►
From YouTube: Package Community Office Hours: Debian MVC
Description
The Package Group meets with Mathieu who has been working on a contribution to add support for Debian to the GitLab Package Registry.
A
Hi
everyone
I
scheduled
a
synchronous
meeting
to
meet
with
matthew
who's,
been
working
on
a
new
community
contribution
for
adding
support
for
dbm
packages
to
the
gitlab
package
registry,
and
I
feel,
like
the
team,
has
all
the
people
in
this
room
have
been,
or
this
virtual
room
have
been
working
together
for
the
past
several
months.
So
I
thought
it
would
be
nice
to
kind
of
have
a
chance
for
everyone
to
get
together
and
ask
any
questions
synchronously
that
they
have
and
also
start
to
get
the
community
excited
about.
A
This
feature
that's
going
to
be
coming
out
soon,
so
we
have
an
agenda
that
we
looks
like.
I
had
the
first
item,
I'm
interested
in
knowing
where
we
are
today-
and
I
guess
how
close
we
are
to
this
feature
being
available
on
on
gitlab.com.
For
for
our.
A
B
Before
with
I,
I
think
that
there
is
the
issue.
B
B
It's
already,
and
there
is
a
draft,
mr
with
all,
the
all,
the
all
those
mr
applied,
and
it's
working
at
least
it
was
working
a
previous
week,
but
now
because
of
the
review,
things
are
have
moved
a
bit,
but
it's
easy
to
to
fix
this.
A
Well,
yeah,
this
this
view
is
really
helpful.
It's
amazing
how
much
progress
we've
made
so
far.
Are
you
blocked
by
anything
or
or
having
trouble?
How
is
everything
going.
B
I
think
I
I
don't
have
anything
blocking
me,
but
the
the
review
level
is
very
high
in
gitlab.
It's
not
the
case
if
in
other
projects,
so
there
are
like
one
two
three
people
reviewing.
I
think
after
this,
because
it's
really
good
it
slows
down
the.
A
Okay,
well,
that
was
just
my
high
level
question
steve.
Do
you
want
to
jump
to
your
your
question
about
group
distributions.
C
Sure
can
I
react
on
two
things:
oh
yeah,
of
course,
so
yeah
the
review
time
is
slow,
but
that's
for
everyone,
even
for
us.
So
that's
that's
expected
and
two.
I
heard
this
line
of
preparing
the
code
to
much
complex
situations.
C
Usually
how
we
do
things
at
gitlab
is
really
we
go
for
the
simple
version
first,
so
it
can
be,
it's
not
dangerous,
but
you
could
invest
a
lot
of
time
in
trying
to
solve
a
future
problem.
That
might
not
happen,
and
usually
what
we
do
is
that
we
just
focus
on
on
the
problem
at
hand
and
just
iterate
on
it
in
if
we
have
a
more
complex
thing
to
handle.
Well,
we
just
iterate
on
it,
so
I
would
say
that
it's
it's
a
balance
about
yeah.
C
I
know
that
this
is
something
we
will
need,
so
I
can
start
working
on
it.
That's
that's
correct,
but
I
guess
it's
a
warning:
don't
fall
in
the
trap
of
yeah.
I
need
to
handle
this
and
this
and
we
will
need
this
and
actually,
after
releasing
the
feature,
this
abcd
problem
do
not
happen
at
all,
and
actually
we
didn't
need
to
handle
them
at
all
and
those
situations
already
occurred.
Sometimes
so
that's
why
I'm
I'm
putting
this
warning
of
not
going
too
far
in
solving
future
problems,
because
it's
it's
not
worth
it.
C
C
We
need
the
group
level
okay,
so
we
will
need
to
implement
the
group
level
and
then
you
wait
for
us
people
needing
the
instance
level,
and
then
you
implement
the
instance
level.
But
this
is
a
really
long
time
frame
and
in
that
time
frame
we
could
have
other
features.
So
other
features
such
as
the
virtual
registries
that
we
have
in
mind
that
could
replace
a
complete
whole
set
of
features
which
in
this
case
would
be
the
group
and
instance
level.
They
could
be
a
good
replacement
for
that.
C
A
A
A
B
I'm
sorry
my
computer
should
not.
C
C
E
Cool,
well,
I
wanted
to
say:
hey,
I
mean
we've
been
chatting
for
a
while
on
reviews,
so
it's
good
to
finally
meet
you
and
I
just
wanted
to
say
thank
you
again
for
all
of
your
expertise
that
you've
brought.
I
mean
I've.
I
come
from
the
world
of
npm
and
rubygems,
where
debbie
or
deviant
seems
very
different.
E
So
it's
been
really
nice
to
have
your
expertise
and
and
help
and
understanding
as
we've
been
working
through
these.
I
had
a
couple
questions
that
I
had
thought
of
one.
It
was
actually
something
that
I
was
trying
to
recall
for
for
david
as
well
was
the
group
distributions
are
not
associated
to
packages
in
any
way
right
now,
and
I
don't
recall
why
that
is.
Was
there
a
reason
or
is
that
planned.
B
B
When
you
create
a
distribution,
you
you
have
a
required
parameter,
which
is
a
codename
and
it
will
match
all
the
project
distribution
with
the
same
code
name
within
the
group
and
it
it
will
aggregate
it.
It
will
merge
all
those
into
a
new
distribution.
E
Okay
yeah,
I
remember
now
yeah
that
I
forgot
the
code.
Name
is
what
ties
everything
together.
Thank
you
for
reminding
me.
Yes,.
B
There
is
also
a
shoot
with
a
sweet
which
is
like
an
alias
for
the
code
name
and
when
you
the
this
is
the
case
you
in
debian,
for
example,
when
you
upload
a
package
you
upload
to
unstable
with
which
is
an
alias
for
seed,
and
you
can
also
less
often
upload
to
stable,
which
is
an
alias
for
currently
buster
and
the
the
idea
is
once
the
package
is
uploaded
in
the
given
distribution.
B
It's
it's
linked,
link
it
to
the
distribution,
even
if
the
shoot
change
the
the
name
changes.
I
don't
know
if
it's
clear.
B
B
Testing
will
map
to
bookworm,
which
is
the
next
next
release
you,
you
can
see
shoot
as
moving
aliases.
They
can
change
during
time.
B
E
Cool,
thank
you
for
the
for
those
details.
E
The
other
curiosity
I
had
just
not
being
super
familiar
with
debian
is
I
see
that
you
know
in
in
the
draft
of
the
docs
and
through
working
through
these.
Mrs,
the
the
distributions
are
all
handled
through
an
api
where
we
make
curl
requests.
How
do
other
registries
work
with
distributions?
Are
those
also
just
usually
using
simple
tools
like
curl,
or
do
they
ever
have
specialized
client
tools
or
cli
tools.
B
B
E
B
B
So,
usually,
everything's
only
two,
yes,
the
this
is
it's
usually
done
by
placing
the
files
and
launching
the
command
which
generates
the
release
files.
So
it's
it's
it's
not
the
same
as
npm
or
compiler
or
thing
like
this.
It's.
B
Yeah
it's!
It
came
from
all
time,
so
it's
in
debian
the
generation
is
done
four
times
a
day
which
is
very
slow.
When
I
upload
package
to
debian
I
I
may
need
to
wait
six
hours
before
it
enters
seed
before
it
is
processed.
So
I
don't
want
to
create
the
same
behavior
in
the
github.
I
want
something
very
instantaneous,
but
the
problem
is
the
way
those
files
are
created.
B
You
create
a
packages
file
for
each
component
and
each
architecture,
so
there
can
be
dozens
of
them
and
then
all
those
files,
those
files
for
depth
files,
udab
files,
source
files.
There
is
a.
B
An
entry
point
file,
which
is
a
release
file
which
points
to
all
those
files
with
the
checksums
and
this
file
is
signed
at
the
end.
Ensuring
with
the
with
this,
the
signature
and
all
the
checksums.
That.
B
The
the
word
thing
is
signed,
transitively
sign
this.
This
is
done
from
packages
to
release
to
package
file
to
a
release
file,
but
when
the
client
requests
the
data,
it
goes
the
other
way,
so
it
first
asks
for
the
release
files,
but
this
file
can't
be
known
until
all
the
other
files
are
already
processed.
B
The
yeah,
the
the
check
sum
of
of
the
entry
file,
is
not
bound
to
all
all
the
all
the
packages
inside
the
repo.
So
yes
it's
so
I
I
at
first
I
thought
I
I
would
keep
the
component
regulate
the
component
files.
B
The
package
packages
only
when
a
package
for
the
corresponding
component
and
architecture
would
change,
but
I
think
it's
it's
complicated
to
to
to
manage
events
and
to
to
know
when
when
each
file
is
outdated,
so
I
change
I
change
it
between
the
start
and
the
current
implementation.
I
change
it
things
a
bit
so
now,
when
one
package
is
uploaded
or
removed
or
something
the
world
distribution
is,
is
marked,
outdated
and
should
be
completely
regenerated.
E
Okay,
yeah
and
that's,
I
believe,
that's
one
of
the
mrs
that
we're
looking
at
or
recently
looked
at
the
regeneration.
That's
interesting,
yeah!
It's
it's
very
interesting
how
that
works
and
how
it's
I've
made
it
dynamic.
Thank
you
for
explaining
do
anyone
else.
Have
any
questions
follow
up
with
that?
E
E
Merge
requests
anymore
and
I
believe
david
will
be
probably
the
primary
maintainer
moving
forward
unless
something
else
changes,
but
just
so
you're
aware,
I
think
that's
kind
of
all
that
I
was
curious
about
and,
of
course
just
to
to
note.
You
know
if,
if
there's
anything
else
that
I
can
do
to
help
with
the
mrs
and
reviews,
because
I
know
sometimes
I'm
the
slowdown
point
of
there's
a
few
days
where
I
get
busy
and
can't
can't
get
back
to
them.
E
D
Of
course,
called
in
the
second,
I
was
taking
a
sip.
Yes,
we
are
actually
at
time,
though
so
I'll
just
verbalize
it
for
the
sake.
What
do
you
envision
will
be
valuable
to
see
in
the
ui
on
the
front
end,
both
in
the
list
and
the
details
page
which
data
is
do
you
think,
is
the
most
useful
when
you
are
looking
at
a
debian
package
to
see
and
if
there
is
any
special
action
that
will
need
to
support.
D
B
I
I've
already
implemented
a
few
ui
stuff
in
the
in
the
working
progress,
mr,
but
it's
only
simple
thing
because
I
don't
know
view
ugs,
so
it's
only
adding
some
fields
but
yeah.
I
think
we
need
a
distribution
view
because
now
it
only
in
the
package
view.
B
B
D
Okay,
thank
you.
I
think
alphabetized
team
wrote
that
we
should
have
a
follow-up
with
ian.
That
is
our
team
designer.
I
think
the
sooner
we
involve
him
and
this
whenever
we
open
a
conversation
around
him
around
this,
the
better.
It
is
more
things
we
can
discover.
B
B
B
C
A
A
A
A
I
heard
from
a
customer
recently
that
they
that's
that
that's
not
the
preferred
way
and
they
still
want
to
use
killer
and
older
versions
of
helm.
So
I
think
there's
some
open
questions
on
how
we
pursue
pursue
that,
but
generally,
that
is
something
that
we'd
like
to
add,
and
we
are
thinking
about
probably
the
the
latter
half
of
this
year.
B
A
Cool
well,
this
was
awesome
and
if
I'm
happy
to
do
this
again,
if
need
be,
but
yeah,
let's
stay
in
touch,
and
let
me
let
us
know
if
you
need
anything
and
it's
been.
It's
been
awesome
working
with
you
so
far,
and
thank
you
for
your
time.