►
From YouTube: WG Component Standard 20191008
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
Okay,
welcome
everyone
to
the
tuesday
october
8th
working
group
component
standard
meeting.
We
had
a
couple
things
on
the
agenda
today.
Let
me
share
my
screen
here.
A
So
the
first
thing
that
I
wanted
to
make
sure
we
discuss
is
just
you
know
with
regard
to
our
mentorship
program
kind
of
like,
what's
the
best
way
to
manage,
you
know
getting
face
time
with
mentors
and
kind
of
check-ins
and
support
like
we
have
the
custom
slack
channel,
we've
got
the
email
list.
You
know
people
seem
to
be
using
that
to
ask
questions.
That's
good!
A
I
guess
my
main
question
was
like:
is
this
weekly
meeting
going
to
be
enough
face
time
for
everyone
to
get
the
support
that
they
need
and
if
not,
maybe
we
should
set
up
kind
of
like
an
office
hour
session
later
in
the
week
or
something
like
that.
So
what
do?
What
do
people
think
about
that.
A
You
know:
should
we
just
go
ahead
and
set
it
up
now,
should
we
kind
of
wait
and
see
whether
people
are
able
to
get
enough
out
of
this
meeting.
B
Well,
I'm
not
sure
about
the
new
contributor
since
last
week
I
can
speak
from
my
experience
that
I
needed
like
a
bit
more
time
for
like
with
lee,
to
speak
me
through
some
issues,
just
like
some
cold
war
and
ask
some
questions,
and
I
doubt
that,
like
this
meeting
is
only
like
half
an
hour,
I
don't
think
it's
it's
enough
to
clarify
those
questions
and
I
don't
think
it's
the
right
kind
of
place
for
it
to
really
dive
deep
into
these
topics.
B
But
on
the
other
hand,
I'm
not
sure
what
the
outcome
has
been
since
then.
I
think
seen
a
couple
of
people
ask
questions
in
the
stack
channel,
but
I'm
not
sure
if
there
was
anyone
yet
who
said
okay
can
we
do
like
a
cold
walk
or
something
like
that?
B
But
I
do
really
like
the
idea
of
office
hours
just
for
like
regularity,
and
you
know
people
to
show
up
maybe,
but
of
course
it
would
depend
on
the
availability
of
people
and
stuff.
A
Right
yeah,
we
have
to
figure
all
that
out
all
right.
You
know
I'll
I'll
talk
to
paris
and
see
what
we
can
set
up
and
you
know
we'll
have
to
find
a
time
that
works
for
most
people
and
all
that
as
well.
But
it's
probably
I
think,
it's
probably
better
to
set
it
up.
A
B
B
Yeah,
so
basically
just
a
quick
update,
I
mean
mike
you've
been
commenting
on
that.
Thank
you
very
much.
I
think
the
like
blocking
issue
was
from
jordan
saying
that
we
implement
like
some
lenient
path,
basically,
which
only
exists
for
b1
beta1,
I
think,
and
we
emit
mornings
and
as
soon
as
that
one's
deprecated,
we
can
remove
that
I've
added
that
one.
B
C
B
Like
a
string
gets
returned
with
the
error
inside
of
it,
so
we
can't
really
check
it
in
tests.
I
set
up
errors.wrapper,
which
is
a
third
party
library,
and
then
I
tried
to
do
you
know
the
new
go
113
style
with
the
percentage
w,
which
I
think
is
not
supported
yet
so
my
question
would
be
at
this
point:
do
we
want
to
use
the
arrows
rep
for
time
being,
and
then
you
know
once
we
have
go
113,
we
basically
patch
that
one
again
or
what
would
be
the
best
way
forward.
A
Yeah,
probably
just
use
the
third
party
library,
the
one
thing
that
I
was
a
little
concerned
about,
just
because
it
looks
fragile,
is
here
expecting
to
only
have
to
call
unwrap
once
and
then
check
that
it's
that
kind
of
error.
It
would
be
better
if
you
could
do
a
recursive
unwrap
or
something
to
make
sure
that
it.
B
A
A
B
A
Good,
the
other
thing
I
think
was
this,
so
I
don't
think
we're
blocking
on
having
to
do
a
v1
beta
2.
Now
that
we
have
the
lenient
path
right,
but
I
think
we
should
talk
about
what
that
would
look
like,
because
the
stretch
codec
has
no
like
positive
effect
on
users
using
v1
beta
1
in
when,
once
we
had
the
linear
path
right.
A
B
To
do
it
all
right,
I'll
find
it
in
there
yeah
yeah,
so
basically,
logging
logging
warnings
now
and
it
shows
up
and
everything
but
yeah.
I
don't
know
I
mean
like
what
would
be.
Is
there
another
way
to
handle
this?
I
mean
at
some
point.
We
will
have
to
first
introduce
warnings
and
then
at
some
point
you
know
just
tell
them
no.
A
Well,
yeah,
I
mean
that's,
that's
one
way
to
do
it,
but
that's
kind
of
antithetical.
It's
what
we're
trying
to
do
with
component
config,
which
is
guarantee
the
stability
of
those
apis
like
the
provide
warnings
and
then
eventually
yank.
It
was
the
flags
approach
with,
like
you,
deprecate
a
flag
and
then
yank
it
right
and
we're
trying
to
get
to
a
place
where
those
kinds
of
changes
are
reflected
in
api
versions,
so
that,
if
you
build
on
top
of
an
api
with
a
given
version,
as
long
as
that
version
is
available,
it's
stable.
A
A
However,
it
is,
you
know,
in
principle,
more
correct
to
this
kind
of
changing
the
v1
beta
2.
So
I
think
that's,
okay,
again
we're
not
blocking
this
prnet,
but
we
should
figure
out
what
that
transition
looks
like.
Oh,
there
are
pre-existing
examples
in
lots
of,
like
other
actual,
like
api
server.
Apis
we've
gone
through
that
kind
of
transition,
so
we
might
look
to
that
for
some
instruction
on
how
to
actuate
that
change.
A
We
have
to
go
look
at
some
history
or
talk
to
jordan
about
it.
He's
probably
been
involved
with
more
of
those
transitions.
B
Yeah
all
right,
so
the
cube
scheduler
merged.
So
we
have
the
strict
ones
and
cube
scheduler,
which
I
think
is
nice
and
then
there's
one
for
q
proxy,
which
I
haven't
been
able
to
get
any
traction.
I
it
on
the
cube
network
slack.
If
you
go
to
dpr,
you
can
just
see
the
issue
and
then
you
can
get
to
the
q
proxy
one,
but
to
the
description.
B
So
I
I've
been
messaging
tim
and
caesar.
I
think
his
name
on
dpr
on
slack
they
haven't
gotten
back
to
me,
no
replies
from
the
cube
from
the
sick
networking
channel.
I
tried
to
assign
a
couple
more
approvers
that
was
just
yesterday,
so
basically
yeah
I
haven't
had.
I
haven't
gotten
any
any
feedback
for
the
past
14
days.
So
maybe
I
don't
know
if
you
have
any
idea
how
to
accelerate
this
a
little
bit.
A
Yeah,
I
think
there
is
also
a
question
with
this,
of
whether
we
need
to
do
the
same
medium
path
thing
right
or
if
you
want
altitude
because
even
like
there
was
an
argument
of
even
though
it's
alpha
a
lot
of
people
use
just
built
on
top
of
the
alpha
api,
which
like
puts
us
in
a
frustrating
place,
because
they
shouldn't
have
built
on
the
alpha
api
if
they
didn't
if
they
wanted
a
stable
api.
A
But
we
have
to
respect
the
reality
that
breaking
a
lot
of
people
is
bad.
So
pam.
B
I
mean
it
would
be
nice
to.
Maybe
I
don't
know
if,
like
the
the
meeting
list,
is
a
good
place
for
that,
we
like
started
the
discussions
like
hey
guys.
How
should
we
proceed
with
these
issues?
We're
working
on
those
it's
kind
of
blocking
us
at
the
moment
and
then
maybe
get
some
responses
from
jordan
or
so
and
then,
if
we
have
it,
I
would
talk
to
rosty.
A
As
well
yeah
about
the
proxy
stuff,
because
he
had
that
move
cute
box
to
the
beta
cap
that
involved
the
refactoring
the
modes
stuff,
I
think
he
was
going
to
start
doing
that
in
a
v1
alpha
2..
So
we
might
be
able
to
piggyback
on
that
and
enable
the
script
decoder
and
human
alpha
2.
As.
C
A
Yeah,
I
think
it's
mostly
in
principle
that
behavior
changes
in
a
way
that
the
same
configuration
would
work
in
one
person
and
not
another.
A
A
Right
so
the
the
solution
that
jordan
gave
for
that
was
to
have
a
multi-codec,
basically
that
where
one
is
a
non-strict
codec
that
understands
only
the
prior
version
and
the
other
one
is
a
stretch.
Codec
that
understands
all
versions
and
if
the
strict
codec
fails,
you
fall
back
to
the
non-script.
A
A
I
I
mean
it's:
it's
also
like
people
in
theory
can
have
a
typo
key
like
yeah,
it's
a
mistake,
but
if
everything
in
their
production
is
working
correctly
with
that
type
of
key,
we
do
risk
breaking
them
by
implementing
a
strict
codec.
C
Yeah
and
I'm,
I
can
appreciate
that
we
are
thinking
about
it.
I
think
it's
it's
a
good
thing
to
be
very,
very
considerate
of
users,
especially
when
people
are
using
kubelet
component
config
and
could
proxy
component
config.
C
But
yeah
I
mean
just
like
as
a
I
like
putting
myself
in
the
shoes
of
operating
a
cluster.
If
I
have
keys
in
my
config
that
do
not
have
effect
and
the
behavior
is
that
the
component
doesn't
start
when
I
do
an
upgrade,
and
it
gives
me
an
error
message
that
says
hey.
This
key
is
like.
That
sounds
great.
That
sounds
much
better
than
before.
A
I
think
yeah,
it's
it's
definitely
a
better
behavior.
I
think
we
have
to
separate
the
you
know.
I've
opted
into
the
better
behavior
from
you
know.
I
didn't
change
anything
and
now
I'm
broken
because
you
implemented
a
better
behavior
so
making
making
it
opt-in
by
upgrading
to
you
know
a
new
api
version.
That's
basically
the
same
as
the
existing
one
would
be
a
like.
That's
a
nice
way
to
enable
people
to
make
that
transition
on
their
own
terms.
B
I
I
was
I'm
not
sure
if
I've
done
it
in
the
correct
way
and
it's
it
may
be
a
bit
hacky,
it's
like
when
you
get
configs
loaded
or
when
you
load
the
config
and
you
get
a
strict
error,
then
you
know
decode
it
with.
We
want
better
one
instead
and
give
a
warning.
Basically.
So
just
you
know
just
for
this
version.
It
works
it's
not
like
somewhere
central
with
kodak
factory.
It's
just
like
in
the
logic
where
the
cubelet
config
gets
loaded.
B
I
mean
I
think
it's
like,
because
we've
been
talking
about
this
like
for
a
while.
Now
I
don't
know
if
it,
you
know,
make
sense
to
you
know
bringing
this
discussion
somewhere
publicly
in,
like
you
know
the
mailing
list
and
you
get
some
people
in
who
have
you
know
like
like
jordan
or
like
people
who
work
on
this
api
machinery
and
say:
okay,
this
is
what
we
do
so
we
agree
on
this
and
then,
if
we
have,
you
know
prs
in
the
future,
we
can.
We
can
link
to
this
and
say
look.
B
We
know
we
talked
about
this.
Can
we
do
it
like
that?
I
don't
know
if
it
makes
sense.
You
know
just
to
to
avoid
these
discussions
in
the
future.
When
we,
you
know
open
new
prs
and
then
we'll
be
at
the
same
point
I
was
like:
oh
should
we
really
do
this.
A
I
think,
like
even
the
cubic
pr
right,
where
we've
had
that
discussion
is
sufficient.
That's
an
open
discussion
like
okay,
several
community
members
outside
of
this
working
group
are
involved
in
that.
So
in
that
case
I
don't
know
if
there's
a
reason
to
start
an
equivalent
discussion.
A
B
A
A
So
it's
probably
the
reasonable
thing
to
do
all
right
sounds
good.
B
C
Yeah
just
basically,
I
always
join
that
meeting
and
report
on
progress.
What
people
are
up
to
what
we
need
to
help
on.
A
Sure,
when
is
it
it's.