.
The
test
is
just
to
have
a
series
of
applies,
like
you
specify
the
work
workflow
X
does
applies
that
work.
Well,
why
does
this
and
then
a
and
then
between
each
one?
A
D
A
D
A
B
B
A
E
B
B
A
A
Yeah
well,
there's
that
too
yeah,
okay,
I
think
we
are
out
of
time.
So,
let's
drop
off
now
and
me
work
out.