►
From YouTube: Kubernetes Community Meeting 20160630
Description
We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.
https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY
1.3 release watch; 1.4 planning and feature workflow; SIG-cluster-ops and SIG-cluster-lifecycle; Fixit update.
A
Alright
good
morning,
folks
or
good
local
time
to
you,
if
you're,
not
on
the
u.s.
west
coast,
this
is
the
kubernetes
community
meeting
for
june
thirtieth
2016
and
we
don't
have
a
demo
this
morning,
the
last
minute
actually
twice
in
this
week.
Two
different
demo
errs
have
canceled.
So
we
have
no
demo
this
morning,
but
we
will
be
getting
updates
and
having
discussions
as
usual.
So
we
have
1.3
release,
watch
update.
A
B
One
of
the
note
is
I
talked
to
Aparna,
who
is
working
on
some
of
the
sort
of
PR
and
media
work
around
1.3,
and
she
suggested
that,
since
it
is
already
thursday,
the
the
advice
is
not
to
do
a
PR
launch
sort
of
on
a
friday
and
also
with
this
long
weekend
in
the
US
coming
up
with
the
July
4th
holiday.
It's
not
a
good
idea
to
do
that
right.
After
that,
the
PR
plan
is
for
Wednesday
July.
Sixth,
so
we
would
still
generate
the
build.
A
C
A
B
We
do
continuous
testing
on
the
1.3
branch,
and
so
since
the
large
cherry
pick
PR
got
in
we've
gotten
lots
of
cycles
on
that,
at
least
so
the
diff
between
that-
and
this
is
small
enough-
that
we
don't
think
it's
fully
necessary.
That's
it!
Obviously,
we
will
have
several
days
of
cycles
before
the
the
PR
part,
next
Wednesday.
Okay,
so
that's
that's.
C
A
B
A
A
Yes,
okay,
sorry,
I
was
having
microphone
issues
earlier
and
yesterday
brendan
and
I
spent
a
good
10
minutes,
debugging
and
hand
signals
yesterday
and
so
I'm
a
little
bit
worried
about
this
these
days,
so
no
more
on
1.3.
So
then
we
will
get
this
out
the
door
and
make
a
big
big
noise
about
it
next
week,
but
we
can
talk
now
about
1.4
and
feature
workflow,
so
evil
or
you
are
going
to
be
up
for
that.
Yes,.
D
Yes,
so,
as
most
of
you
I
remember,
we
started
collect
and
they're
wonderful
items
in
tracking
dashboard
items
have
been
a
22
wonderful
page
and
this
patient
assistance
smacking
to
generate
a
lead
one
to
review
the
get
Capri
for
during
the
next
days
and
on
the
next
week
after
will
respond
to
three
I,
don't
have
updated,
it
will
ensure.
What
do
we
have?
Wonderful
speaking
of
our
decision.
E
G
F
B
H
B
B
Basically,
we
just
talked
about
how
we
want
to
do
a
lot
of
prep
work
and
it
really
use
help
getting
that
at
time,
along
with
sick
docks,
so
I
think
really
where
it's
ever
now
is.
We
just
need
to
do
some
coordination
on
the
back
side
to
make
sure
that
we
were
all
aligned
and
not
just
tacking
on
Doc's
that
aren't
necessary
or
that
are
maybe
not
relevant
and
so
I
don't
know
who
is
a
leading
sig,
docs
John.
B
G
A
I
Sorry
I
was
late,
but
but
my
timing
was
great
okay,
so
the
cluster
ops
sig
is
working
on
reference
architectures
and
we're
hearing
pretty
overwhelmingly
in
several
quarters
that
documentation
for
how
to
install
how
to
install
is
important,
and
so
we
do
want
to
actually
have
a
hack
day
on
it.
We've
we've
sort
of
narrow
scope
down
to
just
hacking
on
the
reference
architecture,
from
hacking
on
docs
in
general,
Jason.
I
I
I
The
cluster
ops
sig
is
building
a
reference
architecture
document.
Let
me
find
it
and
I'll
post
it
in
the
chat,
but
basically
we're
documenting
all
of
the
components
for
running
kuber,
Nettie's
and
install
guidance.
Like
you
set
this
up
open
these
ports.
It
does
these
things
it
has.
It
has
these
prereqs
on
it.
This
is
how
you
build
an
op
cluster.
Would.
C
J
Guess
one
though
I'm
looking
not
sure
how
to
interpret
the
term
reference
architecture
in
some
communities.
That
means
this
is
our
preferred
way
to
do
it,
but
you
could
do
it
other
ways.
It's
more
like
you're
saying
this
is
a
reference,
it's
manual
for
the
pieces
that
explain
how
they
can
be
used.
Yeah.
I
C
You
know
advanced,
install
guide
might
be
one
way
to
do
it
right
or
if
you
just
switch
the
words
around.
So
this
is
an
architecture
reference
all
reference.
Oh,
we
know
I
think
this
is
great
because,
honestly
they
do
look
up
sort
of
the
custom
cluster
steps,
there's
a
ton
of
stuff
there
and
it's
still
not
quite
sufficient
to
figure
this
out
and-
and
I
think
you
know
over
time-
we
can
look
to
actually
make
this
stuff
be
smaller,
simpler
and
you
know
reduce
the
size
of
that
reference.
C
K
I
K
K
Definitely
people
have
said
that
could
be
improved.
Another
thing
I
would
like
to
see
is
for
each
individual
component
like
cube
api
server,
cute
controller
manager,
etc
to
have
sufficiently
useful,
helped
text
it
out,
so
the
people
can
actually
figure
out
what
they
need
to
specify
and
what
the
optional
and
what
kind
of
values
are
expected,
and
things
like
that,
because
we
do
hear
that
people
just
download
the
tarball
and
expect
to
be
able
to
figure
out
how
to
run
it
just
by
doing
help
on
the
binaries.
I
I
C
Know
what
you
know:
the
phases
and
the
tooling
around
those
phases.
That
would
be
huge
right,
because
then
we
can.
We
can
find
the
phases
that
end
up
being
in
common
across
all
sorts
of
install
scenarios
and
really
work
a
streamline
map,
though
I
think
this
will
be
a
great
light
to
shine
where,
where
the
work
needs
to
be
done
to
improve
that
experience,
yeah.
I
Was
I
was
getting
a
little
frustrated
because
I
couldn't
find
a
replacement
so
that
the
purpose
of
the
hack
day
is
for
people
to
come
in?
Who
have
you
knowledge
and
then
have
a
sort
of
coordinate?
What
you
working
on
all
these
pieces,
pulling
in
all
the
data
and
produce
a
produce,
a
reasonable
document
to
go,
and
we've
already
got
a
good
start,
I
think.
But
we're
really
short
on
the
details.
A
I
A
Cool
all
right
are
there
other,
yes,
you're
welcome.
Thank
you
for
bringing
up
what
you
are
doing
in
sick
cluster
ops.
Are
there
other
cigs
that
want
to
talk
about?
What's
going
on
today,
I
sentient
up
this
morning
to
the
cig
leads
mailing
list
asking
that
we
start
doing
this
much
more
regularly,
because
one
of
the
pain
points
in
this
community
has
been
cross
community
collaboration
as
we
as
we
are
growing
so
rapidly,
it's
almost
impossible
for
anyone
to
keep
on
top
of
everything.
A
That's
going
on
so
having
roll
up
summaries
into
this
meeting
as
well
as
notes
from
the
different
sig
meetings
in
2q
Brunetti
7
into
repose
is
one
of
the
things
I
had
asked
asked
for
this
morning.
So
if
anybody
is
game
to
volunteer
to
give
another
another
cig
update
from
what's
been
going
on
this
week,
that
would
be
awesome.
C
Will
I
will
I
want
to
shine
a
light?
I,
don't
know
if
it's
officially
a
cig
yet
or
or
what
have
you
but
there's
been
a
ton
of
discussion
going
on
around
and
it's
very
much
related
to
cluster
ops
and
that's
part
of
the
questions
that
we've
been
trying
to
explore
around
how
to
streamline
and
simplify
the
a.
C
You
know
mechanisms,
and
so
there's
another
cig.
It
was
originally
sig,
install
I,
don't
think
male
was
ever
sent
out
about
it,
and
now
it
got
renamed
to
sing
cluster
life
cycle.
You
know,
I
think,
there's
still
a
question
of
sort
of
how
these
things
relate.
But
the
idea
here
is,
is
you
know,
take
the
idea
around
sort
of
let's
document,
all
the
steps
necessary
to
bring
up
a
cluster
and
then
the
next
step
is
okay.
How
can
we
actually
start
simplifying
that?
C
Where
do
we
want
to
take
that
in
terms
of
of
taking
that,
what
is
probably
going
to
be
a
relatively
complex,
advanced
doc,
and
how
do
we
actually
start
boiling
that
down
having
some
some
reasonable
defaults?
And
so
a
lot
of
that
discussion
is
happening?
Sounds
like
you
should
work
closely
with
my
dog
I.
Don't
have
a
doc
which
talk
are
we
talking
about
unless
there's
another
Joe
and
so
the
oh,
that
was
those
Rob's
doc
yeah.
C
C
Put
a
ton
of
discussion
in
the
signal
cluster
life
cycle
mailing
list
in
and
a
lot
a
lot
of
conversations
going
on
there.
So
if
you
want
to
get
involved,
join
that
mailing
list-
or
you
know,
feel
free
to
reach
out
to
me
and
I
can
start
bringing
you
into
some
of
these
discussions.
Just
quick,
quick
straw
poll,
you
know
put
in
the
chatter
or
speak
up.
Is
this
news
to
anybody?
C
C
Yeah
I
just
want
to
make
sure
that
you
know
our
communication
channel
is
working
okay,
so
it
is
news
to
to
to
let's
see
what
we
have.
One
person
saying
it's
news
to
me:
a
lot
of
other
people
saying
they've
heard
it
so
yeah
I
think
we're
just
trying
to
make
sure
that
we
get
the
get
to
get
the
word
out
there.
Sort
of
you
know:
where's
everybody's
attention
is
that
so
that
you
can,
you
know
pile
on
or
ignore
it
if
you
like
and.
K
The
reason
for
creating
the
new
mailing
lists
and
whatnot
was
to
get
the
discussion
out
of
random
github
issues
and
into
a
form
that
people
can
actually
discover
and
see
so
we'll
figure
out
the
exact
permanent
home
of
where
that
discussion
will
be
happening.
But
you
know
we
do
want
a
forum
where
people
can
participate
instead
of
having
to
go,
find
several
different
github
issues
and
different
repos.
B
One
of
the
differentiators
as
I
understood
it
and
I
just
want
to
confirm
that
I
understand
properly.
Is
it
the
cluster
of
life
cycle
is
really
focused
on
putting
things
in
upstream
that
make
the
install
process
smoother
easier,
like,
for
example,
the
supposed
to
gratis
stuff
that
in
core
OS
m19
and
the
sick
cluster
ops
is
really
about.
We
have
a
working
cluster
now.
How
do
you
actually
maintain
a
upgrade
to
all
the
things
as
an
operator
you
would
be
concerned
with?
B
So
it
seems
like
the
DMark
is:
is
the
life
cycle
of
the
product
as
it
relates
to
being
a
installable
user,
accessible
product
and
then
cluster
opsys?
Now
we've
got
a
usable,
it's
bubble
product?
How
do
we
maintain
it?
How
do
we
make
sure
it's
documented
from
an
operator
perspective?
Is
that
does
ask
consistent.
I
My
video
on
so
there's
there's
multiple
layers
in
this
and
some
of
what
I'm
trying
to
do
is
learn
from
my
history
and
bring
bring
some
of
that
to
bear
I.
What
I'd
like
us
to
do
is
just
sort
of
work
on
the
DMark.
If
we,
if
we
need
demark
and
figure
out
how
what
we
want
to
do,
I
don't
think
it's
decided
yet,
because
I
think
there's
overlapping
missions
and
overlapping
missions
might
be
fine.
I
I
just
I
want
to
make
sure
that
we
deal
with
that
and
then
the
scary
word
vendor,
because
there's
a
lot
of
vendor
interests
in
the
middle
of
this
and
there's
people
being
obviously
paid
to
work
on
this
stuff.
And
so
we
need
to
make
sure
that
we
factor
that
into
the
house.
Discussion
works
so
that
like
clusters,
so
that
I
won't
call
it
11
cluster
or
another.
A
I
A
There
is
a
longer
discussion
that
is
being
planned
for
the
week
of
a
look
at
the
11th
to
figure
out
where,
where
and
what
sig,
cluster
ops
and
cluster
lifecycle
missions
fit
mesh
together,
split
apart.
So
if
you
have
interests
in
that,
please
let
me
know
because
I'll
send
invites,
but
the
idea
is
making
sure
that
saying
what
whomever
is
working
on
the
composable
pieces
of
making
the
install
easier
ends
up
also
getting
feedback
from
people
who
are
installing
engaging
with
it's,
not
you
know,
developed
from
an
ivory
tower,
etc.
A
So
this
is
clearly
a
brought
up.
The
fact
that
we
need
a
little
bit
more
information
about
cigs
with
end
and
share
information
about
what
everyone's
working
on
getting
that
collected
and
coordinated
as
I
mentioned
earlier,
so
we're
starting
to
I'm,
starting
to
put
in
place
a
few
more
processes
around
hey.
We
want
you
to
be
sending
information
back
up,
sending
up
to
specific
places,
starting
to
standardize
that
to
Doug's
point
about
too
many
methods
of
communications
to
follow.
I
tend
to
use
Cooper
Nettie's
dev.
A
As
a
summary
location,
discussions
seem
to
happen
in
slack
I'd
like
to
finish
and
follow
up
and
formalize
all
of
this
in
some
way,
so
that
it's
at
least
that's
the
expectation
and
and
that
we
know
where
we
can
go
to
get
a
broad
overview
of
the
project
and
then
where
we
will
get
pointers
to.
For
example,
the
communities
community
singh,
foo
repositories
will
have
or
directories
in
the
community
repository
will
have
information
about
what's
happening
in
the
specific
cigs
as
well.
C
Even
one
of
the
traps
with
open
source
projects
and
commercial
entities
is
that,
if
the,
if
a
commercial
energy
entity
solely
makes
money
off
of
providing
support
on
a
product,
it
creates
this
perverse
incentive
that
making
the
product
easier
to
use
and
easier
to
install
actually
leads
to
reduced
revenues
for
that
company,
and
we
really
need
to
avoid
that
trap.
And
so
you
know
if
anybody
on
this
call
is
planning
to
build
a
business
around
the
fact
that
Cooper
entities
is
hard
to
use.
It
should
be.
C
A
E
E
You
know,
use
those
and
so
on
and
so
forth,
and
and
and
we
want
those
tools
we
want
those
up
streamed
and
helping
people,
and
things
like
that,
so
I
I
totally
agree
with
Joe's
point
and
and
I
think
it's
just
up
to
us
to
have
a
transparent
discussion
here
are
the
areas
that
need
some
help
and
we'll
take
all
comers.
You
know
monkeys
at
keyboards
or
tools
or
whatever
it
takes
to
make
those
better.
A
So
back
to
the
agenda.
Are
there
other
cigs
that
want
to
talk
SIG's
only
one
twice:
alright,
we're
going
ending
up
without
a
demo
we're
going
to
end
up
with
a
relatively
short
meeting
this
week.
A
So
we
will.
We
will
continue
with
that.
Tomorrow
is,
of
course,
the
community
retrospective,
41.3
and
I
haven't
pulled
up
this
week's
numbers
or
the
to
date
numbers
yet
for
the
fix
it,
but
we
were
actually
trending
with
more
issues
closed
and
then
opened
earlier
this
week,
which
is
a
positive
and
unusual
trend
right
now,
so
we
want
to
keep
moving
forward
on
that.
There
is
a
fix-it
page
up
in
the
wiki.
A
In
sorry,
not
the
wiki
in
the
Cooper
Nettie's
community,
repo
under
community
fix
it,
which
has
different
places
that
people
can
give
can
actually
give
a
help
and
work
on
the
differ,
focus
areas
that
we're
looking
at
Thank,
You
Suraj,
so
more
information
there.
Also
at
the
bottom,
are
the
the
reference
URLs
that
I've
been
saying
we're
doing
better
or
worse
against.
So
you
can
take
a
look
there.
A
So
that's
a
developer
office
hours,
oh
and
I,
would
like
to
continue
as
a
program
the
developer
office
hours
so
that
we
can
have
people
asking
questions
of
the
devs.
Now
this
could
be
I'm
new
to
developing,
or
this
could
be
I'm
a
user
I'm
really
suck,
but
do
preferably
more
toward
the
developer
to
developer
work.
A
But
occasionally
we
can
answer
user
questions
as
well,
but
I
would
like
volunteers
for
more
people
who
might
want
to
do
in
office
hours
in
the
future
as
far
as
what
their
skills
are
with
the
office
hours,
I
did
with
Phillip
on
Tuesday
morning,
which
is
posted
on
YouTube.
If
you
want
it
with
the
developer
office
hours
that
I
did
on
Tuesday,
we
ended
up
going
back
through
the
slack
channel
and
just
answering
some
of
those
questions
as
we
didn't
have.
A
L
L
And
then
there's
probably
in
the
issues,
there's
probably
like
20
double
gates.
You
know
there
are
definitely
things
that
are
easy
to
fix
on
the
dark
side.
Yes,.
A
That
was
also
one
of
the
things
that
I
had
put
a
link
into
the
fix-it
recommendations
for
was
going
back
and
looking
at
issues
that
were
opened
before
the
first
of
this
year
and
to
see
if
they
are
valid,
duplicates
not
no
longer
valid
handled.
Some
other
way.
That
kind
of
things
well
so
closing
out
issues
is
a
great
thing
to
do
as
well.
Yeah.
K
And
actually
I
think
the
rates
of
filing
issues
has
increased
as
we
added
automatically
filing,
and
things
like
that.
I
probably
prefer
more
focus
on
triaging
recent
issues
like
the
most
got,
a
thousand
issues
that
haven't
been
prioritized,
so
those
are
ones
that
been
filed
within
either
the
one,
mostly
within
the
1.2
or
1.3
timeframe,
that
people
haven't
got
around
to
actually
looking
at
and
deciding
what
to
do
it.
K
So
you
know
a
lot
of
the
old
issues
are
ones
that
have
been
around
for
a
long
time
and
the
people
with
the
context
on
those
issues
are
the
people
who've
been
on
the
project
for
a
long
time,
so
I
was
actually
focus
more
on
the
recent
issues
which
you
know
maybe
were
filed
recently
by
people
who
still
care
about
them.
Then
issues
the
people
have
forgotten
about.
K
Closing
knows
that
also
helpful
yeah
I'd.
So
on
the
on
the
doctor,
ipoh
issue,
and
one
big
challenge
we
have
is
that
we
don't
have
any
of
our
automation
working
on
the
repo.
So
that's
something
we're
going
to
be
working
on,
but
if
you're
interested
in
helping
with
a
sort
of
maintaining
the
contributor
workflow
of
the
project,
that's
definitely
an
area
where
we
could
use
help.
I.
Think
for
this
fix
a
week.
We
split
up
a
list
of
doc,
repo
automation
that
would
help
into
lots
of
separate
issues.
K
So
picking
up
those
would
help
a
lot
more
than
going
through
the
current
pile
of
PRS.
We
need
to
get
automatic,
PR
assignment.
We
need
to
get
automatic
issue
triage
things
like
that
automatic
validation
of
formatting.
You
know
there
there's
a
long
list
of
things
that
would
help
make
the
doctor
ipoh
sustainable
as
opposed
to
just
solving
this
once
and
then
a
month
from
now,
we
have
the
same
problem
again.
A
A
So
for
those
of
you
who
don't
know
it's
what
I'm
talking
about
the
when
the
wind
Kuber
Nettie's
moved
to
the
cloud
native
compute
foundation
in
back
in
march
and,
however
long
it
took
all
of
the
paperwork
to
actually
flow
its
way
through
all
the
lawyers,
we
agreed
that
over
time
we
would
move
to
the
apache
CLA
or
the
apache
CC
LA
and
DC
0
for
each
commit.
So
that
means
that
we
need
to
change
all
of
the
tooling
that
currently
is
automating
our
CLA
signing,
and
that
is
in
the
works.
A
There
should
be
a
test
variant
of
the
new
CLA
bought
from
the
cloud
native
compute
foundation
late
this
week,
which
I'm
presuming
means
late
next
week,
given
the
holidays,
that
we
will
be
able
to
test
in
some
test
repose
and
see
how
that
works
for
our
needs.
Both
the
cover,
Nettie's
needs
and
cnc
f
needs,
and
once
we
have
that
tested.
So,
on
the
order
of
weeks
we
can
start
looking
at
rolling
it
out
to
the
broader,
the
broader
repose.
A
C
It
might
be
worthwhile,
in
light
of
the
CNC
F,
to
actually
open
that
up
and
be
a
little
bit
more
lenient
around
accepting
contributions
around
documentation
that
don't
require
the
CLA,
because
you
know
if
you're
working
at
a
big
company-
and
you
just
want
to
fix
a
typo
in
the
documentation.
You
know
you.
C
A
Ok,
I
will
follow
up
with
the
CNC
F
about
that
see
if
they
care
less
and
maybe
we
can
pull
the
Google
CLA
off
of
the
docs
in
some
other
manner.
If
we
can
find
something
else,
I
suspect
they're,
gonna,
even.
K
C
C
A
Will
check
in
with
more
lawyers,
because
this
is
my
favorite
part
of
my
job
and
I
will
see
what
the
CNC
f
thinks
about
the
CLA
on
the
docks,
repo
and
see,
if
maybe
we
can
get
that
lifted
at
least
and
help,
but
Justin
to
your
point,
help
with
looks
good
to
me.
Yes,
please
I'm
happy
to
review
that,
while
why
don't
you
change
this
picture
to
look
more
like
and
then
resubmit
would
be
very
helpful,
all
right,
any
clear
things
we
want
to
discuss
by.
K
The
way
mirpur
use
in
general,
we
would
love
to
have
more
help
from
the
community.
On
doing
my
reviews,
don't
feel
like
you
have
to
be
anointed
yo
PR.
If
you
are
familiar
with
an
area
of
the
code
or
you
want
to
become
familiar
with
an
area
code,
go
look
at
some
PRS.
We
have
500
open
pr's
in
the
main
repo.
You
know.
H
K
It's
super
helpful
and
we're
gonna
try
to
come
up
with
a
more
official
process
to
help
on
board
people,
and
but
you
know
we
are
PR
review
constrained
that
is
slowing
things
down.
So
if
you
can
get
a
bunch
of
the
basic
stuff
out
of
the
way
before
an
owner
comes
along
and
says
yeah,
this
all
looks
good.
That
would
be
tremendously
helpful.
Yeah.
K
Also
just
helped
new
contributors
with
basic
stuff
like
oh,
this
test
actually
failed.
Here's
how
you
find
the
results
like
that,
one
of
the
other
automation
things
that
we
have
filed
so
most
of
the
automation
issues
get
filed
in
the
contribute.
Oh
because
that's
where
the
current
home
of
the
our
github
bot
is,
but
at
one
of
the
items
there
is
identify
new
contributors
automatically
right,
so
we
could
I'd
automatically
identify
new
contributors
put
a
label
on
it.
Then
you
get
search
for
okay.
K
What
new
contributors
might
need
help
an
easy
way,
so
I'd
love
to
have
help
writing
things
like
that.
You
know,
but
we
need
to
grow
the
contributor
basically
to
help
people
get
their
work
done.
We're
going
to
be
putting
up
some
metrics
Antoine
has
been
working
on
some
great
metrics,
we're
going
to
try
to
figure
out
how
to
get
a
public
dashboard,
but
it
shows
a
sad
story
of
you
know
it's
taking
longer
and
longer
to
get
pr's
right
and
so
I'd.
J
Like
to
emphasize
that
you
were
saying
so,
I've
done
some
pr's
and
they
get
hung
up
on
flakes
and
but
I'm
I
tried
to
track
down
the
flake
on
but
got
stymie
because
I
don't
actually
understand
the
testing
system
very
well
and
wasn't
able
to
find
any
documentation
of
it.
So
documentation
of
what
I
need
to
know
going
to
be
productive
hunting
flakes
I
would
be
very
helpful.
K
Yeah,
maybe
that's
something
we
can
also
use
help
with.
Maybe
we
can
have.
We
actually
have
a
team
of
people
who
are
working
on
fixing
bunch
of
these
issues.
I
think
there
was
a
doc
shared
publicly
about
the
fixing
the
test
issues.
You
know
if
we
can
get
some
more
help
with
that,
then
the
new
people
participating
can
learn
the
ropes
and
help
document
all
the
stuff.
There
was
not
obvious
to
them
getting
some
fresh
eyes
on.
It
is
useful
for
that
sort
of
thing.
Uh-Huh
and.
K
F
F
Representative
of
the
fact
that
development
related
docs
never
made
it
over
to
the
main
docs
repost,
so
you
kind
of
have
to
know
that
if
it's
a
user
facing
feature,
you
should
go
to
Cooper
Nettie's
not
hit
github
that
I,
oh,
but
it's
a
developer
facing
issues
such
as.
How
do
I
get
my
environment
up
and
running
kind
of
testing?
That's
still
in
the
main
crew,
benetti's
repo
yeah.
K
C
C
I
think
one
of
the
things
we
can
do
is
we
just
call
attention
to
some
of
these
docs,
because
some
of
this
stuff
is
hidden
like
I'll.
Give
you
an
example.
Is
that,
like
you
know,
the
cage
Bob
will
say
like
hey
I
can't
merge
this
because
of
this
reason
you
know,
if
you
want
to
you
know,
if
it's
a
flake
then
find
the
flake
and
actually
put
the
issue
in
there.
Well,
the
actual
syntax
that
you
use
to
respond
to
that
bot
is
not
documented
anywhere.
That's
easy
to
find
from
that
thing.
K
We
also
need
to
do
things
like
notify
the
pr
author,
because
they
probably
get
the
notifications
filtered
out
in
their
inbox.
So,
yes,
lots
of
room
for
improvement,
and
we
could
definitely
use
help
for
that.
Like
all
the
maintenance,
I
know,
everybody
loves
putting
adding
features.
Are
we
really
need
help
with
a
lot
of
this
sort
of
projects,
maintenance,
contributor,
workflow
kind
of
stuff?
Would
that
go.
K
A
J
Sarah
Chisholm
is
so
yeah.
I
just
want
to
say
I
read
that
talk
about
hunting
flakes
and
it
didn't
help
me
with
a
problem
I
had,
which
is
a
test
ran
it
produced
some
stuff.
I
didn't
know
how
to
interpret
the
stuff
that
the
test
produced,
I,
didn't
I,
don't
know
any
documentation
about
the
test
that
Ram
I
don't
know
what
the
test
was.
I,
don't
know
how
to
paw
through
the
results
of
the
test,
to
find
or
interpret
the
evidence
of
what
went
wrong.
K
J
J
A
J
F
I,
don't
think
I've
seen
you
show
up
at
the
testing
stage
and
we
would
definitely
love
some
and
user
input
on
where
the
docs
are
deficient
right
now,
because
we're
all
moving
pretty
quickly
and
changing
a
lot
of
stuff
I.
Think
a
ton
of
effort
has
gone
into
improving
the
submit
Q
and
the
overall
velocity
of
the
project,
but
as
a
result,
we've
been
moving
fast
and
not
exactly
documenting
things.
F
I
think
there
was
maybe
a
hope
that
this
fix
it
period
happening
this
week
might
be
later
sort
of
do
that,
but
the
release
has
been
kind
of
dragging
on
a
little
bit.
So
it's
going
to
take
maybe
another
week
for
us
to
kind
of
collectively
catch
our
breaths.
This
sounds
like
a
fantastic
and
discussion
to
have
at
the
testing
thing
go
so.
C
C
A
And
2
I'll
mention
it
again.
All
of
these
signaling
lists
have
now
been
opened
up,
so
if
any
of
them
bounce,
just
let
me
know
and
I,
will
start
chasing
people
as
opposed
to
just
occasionally
sending
mail
I
think
I'm
down
to
two
but
I,
don't
remember
which
two
that
are
not
yet
publicly
available
all
right.
Anything
else.
Folks.
G
A
C
A
I
A
D
Sorry,
sorry,
so
one
item
from
the
agenda
to
be
good,
I,
simply
item
very
good
in
the
community
when
you
crave
it
I'm
great
organized
with
other
guys.
So
this
is
a
google
differs
and
I'm
leading
the
cloud
stream
and
prickly.
We
are
looking
for
their
from
speakers
familiar
with
communities
throughout
engine
app
engine
Hanson,
who
related
cloud
stop
stop
if
some
one
of
you
is
going
to
visit
Ukraine
and
the
beginning
of
September
and
like
to
speak
about
this
item
still
feel
free
to
reach
me
directly
and
and
we'll
see
you
in
stember.
D
A
Did
it
got
busy
I
thought
we
were
gonna
end
really
early
and
now
we're
gonna
be
ending
just
about
on
time.
Okay,
all
right
last
last
last
call
for
real,
otherwise.
Well
I'll,
see
you
next
week
after
the
fourth
of
july
holiday
in
the
US
and
a
normal
monday,
tuesday
and
wednesday
for
the
rest
of
you,
alright
see
you
guys
next.