►
From YouTube: OCI Weekly Discussion - 2021-06-23
Description
Recording of the OCI weekly developer's call from 23 Jun 2021; notes/agenda here: https://hackmd.io/El8Dd2xrTlCaCG59ns5cwg?view#June-23-2021---NO-OFFICIAL-CALL
B
D
E
F
G
C
Okay,
does
anybody
want
to
share
their
screen.
B
A
D
C
I
haven't
touched
windows
in
15
years
and
I'm
using
windows
right
now,
but
I'm
sorry
I
am.
I
am
very
tickled
that
they've
actually
like
truly
I've,
already
been
on
several
like
plugged
right
into
several
different
groups
across
the
company
that
are
they're,
prioritizing
being
able
to
have
like
linux
as
a
primary.
So
even
like
on
our
team.
Everybody
got
got
like
their
laptop
hardware
and
stuff
that
could
run
linux
and
have
good
driver
support
so
because
they're
expecting
even
about
august,
that
folks
will
be
able
to
just
like
reinstall
and
use
linux.
A
H
A
It
down,
but
then
everyone
just
free
formats.
C
But
they
they
they've
actually
got
some
a
lot
of
stuff
locked
down
to
where,
like
you
can't
even
you
can't
check
most
of
the
systems
or
the
sites
or
whatnot
even
like
for
hr
stuff.
You
can't
get
to
it
unless
you're.
C
G
Can
you
all
see
this
thing?
Yes,
yep.
G
All
right,
so
really
quick.
Should
we
close
all
of
chris's
maintainer.
G
Well,
I
think
I
think
what
was
discussed
is
porn.
There
was
a
issue
about
it
and
derek
made
a
pretty
good
comment.
I.
G
A
Yeah,
I
I
think
I
don't
know
I
I
don't
know
if
we
should
close
them
all
out
right
now
or
just
ignore
them.
I
definitely
fall
in
the
same
boat
as
you
or
I'm
like.
I
Those
are
no.
C
Full
requests.
Well
I
mean,
but
that's
that's
where,
like
jason
hall
worked
with
him,
hey
what
I've
worked
with
him,
he's
been
around
for
a
while
last
car's,
obviously
been
around
for
a
while.
It's
one
of
those
like
this
was
more
of
a
conversation
on
the
image
spec
yeah
and
I
think.
A
A
G
F
G
Yeah
I
sent
it
to
you
all
right.
Go
look
at
some
real
stuff,
really
quick
before
we
just
go
into
issues
like
I.
G
It's
no
one
has
pr
this
yet
and
no
one
has
given
certain
opinions
or
talked
about
it
much.
G
Oh,
I
you
may
have
been
missing
the
slack
stuff,
but.
A
Yeah,
I
haven't
been
in
the
slack
and
like
freno
died
and
I
was
in
irc
and
there
was
an
irc
bridge,
but
I
don't
think
someone
set
up
a
new
irc
bridge.
So
I.
D
A
D
A
Six
of
us
in
in
the
libera
channel.
G
Anyway,
yeah
there's
there
isn't
there
isn't
a
pr
for
this?
I
just
what
I
was
trying
to
get
at
is
like
as
we're
doing
this
grooming
like,
should
we
prioritize
based
on
what
we
want
to
get
released.
C
This
whole,
even
sometimes
some
of
these
just
kind
of
sat
around
for
a
little
while
and
just
needed
other
lg
tms
yeah.
I
just
opened
that
one
up,
but
it's
got
like
a
few
comments,
but
did
it?
I
think
they
all
got
it.
A
H
A
G
D
J
G
A
Can
we
run
through
everything
else?
First,
just
because,
like
there
are
certain
like,
I
think
we
can
bang
out
a
bunch
of
these
like,
for
example,
the
oldest
thing.
The
register,
like
the
proxying
yeah,
that
was
discussed
a
lot,
and
I
think
we
came
to
a
consensus
on
it
yeah.
This
is
actually
a
thing
that,
like
derek
and
I
went
back
on
and
we
ultimately
agreed
on.
It
needs
to
be
rebased
and
derek
probably
needs
to
be
here
to
do
anything
on
it,
so
that
can
dangle
the
extension
proposal
thing
that
is
some
contentious.
A
I'm
not
gonna
lie.
I
don't
know
what
the
current
state
of
that
is,
it's
nowhere
near
being
merged,
though,
although
the
way
I
look
at
it,
this
pr
is
probably
the
best
place
just
for
people
to
like
wax
intellectual
on
it.
Yeah.
A
Yeah,
I
think
it's
probably
in
the
critical
path
for
us
before
we
start
adding
new
things
to
the
spec
like
I.
I
think
this
would
actually
be
probably
something
that
we
need
to
agree
on
and
get
merged
for
a
1.1.
C
Would
you
should,
can
we
open
this
up
open
up
a
go,
go
to
the.
C
Yeah,
because
I
don't
I
mean
I
could
open
it
up
also,
this
literally
is
nothing.
A
The
one
right
after
that
is
like
expand
the
allowed
status
codes
is
that
that's
literally
just
changing
the
actual
go
code.
I
think
this
back
right.
We
don't.
C
A
F
The
built-in
github
one
now
okay,
so
so
I
think
I
have
some
settings.
G
Basically,
when
you're
uploading
using
chunks,
there's
apparently
some
ietf
way
to
do
that
and
you
have
to
specify
the
units
of
like.
Are
you
using
bytes
kilobytes
and
it's
every
implementation
has
just
assumed
bytes
and
basically
doesn't
follow
the
this
rfc
and
we
shipped
1.0
without
following
the
rfc,
and
I
I
don't
see
a
reason
to
break
everyone
just
because
of
the
rfc.
D
A
I
G
Actually,
this
pr
was
for
the
conformance
test.
So
can
we
should
we
just
ask
him
hey?
Can
we
change
this?
Can
you
redo
this
in
the
markdown
instead
of
or
should
I
just
say,
close
it
and
if
you
want
redo
it
and
the
test
is
optional,.
A
Do
we
think
this
person
is
alive
still
like
this?
That's.
G
A
Okay
in
this
package,
it
doesn't
look
like
they
really
responded
or
interacted
after
just
the
initial
pr
right,
they
yeah
it
looks
like
they
just
work
on
their
own
registry.
They
found
that
this
was
the
behavior
fired
off
a
pr
to
fix
it.
C
A
Yeah,
the
dot
tool
doesn't
matter.
It
imports
this
library
right
correctly
and
then
it's
only
the
standard
library
that
has
dependencies
on.
So
it's
not
particularly
useful
for
that,
and
then
we
have.
H
We
have
there's
test
pkg.
G
A
I
G
That
this
227
is
like
way
out
of
date.
Hey,
I
don't
know
if
this
was
actually.
D
C
All
right,
yeah,
let's
close
it
like
get
out
of
here,
get
this
corn
out
of
my
face.
B
G
Capturing
sebastian's
table
of.
I
I
A
A
So
so
john's
thing
is
how
he
does
his
convert,
how
gcr
does
their
like
backwards
compatibility
and
I
quay
also
does
similar
but
like
it,
also
supports
v1
and
not
just
schema
one
which
even
further
complicates,
but
I
have
no
idea
well
technically,
technically,
quays
fills
all
the
boxes,
I
think
like
we
check
them
all,
but
it's
all
really
jank
how
we
do
it
if
it
makes
no
sense
right
so.
G
Just
I
don't
I,
what
am
I
going
to
say
basically
like?
I
don't
think
we
want
to
lose
this.
It's
obviously
an
issue
somewhere,
but
the
first
word
in
this
document
is
to
do
so.
I
think
someone
does
need
to
take
a
real
pass
at
this
as
a
real
document.
I
just
it's
not
being
used
by
anything.
It's
not
so
I'm
in
support
of
just
putting
it
in.
C
C
C
J
H
C
Hey
I'm
on
these,
I'm
gonna
share
a
screen
now
boom.
Take
you
out
of
them.
C
Oh
all
right,
let's
take
over
okay,
so
I
was
just
looking
at
you
know
we're
on
the
pool,
approved
discussion
and
got
to
look
at
the
branch
rule
stuffs
here.
So
branch
rule
pattern
or
branch
pattern.
A
And
it's
also
like,
if
you
press
the
button
like,
if
you
do
anything
in
the
github
ui
github
like
signs
it
on
your
behalf,
which
is
also
awful
and
totally
it
ruins
that
workflow
negates
it
yeah
yeah.
So
it's
actually
pretty
bad
because
then
you
get
merge
commits
that
are
like
signed
by
github's
key
on
behalf
of
your
user,
which
is
strange.
This
is
also
a
thing
by
the
way
we're
gonna
have
to
deal
with
signing
in
container
specs.
A
So
that's
why
I
created
multiple
routes
of
trust
when
we
did
which
we'll
call
it.
What's
the
signing
spec,
the
tough,
when
we
did
tough,
we
did
that
so
like
you
could
trust
quay
and
the
person.
At
the
same
time,.
B
A
Sign
commits
is
kind
of
useless.
What
you
really
want
to
do
is
check
the
first
box
and
then
it
will
ask
how
many
yep
dismissed.
C
A
Yep
from
code
and
owners
yeah,
so
you
can
actually
have
a
code
owner's
file.
It's
basically
a
github
enforced
maintainers
file
is
this?
Is
this
like
the
owners?
I
mean
like.
A
It's
it
is
literally
called
code
owners
in
all
caps
and
you
can
either
put
it
at
the
root
or
in
the
dot.
Github
folder
claire
uses
this
and
it
like
auto
tags.
A
It
like
auto,
assigns
you
in
the
git
repo
too,
and
it
was
just
spamming.
The
living
out
of
my
github
notifications,
I
had
to
like
ask
people
to
remove
me
from
a
whole
bunch
of
stuff
because
yeah
they
because
every
time
it
forks
it
would
any
pr
anyone
ever
opens.
A
It
would
just
assign
me
immediately
and
give
me
notifications
on
it
and
I'm
like
if
you
actually
need
me
like
at
me
and
I'll
I'll,
see
the
issue
that
I
need
to
jump
on,
but
instead
I
just
get
flooded
with
updates,
but
because
of
because
of
forks,
not
just
forks,
just
pr's
in
general,
I
mean
I
don't
work
on
claire
anymore,
so
I
was
like,
if
you
need,
if
you
need
me,
I
don't
want
to
see
all
the
notifications,
but
if
you
guys
all
sub
to
it
like
I
have
no,
I
don't
know,
but
it
also
would
require
us
to
change
like
what
that
file
is
called
right,
because
we
have
a
maintainers
file.
C
A
A
I
mean
we
might
not
want
to
check
that
because
we
have
to
get
rid
of
the
pool
we
have
to
get
rid
of
a
bunch
of
yeah
have
to
kill
travis
and
stuff.
G
By
the
way,
I'm
working
on
this
kill
travis,
while
this
is
going
on.
A
A
A
G
Vince
helped:
can
you
look
at
the
travis
yaml
for
a
moment.
C
G
A
What
do
you
want?
What
is
travis
even
running
like
make
install.tools,
I
didn't
even
know
we
have
a
make
file
in
this.
Oh,
like
what
I
didn't
even
know
this
was
here.
What
is
this
even
doing?
It's?
Oh,
oh!
This
is
okay,
so
this
is
generating
the
the
yeah.
We
got
a
mic
file,
no,
it
generates
a
pdf
or
whatever.
This
is
all
your
you're
crazy
doing.
Well,.
C
A
G
There's
a
release
one,
this
validates
that
the
I
mean
this
does
that
and
then
it
also
the
conformance
target
will
make
sure
that
the
conformance
thing
builds.
A
C
Not
I
mean
it
does
it
does
the
same
thing:
pando
generated
the
same
html
also,
but
if
there's
a
pipeline
that
produces
that
and
doesn't,
if
there's
a
pipeline
or
action
whatever
that
produces
the
html
and
would
fail
the
build
if
it's
like
some
bunk
formatting,
then
that
would
be
pretty
decent
I'd
get
rid
of
the
pandock.
Also
but
yeah
I
mean
so
far.
All
the
specs
are
using
the
same
pandock
image
or
should
be
so.
This
is
one
of
those
like
image.
Spec
runtime,
spec
yeah
can.
A
A
I
C
A
C
I
A
C
B
G
G
Yeah,
well
that
didn't
what
was
that
coverage
thing
that
just
happened.
A
A
G
G
A
C
C
A
I
A
A
C
J
B
G
I
C
G
All
right,
let's
jenny
or
mike,
if
you
can,
we
try
and
getting
that
in.
So
I
can
see
if.
J
You're
285
he's
using
go
115,
I
think
we're
consolidating
to
116.
Was
it
seven?
Why
why
why
no,
I'm
just
checking
the
other
pr
against
yours,
no
bueno
115
slide.
C
A
The
file
yeah,
in
fact,
that's
legitimately
how,
when
you
create
a
new
repo
and
like
you,
let
go
automatically
figure
out
your
dependencies.
It
finds
the
minimum
viable
version.
Basically,
so
it's
always
looking
for
the
oldest
thing,
which
is
actually
actually
really
annoying.
Sometimes
so
I
just
opened
a
full
request:
basically,
dropping
pull,
approve,
creating
a
code
owners
and
fixing
my.
G
A
No,
no
but
mine's
not
even
supposed
to
end
in
a
y.
It's
supposed
to
be
z,
I
e
l
and
they
said
like
no,
no,
an
e
after
an
I
before
the
e
and
then
they
did
at
the
end.
That's
I
think,
that's
the
story.
D
C
Hey
your
your.
A
C
If
we
could
just
make
this
a
a
quick
thing
like
that,.
A
F
J
Is
it
broken
in
your
your
private
fork
or
is
it
broken
on,
though,.
G
Yes,
it's
broken
on
my
fork,
but
that's
the
change.
I'm
I'm
running
those
exact
names
locally
and
I
don't
have
the
problem.
G
So
mike
the
reason
didn't
mean
to
cut
your
offense
the
mic.
The
reason
I
don't
know
if
we
care
or
not,
you
can
have
a
badge,
that's
tied
to
a
specific
action
and
if
you
don't
keep
them
separate,
then
if,
like
a
pr
fails,
the
badge
will
turn
red.
If
we
don't
care
about
a
badge,
then.
G
G
I
need
to
run
everybody
if
you
send
me
a
message
for,
like
hey,
put
a
check
on
this
I'll,
put
a
check
on
it.
Okay,
all
right.
I
I
C
C
A
A
A
But
if
you
don't
do
that,
this
is
why
I
even
just
made
that
pr,
I'm
pretty
sure
it's
just
like
anyone
with
admin
on
the
repo
can
approve
them.
So
that
means
you
can
get
some
like
drive
by
like
runtime
spec
or
image
image,
spec
maintainer,
proving
something
yeah.
C
A
Know
I
mean
we
can
do
that
with
the
code
on
our
stuff,
but
I
guess
my
point
before
is
like
someone
on
the
toc
or
someone
with
multiple
repository
access
can
could
accidentally
approve
a
thing.
I've
definitely
seen
like
random
people
just
like
submit
their
approval
like
to
be
like
oh
cool.
It
looks
good
to
me,
even
though
I'm
not
a
maintainer.
I
was
giving
my
feedback
on
this
thing,
thus
I'll
stamp
it
when
I
think
it's
good,
which
I
don't
want
to
discourage.
J
Max
make
some
some
comments
about
possible
new
maintainers.