►
From YouTube: WG API Expression Bi-Weekly Meeting for 20210105
Description
WG API Expression Bi-Weekly Meeting for 20210105
A
All
right
welcome
to
vertical
api
expression,
it's
the
5th
of
january
2021.
We
have
nothing
specific
on
our
agenda
today.
I
don't
know
if
anybody
wants
to
bring
something
up.
A
All
right
I'll
I'll
switch
to
the
ga
requirements.
A
A
All
right,
high
bench
fields,
I
think
942-
is
working
on
the
cap.
C
A
That
is
yeah.
I
think
we,
we
should
be
really
good
on
that.
I
don't
know
if
we
have
anything
still
open,
we
need
to
take
care
of
there.
A
I
think
this
just
the
one
point
I
mentioned
with
making
it
easier
for
people
to
know
if
applies
even
available
is
is
more
for
the
clients
and
not
for
this
one.
So
you
should
be
good.
B
A
The
others
might
know
if
it's
me
or
you
can,
can
you
just
understand
me
or
should
I
be
wait?
Can
I
turn
my
volume.
D
A
C
A
A
Yeah,
the
dangling
fields
is
still
only
an
issue.
I.
A
E
A
A
D
B
Okay,
you
move
it's,
it's
clearly
more
complicated
than
I
expected,
so
I
need
to
think
a
lot.
I
need
to
go
through
the
code
that
prunes
the
the
field
and
kevin
I
think
looked
at
it.
So
it's
terribly
complicated.
A
A
A
Okay,
quantity
fails
type
checking.
D
Yeah
that
one
is
for
me,
I
started
looking
into
the
issue
and
antoine
explained
it
for
me
I
didn't
start
the
code,
yet
I
will
work
on
that
this
week.
B
Okay
cool,
am
I
ready?
No,
did
you
say
something
like
that
about
about
this
theology
like
did
you
say,
do
you.
B
D
I
will
I
will
first
implement
a
solution
that
we
will.
We
already
talked
about
the
easiest
one,
but
after
you
see
that
and
it
is
merged,
we
can
think
about
the
difficult
solution.
That
is
the
actual
solution.
I
I
I
just
thought
I
didn't
like
that
easy
solution,
but
I
talked
to
daniel
about
that
and
danielle
was
happy
about
having
the
easy
implementation
first,
so
we
don't.
I
think
we
need
to
go
to
that
difficult
route
yet,
but
in
future
we
might
need
that
yeah.
A
Cool
directly
converts
you,
the
instructional,
schema
to
smd
schema.
F
Oh
hello,
hey,
so
I
think
we're
antoine
with
us
already
we
have
a
pr
out,
for
this
probably
just
needs
daniel
and
stefan
to
look
at
it
the
title's,
a
bit
misleading.
I
think
we
actually
decided
not
to
directly
convert
the
structural
schema
to
smd
schema.
Originally,
it
was
around
fixing
a
bug
where,
when
we
pass
in
the
x
into
or
string
annotation
there'd
be
some
issues,
but
I
think
we
found
a
way
to
directly
fix
that
bug
without
like
going
through
the
conversion
process.
A
B
A
B
I'm
I'm
trying
to
work
on
a
proposal
to
get
rid
each
of
the
open
api
v2,
but
I
mean
this
is
for
the
long
term.
Okay,.
F
A
Yeah
mutating
admission
controls
can
break
many
fields.
I
managed
to
get
a
test
and
I
think
a
fix
I
still
have
to.
We
still
have
to
define
exactly
what
the
validation
is,
that
we
want
to
make
right
now
it's
api
version.
This
is
not
allowed
to
be
empty.
The
fields
type
is
allowed
to
be
empty.
As
far
as
I
read,
because
we
made
an
ex
exclusion
for
that
for
backwards
compatibility.
A
I
remember
something
about
that
and
that
the
managed
fields
in
general,
like
the
psv1,
does
not
allow
it's
not
allowed
to
be
empty.
I
don't
know
what
other
validation
we
might
want
to
do.
We
have
to
talk
about
that
right
now.
I've
had
I
think.
Yesterday
I
had
problems
running
my
test,
though,
because
because
of
some
some
local
networking
issues
with
not
having
docker
so
I'll
do
what
I
usually
do
and
I
get
a
vm
with
docker
on
it
and
I
run
the
test
again.
A
It's
it's
annoying,
but
I'll
get
around
it.
So
this
should
be.
I
I
yeah
I'll,
probably
get
a
pr
out
very
very
soon.
E
A
Expand
coverage
for
atomics
I've
not
seen
this
one
before.
I
think.
F
Okay,
basically,
I
split
this
into
two
parts.
One
part
is
the
open
api
generation
where
we
need
to
generate
the
right
opening
api
spec
when
we
have
these
compositions
of
atomics,
and
the
second
part
is
the
actual
like
for
structure
merge
just
to
actually
handle
this
part.
I
have
most
of
the
stuff
working
for
the
first
or
generating
open
api
spec.
There
are
still
some
cases,
I'm
not
sure
how
smd
would
handle.
So
I
probably
need
to
consult
one
of
you
further
on
this.
A
B
E
A
B
Okay,
I
think
I
think
we're
good
on
that
z,
zaya
ziya.
I
don't
know
how
to
pronounce
your
name.
Do
you
wanna,
introduce
yourself.
E
Yeah
yeah
sure
yeah
yeah,
you
probably
pronounce
it
correctly,
the
you
already
people,
that's
confident,
hi
everyone,
I'm
ziya.
I
have
been
united
into
the
school
for
entrance
and
I
think
to
give
you
some
background
of
what
I'm
doing,
I'm
engineering,
google
and
I'm
working
on
configure
connector
like
a
project.
It's
basically
a
director
directly
to
them
to
help
customers
to
directly.
E
You
know,
provision
their
infrastructures,
it's
kind
of
like
a
terraform,
but
it's
a
more
toward
to
kubernetes
class
users,
and
so
we
have
some.
We
we
use
a
lot
of.
We
have
a
lot
of
our
own
crds.
So
most
of
the
work
that
this
group
is
doing
is
really
interesting
to
us
of,
for
example,
the
server
apply
we
already
leveraged
it
in
our
controllers.
E
So
thanks
for
doing
those
amazing
works
and-
and
I
think
I'm
trying
to
contribute
my
insight
from
based
on
you
know,
the
configured
nectar
experience
around
the
resource
reference
and
also
antoine
will
give
me
a
pretty
interesting
part,
a
stat
project,
to
about
the
immutability.
So
I
will
also
shoot
project
on
my
20
project
and
yeah.
I'm
looking
forward
to
working
with
you
guys.
F
A
You
you'll
you'll
think
up
internally
about
the
client
work
joe's
doing.
G
Yeah
I
just
I
joined
I'm
late.
If
there's
any
questions,
I
missed.
C
G
Yeah,
let's
see
so
I
think
the
big
status
is
the
the
the
cap
is
out
for
review.
I
got
a
review
from
daniel
and
antoine
before
break
and
kevin.
Thank
you.
You
got
you.
I
got
your
review
as
well.
I
need
to
make
sure
I
get
one
from
david
eads
and
then
hopefully
we
can
get
that
in
this
provisional.
The
two
outstanding
things
we
need
to
determine
are
how
we
structure
the
go
client
like
apply
request
itself.
G
I
had
proposed
a
signature
for
that
that
had
some
open
questions
and
then
we
also
needed
to
do
an
experiment
that
jenny
is
going
to
lead
to
see
if
the
community
prefers
the
go,
struct
based
builders
or
the
function
based
builders,
so
we've
implemented
it
both
ways
for
the
entry,
implementation
and
jenny's
going
to
try
it
with
coupe
builder.
G
So
I
hope
to
get
that
kind
of
re-kicked
off
here
at
the
beginning
of
the
year,
but
we're
still
we
still
want
to
do
everything
in
121.,
so
we're
going
to
kind
of
try
and
accelerate
things
and
get
them
in.
But
that's
the
current
plan.
G
And
thanks
again
everybody
for
the
cap
reviews.
If
anybody
has
questions
about
what
I
just
down
brought
up,
I
think
the
cap
has
explained
that
pretty
well
at
this
point
and
feel
free
to
dive
in
there
and
ask
questions
or
do
whatever
needs
to
be
done.
A
G
I
initially
really
liked
the
functions
and
I've
I'm
now.
I
am
now
very
much
on
the
fence.
There
were
some
people
in
the
coop
builder
community
that
were
very
strongly
in
the
ghost
trucks
camp.
So
I
was
hoping
I'm
very
hopeful
that
doing
an
experiment
with
the
community
and
getting
direct
feedback
will
help
us
guide.
That
decision,
but
yeah,
I
know
also
lava
lamp.
G
Daniel
smith
was
very
much
in
the
function
camp
as
well,
so
that
I
I
do
expect
this
to
be
by
far
the
most
contentious
part
of
the
design.
The
good
news
is,
I
have
implementations
of
both
in
tree,
so
it's
really
just
a
matter
of
getting
agreement.
A
F
We
know
that
they're,
not
I
well.
I
follow
up
with
the
release
team
last
week
on
this.
I
think
these
should
be
like
I
think,
they're
finalizing
this
week,
but
most
likely
these
will
be
the
dates
giving
their
current
estimates
right
now.
So
that's
why
I
just
put
them.
E
F
Update
them
probably
next
week
once
we
get
the
actual
dates.
Oh
one
more
thing
is
I'm
starting
to
work
on
updating
the
cap
for
getting
service.
I
apply
ga
based
on
sort
of
what
we
have
in
the
document
so
far,
I'm
just
mostly
compiling
everything.
If
you,
if
we
make
any
like
big
changes
to
the
document,
you
can
just
let
me
know
so.
I
can.
A
Yeah,
I
think
the
only
thing
I
want
to
add
is
is
to
mention
that
when,
when
the
admission
controller
breaks
management,
I
will
reset
them,
but
I
can't
do
that
whenever
you're
done,
I
think
first,
I
have
to
get
proof
of
concept
out
anyways.
Oh
sure
I
thought
that
was
already
tracked.
A
Is
it
not
tracked
in
here
already
there
was
an
admission
kept
pr
that
never
got
merged.
I
think,
or
did
it
about
tracking
the
managed
feeds,
and
I
think
we
abandoned
it,
because
we
had
the
performance
issues
and.
E
A
Okay,
then
we
we
end
early.
I
think
all
right
then
see
you
in
slack
have
a
good
start
in
the
year.
Happy.