►
From YouTube: Kubernetes SIG API Machinery 20190410
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
The
cozy
Megan's,
okay,
hi
everyone-
this
is
the
April
10th
API
Machinery
meeting
we
just
had
Jordan
give
us
a
great
walk
through
of
go
modules,
which
we
published
in
a
separate
recording.
So
go
see
that
if
you
didn't-
and
we
have
a
few
things
on
the
agenda
and
not
very
much
time
so
I'm
not
expecting
to
get
through
this
entire
list,
but
we'll
try
to
keep
it
brief
on
the
caps-
and
maybe
it's
possibility
so
I
placed
at
our
cups
and
put
a
few
on
the
list.
A
That
seemed
to
be
worth
it
to
talk
about.
Let's
do
the
Union
one
of
last,
because
maybe
you
know
take
migrate
or
no,
so
a
server
side
get
to
implementable.
It's
been
open
for
like
a
week.
Nobody
had
anything
had
to
say
about
it,
so
I
just
held
it
and
it
merged
I
looked
at
it,
but
if
you're
interested
in
that
certainly
go
look
at
it
rate.
Limiting
the
the
group
of
there's
like
five
of
us.
A
Six
of
us
that
have
been
having
discussions
about
that
I
think
we
we
have
enough
agreement
to
to
work
on
work
on
the
first
draft
of
this,
and
so
I
just
made
a
feature
branch
for
that
this
morning.
There's
some
caps
outstanding,
I'm,
pretty
sure
it's
my
Kennedy
go!
Look
at
those
and
a
review
or
something
or
very
dry
I.
Don't
know
it's
probably
my
turn
but
they're
there.
If
you
want
to
go,
look
at
it
and
there's
a
feature
branch
that
we've
just
made.
A
So
if
you're
just
contributing,
maybe
say
something
on
the
mailing
list.
Just
in
the
chatter
there
about
the
feature:
branch,
okay
and
last-
there's
other
keps
out
there,
but
the
last
one
that
I
put
on
the
list
is
Union,
one
of
which
is
building
on
those
server
side
by
work.
Building
on
the
notion
of,
as
you
know,
that
we
add
in
there
to
really
implement
a
structured
notion
of
a
union.
B
A
One
there's
this
many
fields,
but
you
can
only
accept
one
of
them
that
most
so
I
don't
know.
We
want
to
talk
about
it
very
much
here,
especially
since
we
don't
have
that
much
time,
but
I
would
appreciate
your
comments
on
it
since
because,
if
we're
applying
looked
in,
it
makes
it
easy
for
us
to
implement
this
and
have
it
just
universally
and
consistently
apply
throughout
the
communities
type
system.
So
even
if
you're
going
through
the
update,
not
like
you
can
do
intelligent
things
now,
he's
got
the
right
tags
in
schema.
A
All
right
cool,
you
can
all
go
look
at
the
caps
whenever
you
feel
like
it.
We
don't
have
to
wait
for
them
to
show
up
on
the
meeting.
Okay,
okay,
today
I
wanted
to
propose
we,
my
team,
fed
his
team
at
Google,
we've
been
looking
at
the
bugs,
and
issues
are
timepi
ours
and
issues
assigned
to
ziggy
machinery
for,
like.
A
Somebody's
typing
in
it's
quite
loud
stay,
be
anything
and
I.
Think
it's
been
helpful,
and
so
the
idea
is
what,
if
we
just
open
it
up
we're
not
doing
a
secret
in
that
meeting.
The
only
thing
is
we're
probably
going
to
have
it
at
times
that
are
convenient
for
us,
so
which
means
usually
it's
like
1:00
in
the
afternoon
or
1:30
my
base
and
Thursdays
Pacific
time
yeah
Pacific
time
at
first.
A
A
A
We
we
have
a
query
document:
the
tracks
where
we
left
last
left.
We
first
looked
through
all
the
open
PRS
find
somebody
to
review
it
and
or
take
it
to
the
next
step
or
say
no
great
nicely
and
then
can
make
it
through
all
the
PRS.
And
we
look
through
all
the
issues
and
attempt
to
find
somebody
to
Shepherd
it
through.
So.
A
D
A
D
A
A
C
Lee
yeah
so
from
the
component
standard
working
group,
we
are
working
on
this
feature.
That's
actually
in
patch,
seven,
four
one
one
one
that
implements
a
strict
encoding
feature
flag
and
this
this
strict
encoding
is
going
to
eventually,
you
know,
be
something
that's
natively,
supported
by
the
decoders
that
are
part
of
API
machinery.
C
But
right
now
we
have
needs
to
have
a
constructor
for
strict
decoders
that
had
these
feature
flags
and
and
talking
with
Liggett
and
and
Lucas
and
Luba
mere
last
week
we
decided
that
we
wanted
a
separate
package
that
kind
of
just
lived
in
component
base
to
implement
this
decoder
constructor.
But
if
you
actually
read
the
interface
in
this
patch,
it's
a
pretty
messy.
C
It
duplicates
a
lot
of
code
from
code
of
factoring
and
I'm,
not
really
super
happy
with
where
it's
going
really
like
some
opinions
on
this,
we
wanted
it
to
be
simpler
and
have
less
code,
duplication
and
benefit
from
the
same
test
suite.
It
would
need
to
be
in
codec
factory,
but
since
these
are
features
that
we
want
to
be
part
of
API
machinery
like
by
default
I,
not
really
sure
where
to
take.
This
and
I
figured
it'd,
be
good
to
bring
bring
up
here
with
the
larger
audience.
A
C
B
C
B
Guess
I
would
be
more
interested
in
seeing
a
direction
that
makes
it
possible
to
cleanly
decompose
different
kinds
of
decoders
like
you're,
adding
another
one
and
you're
saying
it's
difficult
to
do
in
general.
I,
don't
see
that
as
a
reason
to
make
existing
package
larger
I
usually
see
that
as
a
reason
to
say,
okay,
this
is
the
third
one,
let's
figure
out
which
abstractions
we
actually
need.
A
A
No,
no,
this
is
just
no.
It's
no
trouble
I.
Just
when
I
get
you
the
result,
you're
hoping
for,
but
I,
don't
think
we
can
do
that
and
in
the
next
five
minutes.
Yeah.
Let's,
let's
discuss
on
the
mailing
list,
maybe
is
there
I
I
know.
I
saw
some
chatter
about
this
already,
but
I'm
not
sure.
If
it's
because
I'm
subscribed
to
too
many
issues
or
it
there
was
actually
a
minimalist
discussion.