►
From YouTube: WG Component Standard Meeting 20190326
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
This
is
a
follow
up
cap
to
our
earlier
case
at
our
component,
based
one
where
we
basically
said
roughly
on
a
high
level
that
these
are
the
things
that
we
consider
a
component
that
that
component
should
have.
But
we
need
something
more
more
detailed
that
like
what
that
actually
outlines
on
a
component
level,
what
needs
to
be
done
other
than
on
a
code
level
that
this
these
are
the
packages
that
should
be
moved
to
this.
These
places.
B
Yeah
the
scope
for
this
document
is
pretty
wide,
so
I
have
the
general
structure
in
place
and
we've
kind
of
started
filling
in
details
for
areas
that
I
could
find
out
information
about.
But
there
are
a
lot
of
opening
questions
that
I
personally
have
obviously
there's
a
lot
of
special
areas
of
interest
here.
The
this,
a
big
one,
would
be
leader
election,
which
seems
heavily
on
respect
at
the
moment,
and
Timothy
also
brought
up
that
it
could
be
pretty
controversial.
B
He
mentioned
something
about
how
the
existing
packages
should
already
be
factored
to
be,
seeking
import
I'm,
not
sure.
If
that's
true,
obviously
we
shouldn't
be
changing
any
implementation
details,
but
if
we
want
to
refactor
that
package
into
somewhere
else
and
then
if
there's
multiple
implementations
put
them
next
to
each
other,
that
would
be
yeah
leader.
C
Elections
currently
fairly
easy
to
import
a
self-contained
manner.
We
do
it
in
AP.
I'm
are
in
controller
run
time,
I
believe
the
implementations
mostly
sit
next
to
each
other
as
well,
but
there's
really
only
I
think
there's
three
right
now:
there's
one
that
uses
the
leader
election
API,
there's
one
that
uses
config
maps
and
I.
Think
technically,
there's
one
that
uses
that
CD.
D
A
E
A
A
B
A
F
B
A
There
were
questions
or
like
comments
from
segaki
texture
that
we
should
do
this
as
also
a
Doc's
first
or
like
proposal
first
approach.
Besides
the
code
first,
one
with
that
is
component
base.
We
should
also
like
document
the
decide
state
over
component,
and
this
is
more
that,
like
what
is
the
plan
for
logging?
What
is
the
plan
for
need
election,
but
it's
a
register
I.
C
I
think
there's
potentially
certain
classes
of
changes
that
might
want
their
own
cap
so
like.
If
we
make
really
big
changes
to
how
logging
works
we
might
we,
we
would
probably
want
to
do
that
in
a
related
cap
as
opposed
to
a
single
cap,
but
I
still
think
that,
like
they
need
to
like,
there
needs
to
be
a
central
place
where
all
of
that
is
mentioned
together.
C
A
And
that's
less
listing
so
like
without,
for
example,
doing
loads
of
changes
to
none
of
these
areas
really
like
not
like
super
large
changes,
but
more
stating
that
this
is
a
thing
and
then
component
base
is
a
thing
and
then
logging
and
you
should
like,
have
open
API,
endpoints
and
whatever
for
your
component
Co.
Second,
you
should
support
components,
config
and
all
those
things
right.
A
From
from
what
I
heard,
as
everything
else
is
like
caps
today,
I
think
or
if
I
remember
correctly,
the
request
was
that
would
also
have
a
cap
for
this
I'm
like
it's
gonna,
it's
gonna,
say
some
minor
things
that
are
changing
the
existing
behavior.
But
yes,
definitely
like
and
I
think
this.
This
is
a
good
place
that
so
now
we
we
start
in
this.
A
We
start
crafting
what
we
think
a
component
should
function
like
and
then,
when
we
submit
this
proposal,
then
we're
gonna,
we're
gonna,
see
that
oh,
we
we
really
fell
into
a
rabbit
hole
with
regards
to
structured
logging.
This
actually
needs
its
own
cap
and
then
we
break
out.
You
know
so
I
think
that
is
a
fairly
productive
way
of
doing
it.
So,
like
we
start
with
saying
on
a
high
level
this
and
that
and
that
that
and
all
this
needs
to
be
there
and
then
we
see
somebody
it
starts.
A
D
B
Thanks
the
leader
election
section,
there's
a
the
major
point,
though,
is
that,
while
there
is
some
detail
already
in
this
camp,
it's
really
lacking
as
far
as
something
that's
supposed
to
be
a
technical
spec
for
where
we're
gonna
end
up.
So,
if
you're
an
SME
in
any
of
these
areas,
if
you
have
a
special
interest,
please
come
in
and
the
the
hack
can
be.
Anybody
can
edit
it
like
just
make
it
good,
because
it's
not
good.
B
D
B
C
It's
nice
to
be
able
to
reduce
your
set
of
dependencies
when
building
third-party
components
and
I
know
that's
only
vaguely
in
spec
right
now
or
kind
of
semi
outers,
but
like
having
to
pull
in
an
entire
API
serving
tree
of
dependencies,
because
you
needed
to
use
some
other
stuff
for
your
controller,
is
kind
of
annoying
and
DEP
is
pretty
good
at
pruning
unused
dependencies.
But
when
you
have
cross
imports
that
gets
really
hard.
So.
A
So
so
what
I?
What
I
think
makes
sense
and
what?
What
if
I
understand
the
idea
of
component
base
or
the
way
I
thought
about
component
base?
Is
that
we'd
win
factor
out
the
the
HTTP
things
that
aren't
there
aren't
related
to
serving
API
groups?
You
know
so
like
so
a
component
like
you
proxy
that
has
no
API
groups,
let
it
serve,
but
still
it
has
an
HTTP
server
for
metrics
and
config
and
whatever
small
health,
so
that
would
be
in
component
base
that
minimal
HTTP
s
serving
code
and
then
API
server
would
have
like.
A
A
A
It's
great
and
we're
about
six
minutes
still.
F
B
A
Haven't
actually
seen
Daniel's
comments
about
K
flag,
but
with
regards
the
component
base
from
what
I
understand
we're
gonna
move
already
existing
packages
just
kind
of
assess.
You
know
that
is
a
plan
and
we
move
the
message
like,
for
example,
leap,
client-side,
leader
election,
sorry,
a
client-side
authorization
move
that
component
base
and
to
know
nothing,
we're
experimental
is
is
mentioned.
Guess
it's
code
that
already
is
it's
used
and
so
on,
but
I.
F
A
We
should
not
talk
about.
The
cold
chain
is
really
detailing.
That
is,
for
the
other,
kept
the
component
base
kept.
That
should
touch
them
more,
like
this
package
there
and
this
package
courseware,
but
this
new
cap
that
we're
talking
about
now
should
more
paint
out.
The
vision
of
this
is
what
a
component
should
look
like
in
three
years.
You
know
thing
like
that,
like
in
three
years,
I
want
the
cube,
let
the
function
like
this
or
they
keep
proxy.
Do
you
agree
with
that?.
F
A
Yes,
so
so
that's
a
follow
up
fully,
so
it's
like
it's
more.
This
is
painting
the
vision,
rather
than
actually
diving
into
the
implementation
details
and
we're
mentioning
that
component.
Config
is
a
big
thing:
leader
election
authorization,
client-side
HTTP
serving
that
and
then
logging
like
nothing.
B
A
A
B
Yeah,
we're
good
merging
Stuart
also
had
the
global
flag
retractor
for
the
booklet,
so
that's
also
been
open
for
forever.
That's
71
for
91
and
then
I
opened
up
these
two
PRS
that
begin
to
break
out
Lucas's
older
PR
that
had
a
bunch
of
work
in
it.
So
that's
that
stuff's
now
building
and
it
has
some
to
do-
is
kind
of
sprinkled
you're
out
there.
I
haven't,
but.
A
A
So
yeah
Stuart's
PR
and
let's
at
least,
let's
get
that
in
that
is
fairly
simple
and
and
config
serializer
to
start
with
and
I
can
review
your
or
your
two
new
breakouts
from
my
PR,
but
that
is
gonna.
Take
a
bit
longer
yeah
the
two
follow-up
items
for
this
or
three
follow-up
items
this
week
is
take
a
look
at
the
cap
and
then
7411
one
and
seventy
one.
Forty
nine
one
four
should
ideally
be
much
for
all
for
next
week,
so
we
can
keep
moving
now.
That's
called
traces,
as
listed
doesn't
make
sense
to
everybody.