►
From YouTube: CentOS Automotive SIG - April 5, 2023
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
Right,
great
yeah
that
that
was
I
think
what
I
was
trying
to
get
to
is
how
many
external
contributors
do
we
have
in
autosd
in
the
automotive
Sig
and
then
for
the
participants
in
the
automotive
Sig.
If
we
are
using
something
from
Upstream
whatever
it
is,
it's
OS
build
or
sentos
or
whatever.
If
we
want
to
change
something
for
our
benefit,
we're
doing
that
in
the
upstream
and
then
it
comes
down
yeah.
Yes,.
B
B
A
Right
great
and
we're
trying
to
do
we're
trying
to
mimic
that
with
docs,
so
we're
trying
to
do
access
code.
We
do
the
same
thing
so
so
I
guess
then
the
the
next
question
is
then,
when
things
occur
in
the
Upstream
and
Leonardo
kind
of
gave
me
the
impression
that
it
kind
of
sounds
like
it's
a
monitoring
manual
sort
of
thing.
A
C
If
you're
attacked
like
a
packaging
updating
from
Centos
upstring
and
that's
something
that
we
use
the
same
build
system,
so
we
get
chose
updates
for
free
in
this
case,
because
it's
all-
and
it's
got
to
be
the
same,
build
system
for
us.
But
let's
say:
will
you
get
to
us
build
updates
on
the
same
way,
because
it's
built
on
our
build
system
as
well?
So
when
you
you
just
we
just
have
to,
we
don't
need
to
worry
about
osbuilder
updates.
C
If
we
do
add
updating
in
the
in
our
developer
machine,
we
get
those
updates
automatically,
but
but
with
if
we
so
we
get
updates
for
free.
C
B
It
may
get
a
little
bit
more
complicated.
Also
Sophie
is
now
has
a
published
specification,
and
so
up
until
now,
we
have
not
really
been
complying
with
a
spec,
because
there
hasn't
been
a
spec
to
comply
with.
We've
been
really
following
what
they've
been
doing
with
their
reference
implementation,
ewall
and.
B
B
I
know
Eclipse
also
has
aspirations
towards
something
along
those
lines
as
well
they're
at
least
go.
They
will
be
doing
a
reference
implementation.
It's
called
it's
called
Lida.
The
I,
don't
know
whether
they're
going
to
be
focusing
on
compliance
tools
necessarily
because
there
is
there
probably
won't
be
a
written
spec
in
the
same
way
that
there
is
with
Sophie.
D
C
B
C
But
then,
now
that
they
have
a
specification,
you
kind
of
have
a
reference
to
what
the
image
should
from
a
document
standpoint
and
then
they're
already
working
on
tools
to
test
some
of
their
stack,
so
the
spec
more
tools
to
they're
also
open
to
contributions
from
other
groups.
Just
like
ourselves.
It
shows
tools,
but
that's
very
similar
version
with
us
built.
We
would
contribute
upstream
and
then
we
get
kills
back
manually
later.
There's
no
automated
process
for
it.
B
B
B
Just
an
agenda,
and
maybe
one
technical
status
slide
today,
we
we
actually
just
talked
all
about
documentation
resources
as
far
as
Sophie
images,
I
did
want
to
bring
up
the
the
Sophie
specification,
which
is
here.
B
This
in
the
little
chat
here
this
is
what
the
specification
looks
like.
You
can
see
that
it's
got
the
architecture
overview
and
a
whole
bunch
of
placeholders
there's
a
lot
of
challenges
ahead
that
are
not
not
yet
completed,
but
you
can
see
that
there's
also
a
section
called
devops
in
the
cloud,
so
so
I
think
that
that's
going
to
be
it's
going
to
be
a
very
interesting
ride
with
Sophie.
One
of
the
difficulties
of
developing
a
specification
is
that
you
can
create
all
the
code
you
want
in
your
reference
implementation,
but
that's
specific.
B
That's
just
one
thing
that
you're
building
and
a
specification
needs
to
be
more
General.
You
need
to
allow
it
needs
to
allow
the
flexibility
of
other
implementations
that
also
comply.
For
example,
Sophie's
reference
implementation
is
Source
based
it.
You
build
everything
from
the
ground
up
using
yaco
project,
and
you
know
that's.
That
is
the
traditional
way
that
embedded
systems
have
been
built
for
decades
and
the
it's
not
the
way
that
we
build,
that
we
build
Auto
SD
at
all
or
any
of
the
red
hat
stuff.
B
So
making
our
stuff
comply
with
their
stuff.
Is
our
stuff
comply
with
this
specification
is
is
sort
of
a
challenge.
It's
also.
That
means
it's
incumbent
on
the
folks
writing
the
specification
to
write
it
in
such
a
way
that
it
can
accommodate.
You
know,
ideas
other
than
the
ones
that
they
come
up
with
so
anyway,
so
that
is
available,
see
a
lot
of
folks
joining
I
I
really
want
to
apologize
again
for
the
mix-up
with
the
the
calendar
entries,
the
I
know
the
calendar
entries
were
they
were.
B
B
E
E
B
E
Might
be
what's
going
on
for
other.
Folks,
too,
is
the
original
wasn't
wasn't
replaced,
and
so
you
know
I
just
naturally
joined
when
I
saw
the
the
reminder
pop
up
so
that
that
might
be
what
everyone
needs
to
do
is
just
delete
the
one
that
starts
at
the
top
of
the
hour
and
keep
the
one
that
half
past
the
hour.
B
I
will
go
back
and
okay
and
Steve
only
sees
the
one
for
the
top
of
the
hour.
So
I
will
go
back
and
delete
everything
and
and
reshoot
it
I'm
not
exactly
sure
what's
going
on,
but
you
know
what,
if
calendar
entries
were
easy,
then
we
probably
wouldn't
have
jobs.
E
B
That
was
the
intent.
If
people
don't
think
that's
going
to
leave
enough
time,
you
know
I
I
can
probably
still
maintain
it
at
the
earlier
hour.
I
just
I
have
a
conflict
now
that
normally
pops
up
at
7am
on
Wednesdays,
so
at
least
at
7am
my
time
so
but
I
will
yeah.
Let's,
let's
it
might
actually
be
easier
to
just
go
back
and
do
it
at
continued
at
seven,
so
I
may
have
to
go
and
move
my
conflict.
B
Into
the
automotive
calendar
within
Red
Hat,
you
mean
yeah.
There
actually
is
a
copy
of
this
within
the
red
hat
calendar
and
I.
Remember
changing
that
one
as
well,
but
the.
B
All
right
well,
apparently,
my
my
calendar
Foo
has
failed
me
dramatically
today.
Thundering
is
hard,
it
shouldn't
be,
but
it
is,
it
shouldn't
be
hard
in
the
21st
century.
That's
for
sure.
B
So
Pierre
and
Leo
we
have
room
for
a
technical
update.
I,
don't
think
that
the
slide
has
been.
We
don't
have
anything
on
the
slide,
but
if
there's
anything
that
you'd
like
to
walk
us
through
with
recent
changes
to
Centos
or
to
autosdean,
we
would
love
to
hear.
C
Sure
so
I
have
some
updates
with
Sophie
and
integration.
Library,
so
I've
been
I've
been
working
testing
out
Pasty
in
southeast
integration
lab
with
their
CI
system.
That's
called
Leva,
so
lava
works
on
a
different
way.
You
can
actually
attach
hardware
to
it.
Wherever
that
Hardware
is
so
I've
been
using
a
remove.
C
Just
to
press
the
drop
definition
itself,
it's
work
the
job.
The
job
itself
is
working
at
this
point,
so
we
are
now
working
on
using
a
Ava
platform
with
some
a
machine
from
that
arm
will
make
available
to
us
to
to
run
to
run
tests
which
using
Arm
System
ready
and
it's
what
been
using
selfie
or
you
all,
building
and
testing
the
machine
is
being
shipped
to
the
US.
C
At
this
point
it
may
take
a
at
least
two
weeks
to
get
everything
sorted
out,
but
what
base
battery
plan
also
have
all
set
in
a
data
center
is
to
hook
that
machine
up
to
to
the
integration
lab
and
run
tests.
Then
we
can
run
automated.
Ci
tests
beat
nightly
every
pull
request
for
our
software
image
and
so
on,
and
at
the
same
time,
that
chose
test
will
be
using
scripts
and
tools
that
are
provided
by
linaro
itself
and
Sophie.
C
So
that
means
we
will
be
going
to
test
our
images
using
Upstream
Sophie
tools
which
aligns
with
the
goal
of
making
our
selfie
autosd
software
image
and
solely
compliant
according
to
their
specification
and
referencing
limitation
I
just
I'm
also,
we
are
I'm,
also
packaging
lever
itself
in
Fedora
they
have
they
have
so
labor
is
basically
they
have
a
so
the
way
that
you
hook
up
your
Hardware
with
their
status
to
have
to
use
what
they
call
a
labor
dispatcher
with
biscuit
demon
that
you
read
on
your
on
your
Builder
machine
and
that's
when
available
on
Debian,
so
I'm
packaging.
C
My
last
update
for
stuff
is
that
the
mouse
work
in
the
blue
on
a
softly
blueprint
to
use
a
kubernetes
files
to
the
file
application
workloads.
So
that's
basically
getting
what
we
do
with
quadlet,
but
abstracting
into
a
blueprint
and
just
amplifier
how
we,
how
you
can
use
the
same
workload
to
test
your
application
in
the
cloud
run
it
on
ewall
using
k3s
and
running
in
alt
SD
using
quadrilateral
Heritage.
C
So
that
means
it's
a
blueprint
about
user
kubernetes,
yaml
files
to
deploy
application,
different
environments
run
times,
implementations,
Etc,
and
that's
also,
that
will
also
be
used.
Integration
web
under
testing
tools.
So
it's
a
it's
a
bit.
It's
kind
of
a
big
blueprint
and
but
most
multiples
are
interested
and
I
know
that
I
know
that
Leonardo
is
is
has
started
to
use
a
bottoman
and
quadlet
or
some
of
their
tests,
because
it's
way
more
lightweight
than
k3s.
F
C
Update
for
Eclipse
with
and
basically
I
started,
eclipses
try
to
build
a
an
end-to-end
story,
using
cooks
and
at
The
Chariot,
quite
early
stage
test,
and
so
on.
So
I'm
take
a
look
at
those
troop
projects.
What
what
it
takes
to
to
build
test
and
run
them
in
in
our
alt
SD
images.
So
this
most
likely
they
chose
projects
of
your
our
next
targets
to
enable
in
our
out
as
the
eclipse
as
the
ve
image.
B
That's
that's
fantastic.
You
know,
working
with
external
communities
like
this
is
one
of
the
Prime
goals
of
of
this
sig
that
we've
set
out
over
a
year
ago.
B
F
F
I
actually
have
one
thing
that
I
realized
I
should
mention:
iftario
is
now
available
on
Old
Fedora
branches.
So
if
you're
running
federa,
36
37
the
future
38
or
even.
C
D
Yeah
for
Leo,
do
you
have
a
list
of
differences
for
the
pipeline?
That's
building
all
the
images
and
the
the
differences
for
building
the
Sophie
image,
and
is
any
of
that
stuff
that
is
expected
to
be
adopted
by
the
broader
product
pipeline
for
creating
all
the
images
or
is
it?
Are
we
focusing
it
on
just
those
Sophie
images.
C
So
far
the
only
change
is:
is
the
application
stack?
That's
been
there
with
software.
There's
not
much
I,
don't
think,
there's
any
peculiarity
that
would
make
a
difference
for
the
sample
images.
C
Most
if
there
is
if
there
is
a
requirement,
for
example,
like
how
the
kernel
should
or
boot
speed
or
anything
like
that,
that's
required
by
itself,
it's
most
likely
to
be
changed
in
in
the
Upstream
Auto
kernel
anyway.
So
I,
don't
think
there'll
be
any
yeah
differences
from
this.
On
that
level.
From
the
example
image
that
we
currently
have.
C
No
okay,
that's
the
point
of
the
for
Sophie
to
have
testing
tools,
so
they
chose
tools
we'll
test
the
resulted
image
for
evil
or
yakto,
and
the
oldest
D
doesn't
matter
how
it
was
built.
It
matters
as
the
end
result
and
those
tools
will
measure
and
assert
if
it
is
compliant
with
the
specification.
Okay,
great.
D
B
A
good
way
to
to
picture
that
also
is
to
realize
that
Sophie
is
not
going
to
be
solely
open
source.
The
so
the
so
Sophie
itself
is
a
specification
is
open
source,
but
it
would
be
very
easy
for
something
like
qnx
to
also
comply
with
the
Sophie
specification.
B
B
F
B
One
other
thing
about
Eclipse
sdv,
also,
while
we're
still
talking
about
external
communities,
is
that
they
did
just
have
a
community
day.
It
was
actually
two
days
full
of
of
updates
and
information
about
the
the
projects
within
Eclipse,
SDC
and
I
believe
that
those
are
set
to
be
published
between
I
I,
believe
they're
supposed
to
be
up
by
Friday.
B
So
keep
an
eye
on
the
eclipse,
sdv
website
and
I'm
sure
that
that
will
be.
That
will
be
updated
to
show.
B
B
E
B
B
It
is
a
working
progress
yeah
it
is,
and
it
is
a
it
is
a
complicated
process
inside
red
hat
to
to
build
out
this
stuff
I.
C
C
B
One
thing
that
I
could
I
suppose
I
could
throw
out
there
to
folks
like
John,
who
I
know
is
from
TI,
and
anybody
else
who
has
an
interest
in
in
supporting
Auto
SD
on
their
Hardware
is
that
we
could
talk
about
that.
We
can
talk
about
that
in
the
documentation
for
for
autosd,
there's
nothing
Stopping
Us.
B
So
John
I,
don't
know
how
do
you
are
you
having
any
issues
getting
Auto
SD
working
on
the
the
platform
that
you
were
talking
about.
E
Yeah,
so
I
have
a
lot
of
comments
and
and
I
want
to
be
careful
to
separate
the
ones
that
are
specific
to
Auto
SD
in
this
group
and
the
ones
that
are
just
Fedora
in
general.
Okay,
and
just
give
you
a
understanding,
so
I
I'm
attempting
to
get
the
auto
SD
to
work
on
one
of
our
TI
platforms,
and
you
know,
Step
One
is
I,
built
May
made
sure
I
knew
how
to
build
it
correctly
for
Raspberry
Pi
and
then
I.
E
You
know,
went
to
build
it
for
our
platform
and
and
running
into
problems
in
boot,
and
so
I
want
to
debug
and
investigate
and
and
immediately
be
wanting
to
swap
out
debug
versions
of
the
you
boot
or
kernel
or
whatever,
and
all
that
ability
to
to
change.
E
And
that,
honestly,
for
me,
has
been
like
a
completely
new
world
and
a
very
overwhelming
and
in
some
places
great
documentation
in
some
places,
I've
learned
that
you
got
to
go
look
at
the
code
if
you
really
want
to
understand
how
things
work,
and
so
it's
been
a
and
I
come
from
the
yakto
world
or
even
in
yakuto
I'm,
not
an
expert,
but
that's
the
background
I
have,
and
so
it's
been
a
it
ensure
has
been
a
very
steep
learning
curve
is
what
I'm
trying
to
say
now.
E
You
know
I'm
sure
people,
other
people
are
going
to
want
to
Port
their
devices
and
probably
going
to
run
into
the
same
questions.
The
same
learning
curve
that
I'm
going
through
right
now.
B
I
think
it's
a
great
great
summary
and
you're
definitely
pointing
out.
You
know
the
it
is
a.
It
is
an
extremely
complex
thing
and
I'm
actually
relatively
new
to
the
red
hat
stuff
too
I,
don't
know
if
everybody
on
the
on
the
call
knows,
but
I
I
help,
launch
and
run
the
Octo
project
for
a
number
of
years,
and
so
I
come
from
an
embedded
background
with
you
know
where,
where,
when
you
build
something
you
build
it
from
Source.
E
B
I
found
it
easiest
to
understand
by
understanding
that
this
whole
system
that
that
red
hat
has
put
together
over
the
last
30
years
is
really
been
focused
on
the
data
center
and
had
the
needs
of
the
data
center
using
devops
tools,
as
opposed
to
the
embedded
World,
which
is
building
one
tiny
little
thing
and
then
replicating
it
many
times
so
so.
B
Yeah
and
I
think
that
it,
it
would
be
great
for
us
to
to
develop
some
some
documentation
specific
to
Auto
SD,
that
maybe
kind
of
covers
that
I
see
Laura
on
the
call.
If
we
can
talk
about
different
ways
that
we
could
approach
that
and
maybe
put
together
a
tutorial
that
can
walk,
walk
through
what
it's
like
to
Port
from
one
to
one
thing
to
another.
E
Yeah
and
I'm
sure
I,
just
think
of
some
common
scenarios,
like
I
I'm
sure
that
people
who
are
developing
run
into
a
niche
and
go
oh
I,
need
to
modify
I
want
to
swap
out
the
kernel
of
one
I,
just
rebuilt.
It
has
some
instrumentation
or
some
print
app
statements,
whatever
add
it
in
there
and
you
want
to
easily
swap
those
out
into
the
image
to
build
and
and
build
it
and
run
and
debug.
E
F
F
Based
on
the
discussion
we
had
I
had
in
my
to-do
notebook
to
write
something
for
you
and
I
have
something
it's
it's
a
page
long
at
the
moment.
C
B
F
And
once
we
have
something
we
I
may
actually
ask
you
to
pull
in
your
notes,
and
so
it's
just.
C
B
G
Yeah
I
know
I've
always
found
it's
actually
a
little
easier.
If,
for
me,
everyone's
different,
but
I've
always
found
it's
a
little
easier
when
doing
arm
enablement.
If
you
enabling
Fedora
first,
because
obviously,
we've
been
Upstream
first
Focus,
but
that's
not
the
only
reason.
I
I
actually
find
it
easier
because
you
did
the
recent
kernels
there's,
there's
more
tools
available.
G
E
Least
yes,
so
I
I
did
actually
take
that
approach
and
and
I
do
have
the
door
working
well
on
this
platform.
I
did
that
first,
but
now,
when
I
build
with
the
OS
build
flow,
there's
something
different
and
I
haven't
figured
it
out
yet.
F
G
Yeah,
you
have
to
do
it
for
Fedora,
and
otherwise
do
you
have
to
do
it
for
both
sometimes
I
just
find
there's
less
hurdles
in
fedora,
because
the
software
is
more
up
to
date
and
actually
there's
a
larger
selection
from
of
tools.
People,
people
there's
there's
about
I've,
seen
people
build
Fedora
with
about
three
or
four
different
tools.
G
E
Yeah,
yeah,
and
and
actually
to
be
more
precise
in
my
statement,
you
know
you
say
this:
what
I
did
on
Fedora
actually
is
I
took
I
can
use
the
pre-built
images
for
arm
platforms,
particularly
targeted
for
the
Raspberry
Pi
and
I
can
get
them
to
work
well
on
the
ti
platform,
I'm,
not
rebuilding
the
Fedora
image
and-
and
that
probably
would
be
a
good
exercise
to
to
do
because
I
suspect
it's
probably
some
issue
in
the
way
I
am
building
it
for
the
autosd
that
that's
creating
this
problem.
That
I'm
saying.
F
It
may
actually
not
be
that
that
useful,
because
I
expect
I'm
I
expect
hedora
doesn't
choose
the
same.
Doesn't
choose
USB
to
build
its
images
by
the
public
one.
The
one
you've
tested
with
isn't
built
with
this
build,
but
I
believe
there
is
a
support
for
federal
images
in
osbuilder,
but
I
don't
know
that
they
would
give
you
the
so
image
Builder
sorry
image.
F
Builder
is
the
layer
on
top
of
os
build
that
provides
the
nice
user
interface
and
everything,
but
I
I,
don't
know
if
revealing
the
federal
image
the
federal
iot
image
would
actually
get
new
match.
That's
what.
B
B
B
You
know
rmg5
rv7
are
completely
different
architectures,
so
so
yeah,
it's
a
it,
will
be
good
for
us
to
to
kind
of
get
this
stuff
down
on
paper
or
in
in
words.
I
guess
is
a
better
way
to
put
it.
B
Fantastic
all
right:
well,
we
have
actually
covered
a
lot
of
the
stuff
in
the
agenda.
Is
there.
F
B
I
did
have
extended
use
cases
down
there,
I'm,
not
sure
I,
don't
actually
have
a
whole
lot
baked
to
announce
or
discuss
I,
don't
have
anything
to
announce
actually,
but
I
do
want
to
just
mention
that
it
has
been
really
interesting
that
we
have
had
some
queries
about
Auto
SD,
because
it
is
essentially
a
streamlined
boiled
down
version
of
of
Centos
stream,
with
a
very
specific
kernel
that
that
particular
use
case
is
very
interesting
to
other
things.
Besides
Automotive,
it's
you
know
it.
B
It's
not
it's
not
what
I
would
call
small
enough
to
fit
into
an
iot
use
case,
but
it
certainly
is
small
enough
to
fit
into
an
edge
use
case
where,
if
you're,
building
an
edge
device
that
is
going
to
have
some
intelligence
out
at
at
the
edge
that's
going
to
communicate
back
to
a
a
cloud-based
resource
that
you
can
split
the
workload
between
those
things
and
it's
actually
much
easier
to
do
that
with
a
small
streamlined
operating
system.
That
only
has
the
basics
that.
B
B
So
so
we're
we're
proceeding
and
you
know
certainly
Auto
SD-
is
available
and
can
be
used
for
whatever
anybody
wants
it
to
and
I
don't
know
if
any
of
the
people
who
contacted
me
over
the
last
couple
of
weeks
are
on
the
call.
But
if
you
are,
you
know,
let's
talk
about
your
use
case
we'd
be
happy
to
pull
it
out
and
and
disassemble
it
and
see.
B
If
there's
anything,
we
can
do
in
Auto
SD
to
help
enable
it
and
otherwise
I
I
think
that
we've
pretty
much
got
our
agenda
for
the
month.
B
I
will
fix
the
calendar
and
if,
if
anybody,
who
has
anything
that
they
would
like
to
add
to
it
or
if
you'd
like
to
bring
something
up
to
the
meeting
to
discuss
just
send
me
a
note
drop
me.
A
note
in
chat
I'd
be
happy
to
put
it
on
the
agenda
for
next
month
and
with
that,
I
will
give
everybody
back
two
minutes
before
their
next
meeting.