►
From YouTube: cf-for-k8s SIG [April 2020]
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
A
A
Okay,
so
officially
welcome
to
this
week's
instance
of
the
cloud
foundry
on
kubernetes
special
interest
group
call,
as
mentioned
before,
the
recording
was
actually
started.
This
time
we
are
going
to
talk
about
UAE
and
the
ideas
the
USA
team
has
in
the
context
of
cloud
foundry
on
kubernetes,
and
we
have
Jeremy
from
the
team
with
us.
B
C
You
probably
know
we
had
a
spike
recently
to
integrate
native
UAE
into
cube
CF,
and
if
that
one
that
went
pretty
well,
we
were
able
to
make
it
work
and
we
consumed
the
docker
image
and
removed
the
boss
version
from
qcf
used
the
docker
image
and
everything
worked.
Fine,
though
the
only
thing
that
we
found
was
configuration
so
right
now
you
have
to
mount
the
UAE
llamó
config
file
into
the
container,
and
we
were
thinking.
Is
it
possible
to
do
configuration
environment
variables
and
could
that
be
like.
C
C
B
B
B
Know
that
the
UA
a
would
support
that
for
a
reason
of
practicality,
the
UA
is
highly
configurable
I.
Think
in
the
Bosch
release
there
are
over
200,
configurable
properties
and
I
think
under
the
hood,
there's
even
more
so
that
would
be
hundreds
of
environment
variables,
I
think
the
way,
but
that's
handled
right
now,
although
this
is
not
available
in
the
Cates
release
and
I'm,
not
even
sure,
if
it's
available
in
the
Box
release.
B
Things
being
on
the
file
system
versus
environment
variables,
I
also
know
that
I
could
I
could
speak
a
little
bit
to
the
security
aspect
of
that
we've
actually
chosen
with
respect
to
managing
secrets
within
the
UA
to
use
the
to
use
the
file
mount
like
so
the
secret
mount
writes,
that's
a
ram,
backed
mount
as
opposed
to
environment
variables,
I
believe
that's
kind
of
the
human
Nettie's
communities.
There's
a
preference
within
the
community
to
do
that,
because
it
is
perceived
to
be
more
secure
than
environment
variables.
B
B
Cam
yeah
sorry
early
morning
for
me,
and
so
it's
there's
your
secret
right,
which
is
to
me
not
that
big
of
a
difference
between
being
like
getting
access
and
then
I
just
go
and
look
at
where
the
secrets
been
mounted
into
a
ran
back
filesystem.
But
nonetheless,
that
is
I.
Think
the
fact
that
it
gets
backed
by
a
an
on
like
by
a
volatile
memory
source
is
what
people
perceive
to
be
more
more
secure
than
putting
it
into
the
environment.
B
E
The
other
hand,
while
I
was
like
because
I
had
a
specific
question
towards
a
property
that
is
exposed
to
the
Bosh
release
and
not
yet
in
in
Africa's
variant
and
wanted
to
ask
like
how
open
you
would
be
for
a
PR
to
expose
that
I.
Just
looked
up
the
things
and
the
UA,
a
young
is
actually
a
conflict
map
yeah.
You.
B
B
B
B
And
two
there's
two
phases
to
the
answer:
there's
two
parts
of
that
question.
The
first
is
totally
open
to
a
PR,
but
I
would
encourage
you
to
open
an
issue
first
and
start
a
conversation
about
the
change
to
make
sure
it
actually
aligns
with
what
we're
trying
to
do.
I
think
we
can
look
at
our
case
releases
like
a
zero
one.
Zero
release,
I,
wouldn't
exactly
call
it
stable
and
I
also
think
it's
I
should
highlight
that
we
have
some
turnover
in
our
product
work,
so
it
would
be
helpful
to
run
those
things
through.
B
B
More
involved
yeah
absolutely
open
to
contributions
once
we
get
through
that
initial
conversation,
and
we
all
agree
on
an
approach
with
respect
to
configuration,
that's
being
exposed
by
the
UA
yeah.
It's
it's
a
config
map
right
now.
It's
a
very
minimal
config
map,
like
the
number
of
properties
that
we
expose,
our
it's
small,
that's
actually
intentional.
B
The
idea
is
that
for
anything
more
complicated,
it's
bring
your
own
config
map
just
override
it
and
provide
a
different
configuration
you're
looking
for,
and
that
was
an
explicit
design
choice
that
we
made
for
now,
because
we
we
actually
don't
we're
not
happy
with
the
way
the
bot
release
works.
Like
the
configuration
for
that
the
configuration
story,
there
is
a
little
overwhelming
it's
hard
to
consume.
B
People
constantly
have
questions,
it's
it's,
it's
pretty
unapproachable
and
we
didn't
want
to
replicate
that,
but
we
also
didn't
want
to.
We
didn't
want
to
prevent
people
from
being
able
to
configure
the
UA,
which
is
why
we've
opted
to
make
the
UA
available
or
the
to
expose
the
configuration
in
the
way
we
did.
E
Okay,
so
could
you
think
about
actually
finding
some
way
to
overlay
configuration
done?
I
mean
maybe
not
with
ytt
overlay
templating
functionality?
Well,
that's!
Otherwise
it
would
be
a
hard
time
to
track
the
actual
default
coming
from
upstream,
if
there's
just
one
property
that
you
want
to
set
differently
from
the
I.
B
B
B
B
Think
for
us,
like
we
made
a
choice
and
we'd
like
to
validate
it
further
and
see
if
it's
easier
to
work
with
like
one
thing,
I
learned
is
that
we
bundled
the
log4j
properties
file
in
the
same
config
map
as
the
UA
amo
config,
like
file
and
I.
Think
that
might
make
things
a
little
trickier.
So
there's
probably
some
improvements.
We
can
make
along
the
way
for
sure
and
we're
definitely
open
to
feedback,
but
yeah.
The
idea
was
bring
your
own
UA,
a
llamo.
E
You
definitely
can,
although
it
gets
a
little
difficult,
because
the
the
actual
llamo
is
actually
part
of
the
y
TT
function,
but
like
by,
like
figuring
out
how
to
how
to
configure
copy
to
run
with
an
external
blobstore.
We
already
figured
out
that
y
TT
has
functionality
to
actually
change
the
content
of
y
TT
functions
by.
F
E
Intent
c
realizing
it
again,
but
that
was
like
not
a
way
of
overlay.
I
actually
meant
I
meant
something
like
there's
like
confit
directories
in
in
Linux,
where
multiple
config
files
would
be
there
and
there's
some
kind
of
handling
of
like
what
has
precedence
over
what
so
some
more
like
a
runtime
overlay,
not
the
templating
understood.
B
The
UA
a
actually
does
support
something
like
that,
although
that's
also
kind
of
scary
Oh
in
the
UAE,
it's
like
from
a
legacy
perspective,
we
explicitly
chose
not
to
expose
that
and
UAA
has
a
set
of
like
I
forget
how
many
environment
variables
you
can
tell
it
like
three
environment
variables,
like
eight
or
nine
different
places,
to
look
for
config,
and
it
will
kind
of
overlay
them.
It's
unclear
what
the
order
of
precedence
is
or
who
wins
or
what
your
actual
yeah
mole
looks
like,
which
is
why
we
don't
recommend
people
do
this.
B
Similarly,
the
secrets,
implementation
that
we
currently
have
kind
of
does
something
a
little
bit
similar
where
it's
overlaying
the
the
y
ammo
from
the
secrets
on
to
that
that
final
map,
so
those
I
know
some
communities
do
exist
again.
I
think
we
have
chosen
intentionally
not
to
expose
a
lot
of
that
behavior
just
yet
because
well,
because
it's
it's
led
to
really
confusing
experience
for
most
of
our
users
and
we
we
would
like
we'd
like
to
start
making
some
improvements
there.
So.
B
D
E
Ytt
overlay
is
the
way
to
go,
then
I
think
putting
the
config
amyl
inside
of
a
YTD
function
is
a
bad
choice.
If
you
actually
want
people
to
use
ytt
overlays
to
do
that
kind
of
configuration
which
works
like
like
even
better
than
ops
files
in
the
Boche
world,
because
ytt
has
like
more
features
regarding
this
overlay
kind
of
thing,
but
splitting
it
in
a
function.
Actually,
it
just
adds
an
obstacle
and
yeah.
G
B
Think
I'm
I
would
like
I,
prefer
an
issue.
I
think
we're
trying
to
you
know:
rehabilitate
our
open
source
presence
a
little
bit.
So,
let's,
let's
move
it
rollin
I
would
also
encourage
you
to
look
at
the
CF
for
Kate's
repo
and
see
how
they
interact
with
the
UA
a
config
map.
I,
don't
know
that
they're
applying
overlays
to
the
function
I
believe
they
might
be
applying
them
to
the
function.
Outfit.
Okay,
so
that
there
may
be
other
options,
but
don't
take
my
word
for
it.
C
So
if
something
like
that
was
available,
would
that
be
an
option
where
you
can
keep
the
config
llamó
and
also
do
environment
variables,
I'm
really
interested
in
exploring
a
common
way
to
configure
all
the
components
again
where
you
don't
have
to
keep
up
with
each
individual
for,
and
you
know
whether
they're,
overlays
or
I,
have
to
bring
my
own
config
config
map.
It's
like
I
have
to
maintain
a
piece
of
you,
a.
C
C
B
B
B
We
we
end
up
fielding
a
lot
of
product
questions
from
not
just
Cloud
Foundry,
but
other
products
that
have
used
the
UAE
as
well.
So
we
have.
We
have
to
keep
like
this
kind
of
general
general
list.
Point
of
view.
My
understanding
for
the
way
CF
for
Cait
is
consuming
the
UA.
A
config
is
like
they
will
expose
like
a
a
point
of
view,
kind
of
through
the
Y
mo,
and
you
can
modify
that
through,
like
y2t
overlays
or
whatnot
right,
so
they
it's
like
a
really
yeah
mo
based
approach.
C
Right
so
yeah,
of
course,
and
if
you
are,
if
you're
trying
to
build
a
cloud
foundry
foundation
from
multiple
components,
you
want
to
make
sure
that
the
configuration
is
easy
for
each
of
them
and
not
just
specific
to
Whitely
team.
Sure
you're,
saying
UAE
is
opinionated
about
defaults
and
you
only
expose
smaller
set
of
configuration
files
by
default
in
your
config
map.
That's
where
I
would
have
to
bring
my
own
config
map
make
sure
that
I
have
a
template
for
it
in
helm
or
whatever.
C
It
is
make
sure
that
I
always
always
understand
it,
and
that
I
keep
up
with
your
changes
when
it
comes
to.
You
know,
bumping
to
to
new
releases
sure
so
I
would
like
to
avoid
that,
if
it's
possible
and
the
way
that
you
know
you
can
configure
other
containerized
software
is
through
environment
variables
and
that's
a
generic
way.
I
also
believe
it's
secure,
I
think
it's
at
least
on
par
with
with
mountainous,
Rama
fest,
so
yeah
well,
I.
Think
there's.
B
Two
things
in
there
right
like
one
is,
is
that
a
convention
we
want
adopt
across
all
the
components
and
I
can't
speak
to
that
I.
Think
if
we
were
to
learn
that
that
was
a
need
across
all
the
components
that
that
would
be
something
the
UA
a
team
would
then
open
itself
up
to
I
with
respect
to
the
environment.
Variables
I,
like
I,
said
it
yo
a
is
highly
configurable.
B
So
I
have
my
doubts
that
how
how
much
Wow
I
would
I
would
need
more
of
a
conversation
around
how
manageable
that
would
actually
be
right
and
whether
or
not
that
actually
eases
your
pain
of
needing
to
understand
how
to
provide
that
configuration
via
llamo
and
understanding
that
structure
versus
managing
all
of
this
plethora
of
environment
variables
and
keeping
up
with
that
as
the
configuration
options
to
the
UA,
a
change
and
whether
or
not
we
even
think
that
would
be
a
manageable
way
to
present
the
configuration
right
like
I
think
we
do
have
some
challenges
there
and
those
those
challenges
might
be
germane
to
a
conversation
about
a
consistent
configuration
experience
across
all
of
pod
boundary
like
yeah
I.
B
B
C
C
B
C
B
A
little
less
worried
about
the
implementation,
I'm
more
worried
about
the
UX.
To
be
honest,
so
I'm
thinking
about
it,
use
your
orient
perspective,
which
is
why
I
was
like.
Oh
let's,
let's
try
to
if
we
think
that's
important.
Let's
get
that
conversation
going,
but
I
think
it's
a
and
I
want
cqe
on
the
call.
So
it's
like
I
think
it's
like
a
product
kind
of
experience,
question
and
less
of
a
you,
a
a
question.
G
The
question
about
environment,
about
having
way
too
quickly
without
rewriting
the
whole
configuration
file,
because
I
mean
it's
with
characters.
For
example,
when
you
take
a
himself
from
from
whatever
registry,
usually
you
know
nothing
about
it,
and
you
start
tweaking
only
the
things
you
want
to
tweak
right.
So
usually
there
are
variables
exported
and
you
can
only
set
the
ones
you
care
about
and
I
think
what
blood
tests
accomplish
here
is
having
a
way
to
change
the
configuration
of
a
without
known
every
every
option
and
rewriting
a
whole
file
right
right.
E
B
E
B
May
very
well
know
something
I,
don't
I
I'm
not
gonna
propose
to
be
an
expert
on
the
way
they've
consumed
it
that
doesn't
like
I
said
it
doesn't
align
with
the
conversation
I
had
with
them
about
using
some
of
the
default
like
some
of
the
templated
values
in
the
UA
versus
overlays
they're,
like
oh,
we'll
use
the
the
UA
a
templating
instead.
So
that
surprises
me.
That
said,
we
don't.
B
Was
news
we'd
not
believe
that
we're
done
with
the
configuration
story
of
the
UA,
so
I
just
wanna
make
abundantly
clear
to
everybody.
We
don't
believe
we're
done.
What
we
did
was
we
approached.
We
did
the
most
flexible,
minimal
thing
we
could
do
so
that
we
could
focus
on
operator
experience
and
making
sure
that
the
UA
a
was
operated
and
maintainable
in
a
foundation
and
return
to
the
cut
to
the
topic
of
configuration
and
start
gathering
more
information,
figuring
out
the
right
way
or
right
ways
to
improve
that
experience
for
all
you'll.
B
A
A
Pretty
quick,
I
I
wanted
to
repeat
what
what
do
we
pasted
in
the
chat,
so
I
think
for
the
benefit
of
Jeremy,
but
then
probably
also
others.
It
might
make
sense
to
at
least
reef
Lee,
introduce
yourself
and
probably
also
briefly
mention
the
company,
otherwise
generally
might
kind
of
not
be
able
to
correlate
what
has
been
said
on
this
call,
and
maybe
what
comes
in
by
a
github
issues
etc.
Later
with
that
go
ahead,
learn
yeah.
C
C
It
always
gets
wrapped
I
think
into
something
whether
it's
a
hound
chart
or
could
be
darker
compose
or
could
be
white
tea,
tea
or.
But
the
point
is
all
of
these
integrators
like
the
CF
or
Kade's
team
or
cubes
EF
team,
or
somebody
else
that
uses
you
a.
We
would
all
have
the
same
same
way
of
consuming
that
darker
image.
That's
it.
C
D
Okay,
now
I,
can
you
know,
hear
me
right:
okay,
Marvel,
you
hear
from
Suzy
also
quacks
team
and
I
just
wanted
to
throw
on
the
word
Perfector
in
here,
because
that's
what
wheat
customers
write
as
we
tell
users
of
Cloud
Foundry
to
the
twelfth
act,
two
EPs
and
one
of
the
twelve
factors
is
store.
Your
configuration
in
environment
variables
and
I.
D
Don't
think
every
internal
setting
has
to
expose
this
an
environment
variable,
but
in
general,
I
think
it's
good
advice
in
the
cloud
and
I
think
we
are
just
discussing
this
for
you
a
a
because
you
a
is
the
first
component
out
there,
that's
usable,
and
we
want
to
integrate
that,
and
so
we're
having
this
discussion
now,
but
in
general
I
think
it's
good
advice
for
all
the
components
to
behave
like
normal
cloud
software
and
follow
the
perfecter
approach.
I.
F
You
all
want
to
raise,
or,
to
you
know,
insurer,
supported
and
things
like
keep
CF
antia
frigates,
raising
like
raising
those
issues
in
the
appropriate
venue
on
on
github,
like
especially
given
a
lot
of
the
issues
we
have
with
timezone
and
overlap.
That
seems
like
maybe
the
most
productive
forum
that
we
could
discuss
those
in
Jeremy
did
you
want
to
add
something?
No.
B
I,
thank
you.
I
appreciate
that
again,
I
want
to
make
it
clear,
like
we're
happy
to
have
a
conversation.
I
am.
These
are
broader
conversations
that
prepared
to
answer
off-the-cuff.
I
just
wanted
to
say:
I
lost
the
thread
on
giving
introductions
I.
Think
I.
Could
we
heard
from
two
people
from
Sousa
and
some
some
more
please
for
the
environment,
to
consider
the
environment
variable
which
approach
which
not
saying
no,
also
not
saying
yes,
just
saying
like
wheatless,
let's
talk
wherever
the
right
place.
B
A
B
Actually
I'm,
looking
at
this
through
mirrors,
everything
was
fine
but
yeah.
I
guess
all
I
can
do
is
reiterate,
like
we
took
a
very
a
very
minimal
approach
to
configuring,
the
UAA,
because
it
was
it,
was
either
focus
on
getting
getting
towards
like.
What's
our
deployment
looked
like,
how
do
we
expose
metrics?
How
do
we?
How
do
we
focus
on
logging?
B
How
do
we
focus
on
redeployment,
or
we
spend
all
the
time
that
we've
been
working
on
getting
the
rest,
that
I've
done
on
trying
to
get
the
configuration
experience
right
as
opposed
to
just
leaving
it
open
for
y'all
to
be
able
to
use
this
component
and
I
think
by
making
that
choice?
It
is
one
of
the
reasons
why
this
component
is
usable
today,
so
yeah.
We
definitely
have
an
appetite
to
improve
that
we're
not
done
we're.
Not
we're,
makes.
C
B
B
C
B
F
G
C
So,
as
you
know,
we're
an
OS
company,
so
one
of
the
things
that
we
do
with
cloud
foundry
is
that
we
rebuild
everything
with
Susan
as
a
base.
So
that
means
that
okay
I
got
for
every
component
teams.
You
you
have
to
make
sure
that
the
base
image
is
configurable,
one
up
to
install
breaks
that
so
then,
what
do
you
use
and
Bill
packs
are
are
like
specifically
go
for
this
right.
G
G
One
other
question
have
so
right
now:
if
someone
starts
consuming,
locating
words
for
you,
a
a
does
that
mean
they
always
get.
The
latest
updates,
so
the
same
produces
both
versions
like
it
did
before
so
will
always
be
up-to-date,
so
the
dick
image
gets
released
normally
like
the
boss.
Room
is
right
there
on
par
version,
wise.
B
G
B
They
they
are
essentially
cut.
At
the
same
time,
our
our
image
is
a
little
like
I
said
it's
a
zero
and
zero
release.
Our
image
is
a
little
unstable
in
that.
If
you
use
latest
you
get
what
you
get,
but
when
we
cut
it
release,
we
do
tag
the
that
image
with
the
release
number,
so
you
can
go
and
find
it
and
use
it.
Like
I
said
the
only
thing
that's
a
little
different
right
now
between
Bosh
and
the
Cates
world,
which
is
actually
we're
starting
to
work
on
this
at
the
moment.
B
Is
that
once
that
that,
like,
let's
say
70
for
16
images,
is
uploaded
and
tagged,
it
doesn't
get
updates
to
the
the
base
layer
so
like
if
there
were
security
patches
to
the
OS
or
whatnot
so
like
in
the
Bosch
world,
you
get
like
a
refresh
to
the
stem
cell
every
so
often
we
aren't.
We
haven't
replicated
that
yet,
but
we're
actively
taking
that
on
now.
So
we
should
start
to
see
that
kind
of
start
to
change
as
well.
B
G
B
A
So
probably
a
totally
totally
unrelated
questioned,
like
my
understanding,
is
that
that
UAE
has
undergone
quite
an
evolution
right,
like
I,
think
kind
of
started
to
be
a
couple
of
things
like
an
identity
provider,
et
cetera
itself
and
I.
Think
somewhere
along
the
way.
There
was
a
decision
to
say
that's
kind
of
factored
out
the
identity
provider
piece
and
focus
more
on,
like
things
like
like
authentication,
right
I
was
wondering
if
that
is
something
that's
like
being
continued
in
the
context
of
moving
UA
a
to
kubernetes
and
and
then.
A
B
We'll
do
the
second
Congress
will
do
the
second
question.
First,
there
are
people
doing
that
already
like
the
UA
is
configured
as
the
as
the
issuer
for
how
IDC
tokens
and
whatnot
think
versions
of
the
pivotal
cluster
service
did
that
there's
a
couple
other
places
doing
it?
That's
something
we
would
love
to
do.
I
think
we
have
like
a
long-term
vision
for
the
UA,
a
to
kind
of
have
the
Cates
community
embrace
it
as
one
of
the
you
know,
embrace
it
the
way
they've
embraced
X
so
that
you
have
more
than
one
opportunity.
B
So
that's
kind
of
like
our
internal,
like
objective,
like
our
our
North
Star,
so
absolutely,
and
then
for
your
first
question
yeah,
you
know
I
think
the
way
I've
come
to
think
of
the
UA.
Is
it's
a
bit
of
a
Swiss
Army
knife
right?
It
does
a
lot
of
things,
and
maybe
it
doesn't
them
all.
Well
enough.
Maybe
it
doesn't
do
maybe
doesn't
want
to
do
them
great,
but
most
the
time
it
just
does
them
well
enough,
and
what
we're
trying
to
figure
out
is
like
does
it?
Does
it
really
need
all
those
blades?
B
There's
lots
of
places
where
that
might
happen
like
we
provide
a
lot
of
identity
management
tools,
and
we
kind
of
are
thinking
like
that.
Other
people
do
Identity
Management
much
better
than
the
UAA
would
ever
hope
to
do
it.
So
perhaps
that's
an
area.
We
should
invest
lesson,
whereas
a
with
so
often
Oh
IDC
like
there
are
continually
new
things
with
those
protocols
that
we
should
be
picking
up
like
pixie
support
or
whatnot.
B
So
we
can
then
start
iterating
from
there
and
actually
start
taking
some
action
to
bring
that
to
bring
men
in
control,
so
we're
early
stages
there.
But
we
have.
We
have
some
hypotheses
that
we
want
to
start
testing
and
hopefully
we'll
get
I'm
hoping
we'll
start
to
make
some
progress
on
that
now
that
we,
once
we
get
our
new
product
person
up
to
speed
and
we're
able
to
take
on
those
sorts
of
efforts
again,
I
know
does
that
your
question.
It
was
a
little
high
level,
but.
A
Yes,
it
does
I,
guess
probably
a
question
to
that.
Do
you
also
see
like
I
understood?
Yes,
of
course,
you
can
use
you
AAS
and
hook
it
up
to
your
kubernetes
cluster.
Do
you
also
see
kind
of
mixed
scenarios
where,
like
one
instance
of
you,
a
a
is
both
provider-
two,
for
example,
a
cloud
foundry
on
kubernetes
as
well
as
two
kubernetes
cluster
itself?
Oh.
B
You
know
I
mean
like
do
they
other
the
same
instance
of
different
instances
are
the
yesterday
that,
like
I,
think
the
UI,
a
team
spends
most
of
its
time
just
becoming
experts
in
these
identity
protocols,
which
is
already
a
really
deep
and
nuanced
space,
and
then
it's
kind
of
hard
to
like
pull
yourself
from
this
abstract
world
into
this
other,
like
really
detailed
and
broad
world
of
Cloud,
Foundry
and
I.
Think
you
know.
B
Perhaps
we
need
to
invest
a
little
bit
more
in
resourcing
to
find
people
who
can
actually
focus
on
like
that
identity
concerns
within
Cloud
Foundry,
but
do
I
think
they
should
be
the
same.
One
like
I
would
say
up
lahat
off
the
cuff.
You
know
maybe
for
a
small
footprint
like
foundation.
That
makes
sense,
but
there's
probably
a
bunch
of
security
concerns
around
that
as
well,
because
they're
representing
different
layers
of
access,
you
know,
but
it's
it's
a
it's
an
interesting
conversation
I'd
like
to
have
it
for
sure.
F
Yeah
it's
an
area
where
we're
engaging
relevant
as
well.
It
could
be
productive,
I
think
also,
we've
been,
we've
had
a
little
bit
of
discussion
on
CF
dev
Iran
things
like
cedar
deification
inside
of
the
Apple
in
time.
That's
actually.
Where
did
that
could
potentially
move
in
the
opposite
direction?
Where
you
know
it
actually
makes
more
sense.
A
B
C
C
G
C
F
A
C
C
Didn't
have
a
product
released
for
1.0
and
we
did
one
change
to
support
easier
integration
with
native
components
like
ireenie,
which
is
not
to
prefix
every
stateful
said
every
PVC
with
a
deployment
name
and
when
we
did
that
we
thought
we
could
implement
it
so
that
it's
my
gradable
from
version
1
to
2.
But
if
no
one
cares
about
it,
we
won't
and
it
turned
out
that
nobody
did
so.
We
thought
we'd
save
us
some
development
cycles
and
we
increased
a
2.0
because
assembler.
A
C
C
C
G
C
I,
don't
think
there's
anyone
from
IBM
here,
but
I
know
that
in
IBM's
case
they
tested,
they
have
my
sequel,
that's
outside
of
the
deployment.
So
it's
not
inside
SCF
or
cube
CF
right.
They
maintain
a
separate,
my
sequel,
and
for
that
they
didn't
need
to
do
anything.
They
just
pointed
cube
CF
that
my
sequel
and
to
the
blobstore
that
they
have
and
it
just
worked.
C
A
Yeah
I
mean
sure
yeah
I
mean
I,
think
in
one-one
early
version
of
the
document
I
even
had
something
in
like
this
will
probably
only
work
if
you
move
from
the
very
same
CF
deployment
version
to
like
the
same
target
version.
I
think
Marco
corrected
me
on
that
one
and
said
now.
Probably
it's
not
that
strict
of
a
requirement
at
least
things
should
be
relatively
similar
right
and
database
changes
in
during
that
stuff.
This
is
definitely
not
the.
C
E
C
A
Maybe
one
last
thing:
I
kind
of
tried
to
to
be
a
little
bit
earlier
with
sending
out
the
poll
and
also
closing
it
I
think
that
was
feedback
from
this
round.
Basically
saying
would
be
better
to
give
people
a
little
upfront
time
to
actually
know
that
they
are
asked
to
to
present
in
this
round
and
have
a
couple
of
days
to
to
prepare
upfront,
so
I
hope
I
can
actually
keep
that
promise,
at
least
for
next
time.