►
From YouTube: OCI Weekly Discussion - 2023-06-22
C
B
E
E
Sure
so
I'll
try
to
share.
E
Right
so
first
one
was:
let's:
let's
see
if
we
can
do
the
quicker
one.
B
E
It's
basically
just
removing
this
confusion
and
it
kind
of
like
stemmed
from
Andrew's
last
item
in
the
Milestone.
So
that's
like
duplicated
description
and
artifact
created
that
stuff
changed
so
just
to
keep
it
consistent.
These
two
annotations
will
get
removed,
at
least
from
the
growth
specs.
Okay,
thanks.
E
All
right,
this
is
for
Andrew
so
before,
and
we
want
to
describe
this
or.
D
We
have
a
lot
of.
We
have
a
lot
of
annotations,
some
that
go
online
for
artifacts
some
go
for
traditionally
manifests.
Is
there
a
need
to
align
on
what
those
look
like
doing
a
duplicates?
That's
really
what
you
know.
This
is
looking
at
this
seven
month
old,
so
I
know
a
lot
of
changes
have
occurred
in
the
spec.
Does
this
still
make
sense?
I
know
we
want
to
move
towards
conclusion
for
this
version
of
oci.
How
can
we
go
ahead
and
get
agreement
on
what
the
next
step
should
be.
A
I,
don't
think
there's
any
value
in
duplicating
them,
like
I
think
it
actually
confuses
people
to
be
like
should
I
use,
artifact
description
or
image
description.
The
term
image
is
like
unfortunate
and
a
misnomer,
but
that's
if,
if
that's
what
we
already
have
them
all
named
as
we
should
stick
with
them,
is
what
I
think.
B
D
A
Yeah,
if
you
could,
if
you
could
send
a
PR
to
add
whatever
wording,
you
think
would
be
helpful.
I,
don't
think
it
should
block
1.1
because
we
don't
want
to
add
any
more
things
to
1.1,
but
that
wording
would
definitely
be
helpful
if,
if
you
want
to
take.
E
B
G
Yeah
I
so
yesterday
sat
down
with
Jason
and
John
for
a
while
and
worked
at
these
milestones
and
found
that
we're
actually
closer
than
we
thought
and
were
able
to
close
things
out.
Merge
things
remove
things
that
were
outdated
due
to
changes
with
related
to
artifacts,
manifests
and
all
that.
So
it's
it's
really
like
down
to
these
two
open
issues
and
is
that
who's
still
sharing
Sanjay?
Do
you
want
to
pull
them
up
or
I
can
sure
sure.
C
G
G
So
this
is
distribution
stack,
one
is
related
to
conformance,
and
this
is
in
a
PR.
This
is
that
there's
something
missing
from
conformance,
so
Brandon
I
would
like
to
sit
down
with
you
and
go
over
conformance
in
the
next
week.
If
you
have
time
unless
this
is
already
done,.
G
Okay
and
then
we
also
just
merged
ROM
head
changes
to
add
support
for
references
to
the
conformance
test.
I
think
it's
disabled
by
defaults.
So
maybe
we
can
look
at
that
together
too,
but
that
we
put
that
in
this
was
the
only
one
that
seemed
like
required
discussion
or
if
we
want
to
punt
on
this.
B
G
B
Yeah,
the
key
that
I
was
trying
to
clarify
in
this
for
anybody.
It
doesn't
have
the
backstory
on
it.
Is
that
there's
a
lot
of
places
where
we
say
I,
think
registry
everywhere
and
typically,
when
it's
implemented
it's
by
repository
for
a
lot
of
the
access
request
querying
any
kind
of
digest.
Something
like
that.
It's
at
the
repository
level
Even
in
our
API
calls
that's
how
we
do
it
so
yeah.
G
G
G
This
is
pretty
like
locky's
issue.
Here
is
just
a
little
bit
meta
of
the
Milestone
itself.
I,
don't
know
I,
guess
it
can
stay
there.
It
doesn't
really
it'll
be
closed
once
we're
closed,
and
then
this
is
the
one
we
just
talked
about.
So
I
don't
know
if
we
want
to
drop
this
or
wait
for
Andrew
to
put
a
PR
here,
but
if,
if
we
want
to
drop
it,
we
could
probably
do
an
RC
of
image
spec
today.
A
Yeah
I
think
we
dropped.
We
dropped
the
we
merged
the
pr
to
remove
the
annotations
that
mentioned
artifacts.
I.
Think
the
only
thing
remaining
in
this
issue
is
clarification
of
the
process
and
I
think
that's
outside
of
1.1.
Like
that's,
that's
again
like,
like
the
other
thing,
it's
not
something
that
has
now
regressed
in
this
release.
It's
just
clarification
we'd
like
to
make
in
general,
so
I
would
like
to
move
that
we
remove
this
from
the
Milestone
and
go
go.
C
G
Never
had
it
so
we're
14
minutes
in
I
would
propose
that
we
somebody
opens
that
RC
image
spec
s
out
the
email,
and
we
call
it
early
today
unless
there's
more
to
talk
about
I,
would
also
be
happy
to
you
to
live
code
conformance
then
the
second
half
of
the
call,
if
you
want
to
do
that
Brandon
and
do
a
distribution
today,.
B
G
Moving
past
I
think
we've
I
think
we've
waited
long
enough
and
I.
Think
most
of
the
big
stuff
is
agreed
on
here.
So.
B
B
G
I
would
like
to
put
out
RC's
again
for
the
fourth
time
for
the
fifth
time,
and
then,
if
that
RC
works
with
your
testing,
then
we
can
roll
with
it
as
the
release.
But
I
don't
know.
H
I
would
piss
on
that
Josh
because
we're
we're
at
ECR
we're
iterating
for
HRC,
so
it
would
be.
It
would
be
nice
for
us
to
be
able
to
build
against
an
RC
and
then
test
otherwise
we'd
be
waiting
to.
So
we
would.
We
would
delay
our
own
testing
by,
however
long
it
took
to
cut
an
RC,
so
I
would
prefer
that
if
that's
all
right
with
you
Brandon
and
others.
B
Yeah,
so
even
short
of
doing
implementation,
I
just
want
to
spend
some
time
to
actually
review
the
spec
with
a
little
bit
of
a
little
bit
more
time,
just
sitting
down
and
actually
reading
through
it
making
sure
we
didn't
miss
anything
so
by
the
weekend.
I'll
be
ready
with
that.
E
So
given,
given
that
we
don't
have
any
more
issues
or
any
outstanding
item
to
solve,
can
we
at
least
call
that
this
is
the
candidate
for
RC?
Stop
the
email
process,
get
the
words
going
and
you
can
probably
even
take
I
think
we'll
at
least
have
one
week
before
we
even
cut
the
RC
I
just
start
get
the
process
rolling.
Is
that
okay
or.
B
E
Because
I
think
like
what
just
just
mentioned
is
like
getting
the
RC
out
will
help
some
other
implementations
to
take
a
take,
the
new
fields
and
whatnot
right,
that's
in
the
image
pack
which
they
cannot
take
right
now,
so
that
would
definitely
help
unblock
you.
G
Well,
you're
the
one
who
needs
extra
time
to
review
this.
No.
G
For
your
own
for
your
own
for
us
to
know,
that's
that's
waiting
for.
B
A
G
B
No
it's!
It
is
not
that
I
think
that
we
haven't
had
time
to
have
input
on
an
issue,
it's
that
when
we
merge
a
bunch
of
things
real,
quick,
the
last
second,
it
opens
up
a
lot
of
possibility
of.
We
didn't
realize
that
we
added
one
phrase
here:
delete
one
phrase
there,
and
now
these
things
don't
necessarily
read
correctly
anymore.
Yeah.
A
Another
solution
to
that
is
to
slowly
merge
them
over
the
last
six
years,
instead
of
instead
of
merging
them
in
a
fit
of
peak
or
kick
them
out
of
the
Milestone
right
like
they
don't
need
to
be
changed.
We
don't
need
to
make
changes,
but
I
can't
help,
but
laugh
at
moving
too
fast.
E
G
I
know
how
to
nudge
people
with
the
power.
I
No
one
has
anything
to
say:
I
might
mention
something
that
I
discovered
today,
which
I
mean
I'm
totally
newbie
here,
hi
everyone,
but
I
was
just
looking
at
the
I
was
just
pausing
some
references
and
just
noticed
that
the
the
regular
expression
for
the
repository
names
and
the
distribution
spec
doesn't
seem
to
match
the
practice.
The
actual
you
know
this.
I
What's
going
on
in
the
docker
part
reference
parsing
thing
so
specifically
about
separators,
you
know
the
the
original
allows
a
dot,
but
but
the
docker
thing
allows
like
double
underscores
or
any
number
of
hyphens
and
I
just
wondered
what
the
policy
is
regarding
that
kind
of
discrepancy
which
which
one
kind
of
wins,
because
I
don't
know
whether
I
should
raise
an
issue
or
a
PR,
or
you
know
that
kind
of
thing.
I
I
was
specifically
talking
repository
I'm
talking
the
the
name,
regular
expression.
I
I
G
You,
if
you
want
to
take
a
stab
at
that
regular
expression,
I
think
we
would
accept
that
if
that's
what
this
reality
I.
I
I
That's
right
so
so
it's
basically
you
know
an
underscore.
You
could
have
a
double
end
score,
but
you
can't
have
a
separated
directly
from
the
bar
separator,
so
so,
but
but
it
does
seem
to
be
the
the
comment
in
the
code,
for
that
is
that,
in
order
to
support
valid
host
names
as
name
components,
supporting
repeated
Dash
was
added,
additional
double
underscore
is
now
allowed
as
a
separator
to
loosen
the
restriction
for
previously
supported
names.
I
I,
don't
really
get
what
that's
talking
about,
but
that's
what
the
comment
says.
F
B
I
I
I
just
know
that
I
had
looked
to
the
distribution
spec
in
sort
of
my
implementation,
I'd
I'd
I
was
I
was
you
know,
checking
everything
fitted
within
the
the
other,
the
other.
You
know
the
existing,
stricter
regular
expression.
Then
I
realized
that
if
people
copy
something
from
a
dock
of
England,
it
might
not
be
possible
now.
I
guess.
G
I'm
not
going
to
pretend
I'm
good
at
regular
Expressions.
So
if
somebody
wants
to
Port
that
into
the
spec.
B
Since
you're
here,
Josh
I
know
there
was
a
comment
somewhere
about
us
importing
a
bunch
of
stuff.
Is
it
image
go
that
we
did
this
in
yeah,
we've
got
a
bunch
of
image,
spec
copied
into
distribution,
spec,
slash
conformance
and
a
question
that
had
been
raised
casually
in
some
other
conversations.
Is
why
don't?
We
just
include
image
spec
as
an
external
package.
Instead
of
copying
the
code
over
as
a
profile.
A
So
this
was
on
purpose,
especially
during
the
time.
I
may
have
even
been
the
one
that
copied
it
during
the
time
that
we
were
debating
the
behavior
of
conformance
under
the
in
the
face
of
new
fields,
we
wanted
to
have
the
ability
for
conformance
tests
to
inject
a
new
field,
to
show
like
that,
the
conformance
tests
you
know,
support
or
don't
support
an
image
manifest
that
has
some
extra
thing.
A
I
remember
having
a
week-long
discussion
about
whether
distribution,
specs
should
officially
actually
depend
on
real
image,
spec
versus
other
stuff
I
think
we
should
I,
think
I,
argued
and
will
continue
to
argue
that
they
should
be
separate.
We
should
also
like
make
it
an
internal
package
in
distribution
spec,
just
so
that
somebody
doesn't
accidentally
depend
on
the
distribution,
spec
copy
of
image,
spec.
A
B
A
Think
I
think
a
concrete
action
item
is.
We
should
move
it
to
internal
or
document
that
this
is
not
the
image
spec
you're.
Looking
for
but
yeah.
C
B
The
content
length
distribution
spec
for
for
I
could
put
this
up
on
the
screen.
That's
easier
for
people
wrong
thing
to
move
to
other
screen.
C
B
B
A
B
That's
My,
Hope
and
I
just
don't
know
what
the
right
thing
to
do
in
this
case
is
and
I
don't
know
from
here
or
which
you
know,
I
need
to
go,
get
the
pointer
to
the
other
spec
and
figure
it
out,
but
everything
else
in
here
we
have
to
find
that
you
need
to
have
a
range.
The
range
needs
to
have
this
syntax.
A
Yeah
I
think
I
would
I
would
bias
towards
removing
more
of
that
and
making
more
of
it
a
pointer
to
authoritative,
HTTP
specification,
which
is
a
better
spec
for
everyone,
including
us
who
has
less
to
maintain.
But
that's
probably
a
bigger
change.
I
would
say,
let's
not
add
anything
to
it.
A
Let's
point
to
http,
even
in
in
this
issue
in
this
PR,
if
we
don't
even
like
change
the
spec,
but
in
general
HTTP,
questions
should
be
solved
by
the
HTTP
spec.
G
Documentation
I
think
we
should
accept
both
and
try
parsing
it
Docker
way,
first
or
something.
B
Yeah
I
just
don't
know
what
the
docker
weight
is
for
this
field.
I
mean
we
said
the
content
length
could
be
excluded,
but
we
didn't
say
what
the
content
range
should
be
and
so
without
knowing
that
second
half
of
that
at
all
I,
don't
know
what
we
want
you
here,
so
my
reason
bring
it
up
was
to
see.
Maybe
someone
just
knew
off
top
of
the
head
or
when
you
have
this,
you
just
don't
have
a
range
or
something
but
I
feel
like.
You
still
need
a
range.
G
G
All
right
I
will:
let's
try
to
meet
on
Monday
or
Tuesday.
If
you
Kim.
B
See
ya
Chick-fil-A
all
right,
that's
it!
For
the
week
everybody
rolling
out
thanks
so
much
for
everybody
coming
out
and
yeah.
Let's
hopefully
get
this
RC
pushed
not
too
long
for
now.
A
Going
through
the
exercise
with
Josh
yesterday,
there
were
issues
from
like
three
years
ago
that
if
someone
still
cared,
but
they
would
still
respond
to
this
issue
or
something
in
general
I
know
the
idea
of
bug
bankruptcy
is
traversal.
Maybe.
C
B
When
I
looked
at
information
on
my
own
personal
repos,
the
thing
that
I
specifically
went
to
was
not
automatically
closing
everything,
but
I
would
put
a
tag
or
sorry
a
label
and
by
adding
the
label
onto
there.
That
was
my
indication
that
I'm
waiting
for
that
other
person
to
say
something:
Jason
are
you
with
us.
I
know
your
audio
went
funny
going
in
there.
B
A
A
Old
ones,
it
was
like
you
know
the
the
originator
of
the
issue
and
two
maintainers
think
it's
a
good
idea,
we're
just
waiting
for
any
two
more
of
the
six
remaining
ghosts
on
the
oci
maintainer
list
to
click
a
button
and
if
things
sit
in
that
state
for
a
while,
it
should
not
be
the
originator
of
the
issue's
fault.
It
should
not
be
the
two
maintainers
that
were
doing
their
job
as
well.
It
should
be
like
a
pin
on
the
ghosts.
A
I,
don't
know,
I,
don't
have
a
solution
for
it.
I'm
trying
to
solve
human
problems
with
Technical
Solutions,
but.
A
Yeah,
even
even
just
even
if
it
doesn't
keep
things
moving
along,
but
just
clears
up
the
backlog
of
old
issues.
There
were,
you
know,
discussion
topics
from
before
my
time
that
still
seemed
like
open
discussions,
but
clearly
no
one
is
engaged
in
this
topic.
So
it
should
be
clear
that
this
is
not
an
open
topic.
B
Yeah
we
I
don't
know
if
I
have
any
access,
I
think
that
would
go
up
to
LF
to
do
their
little
flipping
up
or
something
like
that,
but
we
need
to
tell
them
what
we
decided.
We
want
to
do.
B
The
other
one
that
was
on
the
agenda
that
I
completely
ignore,
since
we
were
saying
goodbye
so
quick
earlier,
runtime
spec
is
going
to
be
pushing
for
a
1.1
release
soon.
So
just
as
a
heads
up
for
folks
that
that's
going
to
be
coming
in
the
not
too
distant
future
here.
B
B
I
know
we
raced
through
this
one
quick
and
everything
happened
quick
this
week,
so
any
lastings
before
we
wrap
it
up.
E
B
E
With
you,
after
the
call
right,
if
you
have
time
this
was
to
just
to
confirm
the
the
are,
we
still
continuing
to
toggle
the
version,
or
are
we
just
going
to
do
a
single
comment
to
it?
Rc3
plus
Dev,
on
a
similar
for
the
specs.
B
E
B
But
yeah
the
plus
Dev
I,
think,
is
what
we
wanted.
Instead
of
the
Maya
Steph
and
yeah.
C
B
Right
with
that,
yeah
feel
free
ping
me
if
you
have
any
challenges
and
have
a
great
weekend
folks
and
the
rest
of
the
week,
take
care.