►
From YouTube: OCI Weekly Discussion - 2022-08-18
B
A
B
A
Turned
on
it's
just
that
I'm
sharing
my
microphone
and
camera
with
the
two
computers.
A
I
bought
this.
This
computer
has
been
with
me
for
a
very,
very
long
time
and
I,
don't
know
what
the
status
of
the
Bluetooth
Hardware
is,
or
even
if
it
has
Bluetooth.
B
B
E
C
Okay,
all
right
so
there's
interest
like
you
know,
I'll,
let
you
all
know
about
like
space
and
that
sort
of
thing
as
soon
as
I
have
it.
But
it's
one
of
those
things
where
you
know:
I
I
recognize
that
planning
for
things
in
August
for
a
October
meeting
is
a
little
challenging,
but
this
is
something
likely
coming
cool.
C
I
can
send
out
something
towards
like
mailing
lists
as
well
to
be
able
to
get
like
a
little
bit
more
I
mean
this
meeting
is
somewhat
lightly
attended.
So.
B
C
Making
it
I
wonder
if
there's
a
way
to
be
able
to
do
this
hybrid
again,
because
that's
that's
pretty
much
like
the
the
way
that
we've
done
this
before.
E
C
E
C
C
I
I'm
I'm
reasonably
certain.
We
can
make
this
one
work.
I
just
want
to
be
able
to
double
check.
That's
all.
E
B
All
right
so
other
than
that
we've
had
a
bunch
of
updates
to
our
PR.
Do
we
sort
of
interest
in
going
through
those
I
know?
We
don't
have
I,
don't
see
Mike
over
here,
so
I,
don't
know
how
much
we're
going
to
get
in
terms
of
feedback
for
some
of
the
stuff
we've
been
changing.
A
I'd
be
happy
to
listen.
If
you
can
explain
the
changes,
because
I
was
on
vacation
last
week
and
didn't
get
a
chance
to
review
it.
E
B
So
yeah
distribution
and
image
spec
got
a
handful
of
things
we
have
updated
here
in
the
past
week
or
so
a
couple
of
course
updates
and
whatnot
I,
don't
know
if
there's
a
good
way
to
say
just
show
everything
that
changed
since
the
last
meeting.
I
guess
I
could
go
in
there
and
find
a
specific
commit
and
do
a
diff.
A
You
could
go
to
the
pr
branch,
the
oci
playground,
PR
Branch.
Well,
you
could
have
clicked
there.
A
A
Yeah
in
the
working
group
proposal-
clicky
thing:
okay,
so.
D
B
B
B
B
B
This
shows
what's
in
the
listing,
and
so
each
one's
got
to
be
a
descriptor
with
an
artifact
type,
where
the
artifact
type
comes
from,
so
I
kind
of
gave
mapping
there.
It
comes
from
the
artifact
type
and
the
artifact
of
manifest
or
the
media
type
in
the
image.
Mana,
the
media
type
of
the
config
descriptor
and
the
image
manifest
so
I'm
kind
of
giving
some
guidance
on
where
we're
pulling
the
stuff
up
from
one
thing,
I'm
catching
on
what
I've
been
doing
here.
This
was
pulled
from
some
other
documentations.
We
moved
stuff
around.
B
A
A
It
to
the
you
could
link
it
to
the
image
spec
artifact.md.
B
Yeah
yeah
and
that's
when
I
did
the
definition
of
them
elsewhere.
That
was
what
I
was
doing
is
I
was
thinking
specifically
over
those
files,
so
I'm
debating
if
I
want
to
put
like
a
load
of
just
kind
of
a
oh.
What's
the
right
word
for
it,
something
up
top
just
say
here
all
the
terms
we're
defining
just
include
that
that
way
not
to
keep
keep
redefining
it
every
single
time
go.
A
A
A
B
B
B
Okay,
if
the
query
results
in
no
results,
then
an
empty
manifest
must
be
your
turn.
So
this
kind
of
goes
back
to
the
people
saying.
Is
it
a
four
or
four?
Is
it
some
of
these
other
options
in
there
that
we've
been
looking
at
so
that
kind
of
covers
a
bit
of
that
get
the
chat
window
open
just
in
case
people
start
talking.
B
B
You
have
to
return
a
manifest
here,
but
the
Manifest
can
have
an
empty
list
in
there
or,
if
you
want
to
you,
can
return
everything.
That's
referring
to
that
ref.
A
A
You
don't
mind:
Brandon
I
have
a
general
question
about
the
feedback.
It
seems
to
me
that
most
of
the
questions
we
get
is
about
like
backwards
compatibility
or
upgrade
path.
Do
you
think
that
we
should
explicitly
point
out
like
make
a
document
for
the
upgrade
path.
B
A
A
lot
of
the
comments
happen
go
ahead,
so
there
was
one
question
about
what
happens
if
the
reference
API
doesn't
exist
and
someone
makes
this
kind
a
client
make.
This
client
makes
this
kind
of
call
and
I
think
the
spec
says
that
you
could
you.
If
if
the
API
doesn't
exist,
then
you
return
a
404,
but
it
may
not
have
been
that
or
maybe
it
was
difficult
to
you
know,
I
mean
you
still
have
to
go
through
the
spec
and
find
that.
B
Yeah
I
think
that's
the
default
everywhere
in
the
registry.
If
you
get
a
request
for
something
you
don't
recognize,
it's
just
a
404
at
that
point,
but
we
can
easily
call
that
out.
In
addition,
just
to
say
hey,
if
you
don't
support
this,
it's
a
404
but
the
the
challenge
I'm
getting
to
for
that
is
that
I
think
we
want
to
say
if
you
are
oci
compliant
with
1.1
of
the
distribution
spec.
A
B
See
the
link
header
I
went
through
a
bit
of
that
and
then
how
that
looks,
I
think
we
just
kind
of
add
a
few
more
words
in
here.
B
So
we
go
down
to
those
endpoints
down
below
we'll
see
an
extra
field
in
there
and
then
how
that
applies.
I
did
change.
I,
don't
think.
We've
got
the
right
people
here
for
this,
but
oh,
we
do
have
Michael
Brown
here
really
AWS
Michael
Brown
I
changed
the
annotation
from
references
to
reverse,
just
because
I
want
to
keep
that
consistent
with
everything
else.
We
were
doing
within
this
API.
B
E
B
B
B
B
B
And
let's
see
one
question
we
got
is
what
happens
if
the
registry
are
we
allowed
to
push
a
refers
field
that
points
to
a
manifest
that
doesn't
exist
in
the
registry?
Yet,
and
so
this
was
in
response
to
that
saying
you
can
push
the
signature
before
you
push
the
image
that
the
signature
is
signing
and
that
should
be
okay
for
us,
and
so
we're
saying
should
on
this
one.
Where
we've
been
saying
on
other
places,
our
registry
can
reject.
B
If
you
don't
have
a
reference
already
pushed
and
that's
for,
like
Blobs
of
you
have
an
index
that
has
manifests,
are
missing
or
if
you
have
a
image
that
has
blobs
are
missing.
The
registry
can
say:
yep,
it's
not
there,
you're
going
to
fail
with
a
plop
on
known
error,
and
here
we're
saying
you
shouldn't
do
that
for
this
one
field.
A
So
you,
okay,
you
shouldn't,
do
it.
Okay,
then.
A
B
B
Yep,
which
that
happens
already,
of
course,
they
all
say,
blob
unknown.
Even
when
they're
talking
about
a
manifest
which
confuses
me.
B
Let's
see
our
notes
now
that
said,
hey
girl,
look
at
the
implementation
file
is
gone
because
that
file
is
no
longer
there,
and
so
now.
Instead,
what
happens
is
when
we
get
to
the
end
of
the
push
section.
We
say
if
you're
pushing
something
that
has
a
reverse
field.
You
need
to
do
these
extra
steps
if
this
API
returns
a
404
and
so
I'll
link
over
to
the
API
we're
talking
about.
If
that
gives
you
a
404,
you
must
do
all
these
steps,
and
so.
A
The
same
thing
from
Mayan
I
think
that
refers
has
to
have
a
small
r.
B
B
And
then
yeah,
this
is
the
whole
backward
compatibility
of
you
pull
the
index.
If
you
don't
have
an
index,
you
make
one
that's
empty,
and
let's
see
here,
you
assume
that
it's
empty.
You
verify.
This
is
already
there.
If
your
description
already
there,
you
can
skip
it.
If
it's
not,
you
have
to
add
it
in
there.
You
have
to
pull
up
the
artifact
type,
you
have
to
pull
up
the
annotations
and
you
have
to
push
a
new
index.
So
we
have
those
steps
all
listed
out.
B
We
see
people
kind
of
coming
off,
mute.
There
feel
free
to
jump
in.
Let's
see
on
the
the
order
over
here.
That
was
the
period
from
before.
On
this
one
I
think
we've
already
looked
through
a
bunch
of
this
scriptures,
artifact
type,
yeah
I,
think
this
was
what
we
had
already
looked
at
the
example.
We
already
talked
about
changing
the
s-bomb,
the
link
we
already
talked
about.
B
B
We've
got
deleting
an
image.
So
if
you
delete
an
image
or
an
artifact,
that's
got
the
reverse
yield.
You
should
also
go
through
and
find
anything.
It's
got
that
that
type
schema
for
it,
assuming
you've
got
a
404
for
that
and
then
go
clean
up
the
entry
from
that
manifest
list.
So
that's
just
kind
of
being
a
good,
a
good
Community
member
of
saying,
okay,
my
manifest
is
gone.
Let
me
go
back
and
get
rid
of
everything
that
was
pointing
to
that
manifest.
B
B
So
registry
only
gives
you
some
of
the
apis
or
that
might
be
an
older
version,
and
here
are
the
here's
fallback
procedure,
and
so
this
language
came
out
of
talking
with
Mike
last
week,
IBM
Mike,
where
he
said
start
with,
must
and
then
we
can
always
fall
back
later
on
and
say:
okay,
we
can
make
it
or
should.
If
someone
wants
to
make
a
client,
that's
only
1.1
compatible
and
doesn't
have
any
one
other
support
at
all.
B
We
can
consider
if
we
need
to
make
that
an
option
to
be
within
the
spec
at
that
point
in
the
future.
For
now
we're
saying
it's
a
must:
they
have
to
comply
with
older
registries,
and
so
when
that
happens,
we
Define
here's
where
I
Define
my
terms
before
so
I
might
just
want
to
take
this
stuff
and
move
it
up
somewhere
put
a
section
up
top
and
then
we've
got
the
tag
schema
here.
That
just
says
here's.
What
the
schema
looks
like
here
is
what
the
result
is
supposed
to
look
like.
It's
got.
B
You
know
the
output's
supposed
to
be
an
index
and
the
maintaining
of
this
index
is
a
client
responsibility,
so
they
need
to
make
sure
they're
doing
that
if
they
delete
an
image
or
something
like
that,
they
need
to
update
it
and
it
should
contain.
You
know
anything:
it's
got
the
reverse
field
and
so
I'll
go
through
and
do
a
global
binary
place
for
that
refers
field.
B
Change
that
lowercase
and
let's
see
I,
do
point
out
that
we've
got
race
conditions
with
this
and
so
protecting
against
that
it's
either
responsibility,
client
to
figure
that
out
and
users
figure
it
out
or
you
can
upgrade
to
a
registry
that
provides
the
reverse
API
and
Magic
happens.
You
have
to
worry
about
that
anymore
and,
in
addition,
I'm
saying,
potentially
this
conditional
HTTP
push
with
the
etag
infrared
free
sports
set
great.
You
can
use
that
as
well.
B
B
I
would
not
be
offended
if
someone
says
delete
that
line
out
of
this
file,
but
I
figure.
We
give
people
a
little
bit
of
an
idea
of
potential
other
ways.
They
can
work
with
this,
and
this
was
that
file.
I
was
looking
for
earlier,
which
was
upgrading
what
do
Registries
and
what
not
need
to
do
so.
I
took
out
the
little
client
implementations
in
there,
but
we
go
through
and
say:
Registries
need
to
convert
over
certain
fields,
and
they
must
include
anything
that
has
one
of
those.
B
B
B
B
I
also
made
the
RFC
a
link
and
put
that
down
below
I
really
want
to.
After
we're
all
done
with
this,
do
a
massive
markdown
reformatting
change
this
to
like
backticks
instead
of
bold
things
and
changes
to
you
know,
standardize
how
we're
making
lists.
But
that
is
something
completely
outside
of
what
we're
doing
here.
B
B
It's
your
problem.
If
someone
else
steps
on
your
same
name,
but
if
you
do
then
great
you've
staked
your
claim
to
it,
but
at
the
very
least
anybody
parsing.
It
should
be
able
to
parse
that
field,
and
it
should
look
within
that
four
method
to
file
within
the
RFC,
with,
like
the
application,
slash
something.
B
B
And,
let's
see
same
for
the
refers
said,
there's
a
scripture
instead
of
a
string
and
then
we
standardize
the
definition
for
refers.
So
we've
got
this
defined
two
different
places.
We
use
copied
and
pasted
the
same
definition
in
both
places.
So
please
got
that
down
to
a
single
thing
and
then
I
linked
over
to
the
refers
API
and
said:
hey
look
over
there
to
see
how
we're
using
this
thing.
B
And
let's
see
following
may
be
used,
I
got
rid
of
all
of
our
annotations
here
and
just
said:
hey
go
see
the
predefined
annotations,
so
we
got
all
the
definitions
in
one
place
and
then
why
would
you
have
an
annotation?
Well
I've
referred
you
over
the
refers
API
say
filtering
may
be
possible
on
the
output
of
that
thing,
because
we're
pulling
them
up
so
client
can
do
some
filtering
on
that
and
then
yeah
like
I,
said
Nisha.
B
B
And
one
of
them
was
an
s-bomb
format,
so
yeah
we've
got
two
different
things
in
here:
two
different
fields
with
the
different
media
types,
but
otherwise
yeah
try
to
go
down
to
a
simple
to
a
single
example
of
keeping
these
things
standard
and
the
same
example
we
copied
to
the
different
files.
So
if
someone
looks
from
one
filing
it
goes
looks
the
other
file,
they
should
be
the
same
content
in
both.
A
E
A
The
wasn't
the
descriptor
so
artifact
type
supposed
to
be
in
the
descriptor,
though.
B
And
let's
see
or
CI
artifact
manifest,
so
this
is
us
saying
yes
in
the
index
definition,
the
index
says:
You
must
support
within
the
descriptor
list.
These
manifest
types,
and
so
we
add
in
there
saying
well
in
an
index,
you
must
support
not
only
the
image
manifest,
but
now
you
must
also
support
the
artifact
manifest
we're
debating
whether
it
should
be
a
May
or
something
like
that
and
figured
well,
let's
throw
it
in
there's
a
must
for
the
1.1
version
of
the
spec.
B
Yeah
we've
got
image
index
with
a
couple
examples
here,
so
I
think
we're
just
adding
more
examples
in
the
different
files.
Just
so
people
can
see
what
they
look
like,
and
this
includes
both
image
index
and
an
artifact
index
in
here.
B
B
I
can
copy
two
lines
around
and
there
is
an
annotation
Square
spec,
which
I'm
discovering
these
files
as
we
go
through
and
do
this
stuff
a
lot
of
files,
I
don't
have
to
work
with,
and
so
we
added
the
definition
for
each
of
these
annotations
in
there
as
well
at
a
standard
fields.
For
that
and
that's
what
happened
this
past
week.
B
B
B
A
B
Cool,
well,
that's
been
what's
happening,
I'll
go
through
and
change
the
the
one
comment
you
had
there.
Nisha
I
feel
like
I've,
seen
that
myself
I'm
just
sort
of
changing
reverse
lowercase
r.
That
makes
a
lot
of
sense
to
me.
D
So
what
what's,
what
do
we?
What
do
we
plan
to
do
like
I've,
been
trying
to
Ping
I?
Think
lucky
has
also
been
trying
to
bring
John
to
get
some
feedback.
D
B
And
as
much
as
I
want
to
say,
you
know
is
keeping
them
keep
pinging
them
at
a
certain
point,
we're
just
going
to
have
to
say:
well,
we
we've
got
the
votes.
We
can
merge
this
if
we
wanted
to
right
and
so
the
the
worst
that
happens.
Is
we
merge
it?
Someone
comes
back
and
says
you
fools.
You
messed
up
all
this
stuff.
You
need
to
fix
it
all.
We
say
great,
send
a
PR,
yeah
and,
and
we
can
always
do
another
PR
down
the
road.
D
We
have
time
to
release
the
Qatar
cutter
tag
or
something
like
that.
So
there's
that's.
B
A
that's
a
great
comment,
because
we
absolutely
do
and
we've
been
getting
the
pressure,
I
think
Sasha
you're
on
that
PR
as
well
or
issue
as
well,
where
they're
saying
hey,
we
want
to
1.0.3
tag
of
the
of
the
image.
Spec
I
believe
that
that
was
what
that
was
for,
because.
D
Data
field
got
merged
and
that
should
have
released
a
gold
spec
file
for
supporting
that
field
in
different
places.
And
it's.
B
So,
for
whatever
reason
they
want
that
one
little
small
change
in
there
I'm
like
okay,
you
know
we
we
can
put
that
in
my
thought
is:
let's
finish
up:
we've
got
our
hands
full
right
now,
getting
this
one
solid
as
best
we
like
it
and
pick
some
old
tweaks,
there
make
sure
we
got
the
lowercase
on
certain
fields
and
all
the
other
little
mirror
details.
We
keep
missing
finish
that
get
that
merged
into
Main
and
then
go
back.
B
B
So
that
gives
us
something
to
do
not
next
week,
because
I
think,
if
we'll
have
our
hands
full,
just
hit
the
merge
button
on
this
next
week,
but
the
week
after
it'll
be
good
to
do
a
little
bit
of
maintenance
on
stuff
like
that
and
then
once
we
get
this
done.
I'll,
probably
like
I,
say
I
want
to
do
some
markdown
maintenance,
just
clean
things
up
make
it
look
pretty
so
I'll
probably
be
going
through
the
spec
submitting
a
bunch
of
little
PR's,
no
substance,
no
substantive
change,
just
making
stuff
prettier
standardized.
D
E
A
B
I
appreciate
all
the
help
we've
been
getting
on
making
this
happen,
because
we've
had
a
lot
of
people
putting
heads
down
this
one.
You
know
Josh,
isn't
here
today
to
get
the
Josh
love
but
yeah
a
whole
lot
of
effort
on
this
thing.
B
Cool,
so
we've
got
stuff
to
do
still
little
minor
tweaks
on
this
thing
overall,
I
think
nobody
on
here
should
be
surprised
by
this,
because
I
think
everybody
hears
that
has
been
trying
to
follow.
It
has
been
probably
following
along
with
individual
PR's
yeah
and
then
next
week.
Hopefully
we
get
this
thing
merged.
A
So
do
any
anyone
who
is
well,
okay,
maybe
that's
a
question
for
another
time,
but
what
about
the
current
status
of
the
working
group?
Is
it.
D
D
We
discussed
this
last
week
where
we
have
to
figure
out
whether
we
archive
the
repo
once
this
is
Mars.
What
the
next
steps
are.
I,
don't
think,
there's
an
answer
yet,
but
potentially
there's
not
much
going
to
happen
in
the
working
group,
because
all
the
discussions
happening
in
the
pr
at
this
point,
yeah.
F
Sounds
like
I
mean
it
sounds
like
unless
someone
is
like
wants
to
use
that
as
a
forum
to
discuss
the
pr
there's,
not
really
a
lot
for
us
to
talk
about,
except
you
know,
advocating
for
the
pr
to
folks
and
clarifying
it
wherever
we
can
yep.
B
I
think
yeah,
unless
there's
anything
that
comes
up
between
now
and
then
we'll
just
officially
say
skip
for
now.
E
B
Last
meeting
yeah
and
then
after
this
Nisha
I,
think
the
decision
was
we're,
keeping
it
open
just
in
case
this
doesn't
get
approved
if
we
got
massive
pushback
and
we
need
to
do
a
complete
refactor
of
everything
and
come
out
of
The
Proposal
G,
but
that's
not
happening
I
think
we're
we're
going
to
get
this
pushed
on
through,
with
only
my
tweaks.
So
far
nobody's
really
complained
yet
yeah.
B
A
Did
we
get
any
feedback
from
the
folks
in
the
APAC
time
zone.
A
B
A
He's
welcome
to
submit
the
change
like
suggested
change,
I
I
think.
Maybe
he
may
feel
that
we
would
not
accept
changes
and
perhaps
we
we
are
to
say
no.
No,
we
actually
do
want
feedback.
D
A
B
There
there
was
a
separate
issue
over
in
the
in
the
working
group.
Oh
okay,.
A
Yeah
I
wasn't
on
the
pr
I,
don't
think
akihiro
commented
on
the
pr
I.
D
B
And
part
of
the
challenge,
this
one
I
think
is
when
we
put
it
over
there,
yeah
Josh
jumped
on
there
and
just
said
closing
his
PRS
are
open,
Upstream,
repos
and
so
we've
closed
the
issue.
This
was
issue
41
over
in
there.
So
I'll
link
it
here
and
in
our
mean
notes
as
well.
If
anybody
wants
to
go
look
at
it,
there's
a
bunch
of
back
and
forth
a
lot
of
people
didn't
like
it.
I'm,
not
a
huge
fan.
The
whole
reason
I
open
it
up
was
just
saying:
hey.
B
F
D
The
other
thing
is
these
fields
are
things
that
nobody
should
technically
touch
in
clis
or
experiences,
but
definitely
when
you
write
code,
you
would
have
to
kind
of
do
some
kind
of
buffering
off
the
fields,
and
that
is
maybe
a
quirky
experience,
but
given
that
it's
a
string,
if
they
I'm
open
to
calling
it
whatever
you
want,
I'm
fine,
because
a
big
chunk
is
just
about
whether
the
experience
and
the
behavior
is
good
right.
So
if
it's
just
a
string
change,
let's
go
for
a
dream.
Just
before
we
cut,
we
just
do
a
blanket
thing.
D
I
think
yeah.
D
B
B
Right
all
right,
I'll,
give
everybody
10
minutes
have
a
great
rest
of
your
week
and
talk
to
you
next
week
when
we
hit
merge
on
this
thing.