►
From YouTube: Weekly Package Meeting 07-22-19
Description
We review the upcoming milestone 12.3
A
All
right,
happy
Monday,
it's
nice
to
see
all
your
faces.
Hope
everyone
has
a
nice
weekend
looks
like
there's
only
one
and
a
half
items
on
the
agenda
today,
so
we
could
start
with
the
the
first
one,
which
is
the
twelve
point
three
milestone.
So
key
gate
for
product
and
engineering
is
the
22nd
which
is
today
I'm
supposed
to
have
the
milestone,
the
ideally
everything
in
there
that
we
want
to
work
on
and
then
by
the
first
for
engineering.
A
We
want
to
make
sure
that
we
mark
any
items
that
are
accepted
as
with
deliverable
and
then
that'll
make
its
way
through
the
the
proper
path
so
they'll
make
it
into
a
kick-off
document
and
they'll
make
it
onto
the
kickoff
call
but
yeah
the
first
thing
being
starting
today
and
then
by
the
first,
having
everything
accepted
and
in
the
deliverable
tags.
So
I
have
the
link
there
and
I
won't
share
my
screen,
but
I
could
just
go
over
each
item.
That's
in
here
I'm.
A
A
This
one
is
goes
really
well
with
what
next
working
on
right
now
with
the
multi-select,
but
basically
the
dr.
API
is
limited
in
that
you
can't
select
a
single
tag
for
deletion
with
the
API,
so
a
bunch
of
users,
and
the
issue
suggested
some
workarounds,
where
we
create
a
fake
tag
and
then
mu
and
then
and
then
delete
that
one
sounds
like
this
is
in
progress
and
and
GG's
pretty
far
along
based
on
the
last
stand-up
report
that
I've
seen
from
him.
But
any
any
questions
on
on
this
one.
B
So
he
was
last
up
at
a
quick
chat
with
Gigi.
Yesterday
he
mentioned
that
he
was
sort
of
trying
out
some
ideas
that
he
felt
like
would
be,
would
work
and
he
created
an
mi
yesterday,
just
cuz.
He
who
said
he
has
sort
of
prototypes
that
he
was
playing
with
I,
had
a
quick
chat
with
him
and
we
tech,
we
think
you
and
that
I
think
Steve
just
to
see.
So,
if
you
follow
up
with
him
and
in
the
slack
chat
that
would
be
awesome,
you
probably
heard
it
already
I
just
missed
it.
B
A
Something
done
on
that.
I'm
really
excited
about
that
one.
That
was
one
I
thought
we
were
gonna
have
to
wait
a
long
time
to
get
to
it's
nice
that
we
were
able
to
just
work
around
it
and
and
I
think
it's
gonna
be
really
valuable,
and
especially
with
that
multi
select
delete',
that's
going
to
really
transform
things.
It's
great!
When
you
have
actual
people,
isn't
it
yeah?
A
It's
awesome
the
next
one
I'm
unsure
about
so
we
and
I'm
open
to
conversation
on
those
one
so
that
we
have
the
consumption
that
basically,
the
data
tracking
for
package.
I
worked
through
this
milestone
to
find
out
that
we
are
currently
tracking
page
views
for
gitlab
calm,
but
not
for
self-serve.
So
we
are
tracking
some
pageviews
that
we
can
add
to
a
dashboard
and
I'm
gonna
I'm
gonna
work
on
that
and
look
in
this
week,
probably
just
trying
to
get
so.
A
We
get
some
view
into
our
stage
what's
going
on,
but
there
are
a
bunch
of
other
metrics
that
are
not
paid
use
that
we
want
to
track
things
like
the
number
of
build
or
push
events
that
happen
with
the
container
registry
or
the
number
of
each
time
and
account
of
the
images
deleted
from
the
UI
they're
put
in
priority
order
in
the
issue.
I
bet
for
this
upcoming
milestone.
If
we
work
through
the
p2s
that
would
be
awesome,
I
just
don't
have
a
whole
lot
of
insight
into
how
the
data
tracking
works.
A
What
needs
to
be
done
and
what
the
intricacies
are
so
I
feel
like
there's
a
bit
of
Investigation.
That
probably
needs
to
happen,
but
I
think
it's
important
because,
like
right
now
for
this
milestone,
I'm
trying
to
prioritize
the
Maven
feature
over
an
NPM
feature
and
I,
don't
really
know.
I
suspect
NPM
is
more
popular,
but
I.
Don't
know
that,
for
sure,
so
would
be
good
if
we
could
start
start
to
think
about
this.
B
A
C
I
have,
in
fact
so
the
way
we
did
this
on
release.
We
basically
took
took
the
metrics
that
we
that
we
wanted
to
gather
and
then
try
to
figure
out
how
we
would
get
that
from
the
from
the
app
code.
So
it
was,
you
know
what
kind
of
active
record
query.
Can
we
run
to
to
get
this
data
and,
like
is
that
query
gonna
cause
any
problems?
Is
it?
Is
it
a
huge
selector?
Is
it?
C
Is
it
something
that's
indexed
or
whatever,
and
so
we
I
think
we
picked
five
metrics
or
whatever
it
was,
and
then
we
figured
out
what
those
queries
were
going
to
be
and
we
figured
out
that
a
couple
of
them.
We
couldn't
get
given
the
current
database
architecture,
and
so
we
just
punted
on
those
and
then
then
we
then
well
I've
got
an
M
R
somewhere.
Let
me
see
if
I
can
find
it.
That
shows
how
to
like
wire
this
into
the
usage
pane,
which
is
just
what
I
ended
up
doing
with
those
it's.
C
The
usage
thing
is
actually
pretty
easy
to
work
with,
but
I
think
the
harder
part
is
figuring
out
how
to
get
get
those
those
metrics
that
you
want,
but
then
once
you
have
that
you
can
just
add
it
to
the
usage
thing,
there's
no
approval
or
anything
that
has
to
happen.
You
just
add
the
add
the
item
in
and
it
gets.
B
B
Where
that
we're
in
the
application
code,
we
need
to
add
the
line
for
a
specific
metric
is
kind
of
like
it's
about
understanding
the
code
base
and
I'm,
not
sure
we
understand
the
code
base
at
that
level,
because
we're
all
quite
new
moments,
I,
don't
know
whether
the
I
feel
like
once.
We
see
your
template,
which
again
I've
looked
at
already
I
think
it
was
a
while
ago,
so
I'm
not
saying
I.
Could
they
understand
it
or
anything,
but
it'll
be
that
feeling
of
like
okay,
but
which
metric
is
important?
B
C
Yeah
yeah,
so
so,
and
that's
gonna.
That's
gonna
require
some
input
from
engineers
that
have
probably
worked
on
this
stuff
because
there
isn't
like
a
clear
understanding
of
the
data
model,
at
least
like
if
I
go
and
look
at
something.
I
can
look
at
it
and
say
that
that
seems
to
make
sense
if
I
look
in
the
schema.
But
then
someone
with
some
experience
is
gonna,
say
yeah,
actually
that
that's
almost
right
except
you
need
to
filter
on
this
other.
C
B
C
B
That
what
maybe
make
sense,
I'm
Steve,
given
that
you're
the
person
with
the
hands
in
the
back
end
of
the
code,
the
most
welcome,
maybe
what
makes
sense
is
actually
looking
down
that
list
of
items
for
this
for
a
second
just
ignoring
the
priority
just
for
a
second
and
then
have
a
look
down
and
see.
If
there
any
items
you
feel
like
you
actually
kind
of
go.
B
Oh
I,
I
think
I
know
I've
touched
that
codebase
that
relates
to
that,
and
then
we
can
figure
out
a
way
forward,
just
even
if
it's
not
the
top
rate,
at
least
we're
sort
of
moving
the
bar
and
understanding
how
that
process
works,
and
then,
in
the
meantime,
we
can
also
tie
that
up
against
the
priority
items
as
well
and
say:
okay.
Well,
let's
look
at
this.
A
E
B
Also
a
point
of
reference,
but
again
it's
that
knowledge
of
the
codebase
I'm
aware.
Okay,
so
we
need
a
metric
that
tells
us
how
many
times
a
package
is
downloaded
right,
knowing
where
that
line
of
code
needs
to
go
to
right.
That
somewhere
is
is
actually
the
area
that
we're
trying
to
fill
in
here,
I
mean.
So
that's
why
I'm
suggesting
stick?
Oh
just
sort
of
look
around
those
items.
They
go.
C
I
just
linked
the
release
issue
in
the
chat
here,
so
you
can
take
a
look
at
the
way
we
worked
through
this.
Is
we
pretty
much?
Did
all
this
work
in
the
issue,
but
essentially,
like
I
I,
made
a
first
pass
at
like
here's,
what
I
think
these
things
should
be
and
then
Shinya
and
all
SEO
kind
of
commented
on
yeah,
here's,
here's
how
you
want
to
get
the
data
and
then
here's
places
that
we
don't
have
it
yet
and
then
yeah.
C
So
you
can
kind
of
like
read
through
that
process
and
then
there's
also
I
think
the
M
ours
that
I
added
were
yeah
they're
there
in
the
related
em
ours.
So
you
can
see
the
actual
changes
to
the
usage
bin
classes
to
to
make
those
changes
so
but
feel
free
to
hit
me
up.
If
you
get
into
the
weeds
with
any
of
this
and
have
questions,
that's
awesome.
Thank
you.
Derby
trivia,.
A
Okay,
the
next
two
issues
that
are
on
the
list:
they
are
issues
that
came
from
the
hacker
one
bug
report.
They
have
I'm,
not
sure
we're
supposed
to
talk
about
them,
since
this
video
is
public
and
those
are
confidential
issues
and
but
you
can
open
them
and
you
could
see
the
the
report
and
what
they're
related
to
I
think
they're
pretty
straightforward.
A
But
let
me
let
me
know
if
I
need
to
add
in
any
details
or
anything
like
that,
but
knowing
that
these
were
identified
as
p3
s3
bugs
and
they
were
identified,
they're
due
and
basically
August
in
September
and
I
figured
that
we
should
prioritize
them
now.
So
we
can
meet
that
meet
those
deadlines:
okay,
okay,
then
the
the
next
one.
It's
one
that
Nick
open
it's
it's
kind
of
a
meta
issue.
Right
now,
I'd
say
Nick,
it's
it
improved
the
container
registry
UI.
A
We
have
some
good
suggestions
in
there
and
but
I
think
we
could
do
more
to
turn
it
into
something
concrete
for
this
release
and
open
to
opinions
and
feedback
here.
But
one
thing
dan
and
I,
Daniel
and
I
were
talking
about
was
what's
the
de
tomate
like
what
data
do
we
need
to
surface
in
the
in
the
UI
for
the
different
features?
So
I
was
thinking
about
it,
as
maybe
a
separate
issue
could
be
instead
of
just
sending
the
the
UI
examples.
A
I
thought
from
some
of
our
competitors
to
actually
identify
like
okay,
here's
the
data
we
want.
We
want
the
branch
information
we
want.
The
last
commit
we
want
the
ideally.
The
doctrine
was
the
information
from
the
manifest
in
there
and
then
once
we
identify
that
and
where
that
data
comes
from,
that
could
help
solidify
how
we
design
the
user
interface
and
I
thought.
A
That
would
be
a
good
task
for
Nick
and
Ian
and
I
try
to
work
through
sort
of
like
figuring
out
what
information
needs
to
be
there
and
then
figuring
out
how
to
architect
that
information
into
the
into
the
user
interface
and
then
once
we
have
that
we
can
test
that
with
you,
some
of
the
users
as
well,
because
then
we
can
prototype
it
and
say:
okay,
how
does
this
presentation
look
to
you?
That's
how
I
was
thinking
about
tackling
that
problem?
What
do
you
guys
think?
What
do
you
all
think
sorry,
I.
E
I
was
going
to
agree,
so
I
think
that's
a
pretty
good
approach
that
we
take
there
because
I
this
many
time,
I'm
quite
unsure
about
what
day
is
available
and
how
hard
it
is
to
get
hold
of.
So
if
we
have
the
answer
to
those
questions
that
will
help
drive
UI
decisions
that
we
can
actually
implement,
rather
than
just
dream
about,
so
you
know
I'm.
A
A
E
F
A
I
was
looking
through
that
give
that
I
was
trying
to
figure
out
what
data
was
available
and
I
was
just
looking
through
the
get
lab,
API
itself
and
I
was
like
what
we
have.
We.
We
have
project
and
branch
and
commit
info,
and
all
this
stuff,
that's
available
from
the
API,
but
I
don't
know
if
we're
making
it
available
on
the
container.
F
A
G
Say
something
and
I
decided
against
it,
but
if
we
know
what
data
is
available,
that's
kind
of
the
question
I
had
around
this.
Was
we
already
know
what
data
is
available?
There's
some
really
easy
testing
to
figure
out
how
users
will
prioritize
it.
So
do
we
know
if
that's
a
fine
line
finite
list
of
what's
available
or
is
that
part
of
the
exploration
we
have.
A
Then
we
could
it's
just
a
question
of
okay.
Well,
where
does
all
that
information
belong?
And
some
of
that
will
have
to
say,
like
I
know
for
a
fact,
we're
not
I,
don't
think
we're
storing
the
docker
manifest
info
or
docker
labels.
That's
things
that
could
be
really
useful
for
the
user,
but
we're
not
storing
it
yet.
So
that
would
be
a
separate
issue
to
say:
okay,
start
collecting
and
storing
this
data
and
then
eventually
we'll
expose
it
on
the
front
end.
Gotcha.
G
G
A
D
A
Ask
them
to
click
through
things
and
see
how
it
works.
Okay,
it's
actually
moved
up.
This
card
identify
the
data
required
to
redesign
with
packaged
UI
so
that
that
one's
assigned
to
Ian
and
I
and
I
think
we
have
enough
to
start
out
with
and
okay
the
next
one
is
a
follow
up
to
some
of
the
NPM
authentication
work
that
we
would
that
we've
been
doing
in
12.2.
A
A
If
we
have
to
make
the
same
change
that
we
had
to
the
personal
access
token,
which
I
think
Steve,
we
just
had
to
change
the
header
for
how
it's
delivered,
but
I,
don't
know
if
that
same
change
will
work
for
CI
job
token
or
not
so
I
I
just
have
this
one
up
here:
I
have
what
we
want
to
do,
but
I'm
not
sure
how
would
actually
happen.
Yeah.
D
A
Users
do
have
a
workaround
for
this,
it's
not
ideal
because
they
have
to
use
the
their
own
generated,
Oh
F
token,
and
they
then
they
can't
have
two-factor
authentication
enabled
so
it's
problematic
it'd
be
nice
to
have
CI
job
token.
But
if
it's
a
big
change-
or
it's
really
involved
and
I
would
say
we
could,
we
could
have
them
working
with
the
workaround
for
a
bit
long.
A
A
It's
coming
out
and
stable
release,
I
think
in
August
or
September
and
Microsoft
has
already
worked
with
helm,
to
launch
helm
charts
in
their
container
registry,
and
the
big
change
that
happened
was
helm,
was
using
I'm,
probably
gonna
mess
this
up,
but
helm
was
using
tiller
as
the
backend
to
sort
of
do
the
user
management
and
to
store
the
files
now
they've
just
switched
to
the
container
registry.
There's
a
blog
post,
I
shared
it
in
the
ticket
where
they
talked
about.
A
Why
they
did
that
and
how
really
the
container
registry
just
holds
files,
so
it
can
hold
the
helm,
chart
files,
no
problem,
and
in
this
issue
a
couple
of
our
users
have
been
saying
well,
I'm
already
using
the
container
registry.
This
way,
it's
already
working,
there's
no
problem
with
it.
So
for
this
issue,
I
thought
we
could
go
through
with
those
users
showed
what
they're
already
doing
and
document
how
to
do
it
and
add
that
documentation
to
the
there's.
A
Some
container
registry
examples
in
our
documentation
that
we,
so
we
could
add
a
piece
to
that
and
I
suspect,
there's
some
fernand
changes
that
will
come
out
of
that
as
well.
I
saw
for
the
user
that
they
weren't
having
the
data
populated
for
size,
for
example,
I
think
was
one
of
the
fields
that
was
missing.
So
it's
really
more
of
a
get
it
working
documented
and
then
figure
out
what
needs
to
be
added
to
the
front-end
to
make
it
function.
Nicer,
I.
A
A
Just
so
you
know
thank
you
yeah,
because
it's
yeah
exactly,
but
this
is
a
nice
easy
win
for
us
and
it's
something
that
we
can.
You
know
I,
don't
know
how
big
of
an
announcement
that
helm
v3
will
will
make
in
the
tech
industry,
but
it's
enough
that
I
really
liked
how
github
announced
you
know,
support
for
Swift
in
their
packaged
registry
during
the
Apple
con,
so
I'd
love
to
do
the
same
thing
for
how
announced
it
during
their
conference
that
we
supported
as
well
any
any
questions
on
that
one.
A
No,
okay,
the
the
next
one
is
the
verb,
basically,
the
similar
to
the
work
with
that.
We
just
did
for
that.
Steve
just
did
for
MPM
for
subgroups
doing
the
same
thing
for
maven,
so
so
being
able
to
support
multiple
groups
and
subgroups
for
maven
and
with
that
ideally
would
would
we
would
change
how
we
handle
naming
conventions
for
the
maven
repository
as
well,
and
so
this
issue
itself
is
not
too
popular,
but
the
naming
convention
thing.
We
have
a
couple
of
issues
about
that.
A
A
D
D
B
A
B
A
B
A
Two
more
the
Steve.
This
one
was
your
issue
that
we
brought
forward.
This
addresses
a
problem
with
some
of
the
changes
we
made
recently
to
NPM
with
how
we
handle
the
logic
for
duplicate
names,
and
basically
we
want
to
warn
a
user
when
they
attempt
to
change
the
path
of
projects
and
I
didn't
know.
If
you
wanted
to
bring
this
into
this
iteration
I
think
it
could
wait
to
the
next
one
if
we
wanted
to,
depending
on
your
workload
but
I,
think.
D
There's
not
necessarily
a
hurry.
This
is
this
is
behavior,
that's
always
been
there.
It's
not
something
that's
changed
recently,
but
if
the
user
moves
a
project
or
changes
the
names
of
these
things,
it
will
break
the
connection.
So
that's
the
main
concern
and
they
won't.
They
may
not
realize,
but
that's
why?
Okay.
A
I'm:
okay,
if
we
wait
till
the
next
milestone,
if,
if
it's
feeling
like
we
have
a
lot
going
on
in
this
one
and
then
the
last
one,
is
this
sorting
issue
which
I
know
I'm
pushing
for
and
maybe
a
little
too
hard
for
for
maybe
I
hope,
I
hope
that
we
can
create
value
for
the
user
in
this.
But
there
are
performance
implications
of
sorting
in
the
container
registry.
Ui
we've
never
tested
those.
Really.
We
don't
know
what
the
where
the
performance
problems
start,
and
we
came
up
with
a
purpose.
A
I
came
up
with
a
proposal
for
only
sorting
the
first
page
and
then
anything
with
more
than
one
page,
don't
sort
it
basically
and
I,
don't
have
data
to
say
the
majority
I
would
like
to
say
the
majority
of
users
have
less
than
one
page
of
tags,
but
I
don't
have
the
data
to
say
that.
So
this
is
just
a
hunch,
but
I
think
this
would
be
a
big
value.
A
Add
for
our
users
like
when
in
each
of
the
user
interviews,
when
I
asked
them
the
question
of
how
is
how
are
they
using
the
user
interface?
They
were
like.
Well,
why
is
it
sorted
this
way?
It's
makes
no
sense
like
there's.
I
could
pull
out
a
quote
from
each
video
so
that
this
would
be
something
that
would
be
valuable,
even
if
we
could
solve
a
little
bit.
A
No
I
know
we're
running
out
of
time.
I
think
we
need
some
more
front
end
issues
as
well
and
noticing,
as
we're
going
through.
A
lot
of
these
changes
are
our
back
end
and
I'd
like
to
work
through.
If
I
find
a
couple
more
front,
end
issues
with
the
in
here.
That
might
be
more
helpful,
and
do
you
have
a
chat
scheduled
with
John
already
Tim
coming.
B
A
Yeah
I'd
like
to
that
sounds
good
I'll.
Do
that
and
Ian
can
be
there
as
well,
because
I
think
it's
gonna
be
a
big
piece
of
making
our
front-end
issues,
look
really
nice
and
and
and
work
really
well,
and
but
we
just
need
to
find
where
we
invest,
invest
at
the
time
it
was
good.
Okay,
I
used
up
all
of
our
time.
I,
apologized
and
I
know
you
wanted
to
talk
about.
It's
not.