►
From YouTube: GMT 2017-10-05 Developer Meeting
Description
Agenda and Notes:
https://docs.google.com/a/mesosphere.io/document/d/153CUCj5LOJCFAVpdDZC7COJDwKh9RDjxaTA0S7lzwDA/edit?usp=drive_web
B
C
C
Volunteer
for
1.5
and
what's
the
target
date
I
feel
like
we
are
not
very
diligent
like
setting
our
release
date
and
the
release
on
time.
Things
like
this
I
think
we
should
improve
that
and
I
think
one
way
to
improve
that
is.
We
slept
pretty
expectation
like
when
I
released
ads,
where
we're
going
to
the
cut
things
like
that
and
also
I
think
the
last
I
think
I
discussed
it
out
that
how
do
you
provoke
irony
I
think
we
like
turn
like
a
compacting
force
during
the
release,
voting
process.
C
A
C
I
think
like
I,
mean
a
lot
of
time
that,
like
the
last
time,
I
had
a
chat
with
me
know
when,
like
say,
hey
I
want
to
retarget
this
21.4,
but
believe
that's
not
a
blocker
for
1.4,
so
I
think
we
should
set
up
some
process
like,
for
example,
retargeting
week,
so
that
so
that
we
have
a
rule
for
that.
If
you
miss
that
window,
you
cannot.
We
target
the
ticket.
You
have
two
weeks
window
to
reach.
Are
the
ticket
you
fee?
If
you
feel
that
a
blocker
or
it's
a
very
important
issue
for
you?
E
Agree
on
what
are
the
things
that
can
be
a
blocker
for
release
and
that,
in
my
view,
that
has
to
be
very
minimal
set
of
things
that
could
ever
block
a
release
like
it's
a
digression
such
a
great
blocker.
If
it's
something
that
was
introduced
in
that
release,
this
is
regression.
If
it's
a
new
feature,
it
should
not
be
locker.
E
It's
a
security
issue,
it
might
be
a
blocker,
but
if
it
again,
if
it's
existing,
if
it
has
like
it,
should
always
whether
it
has
to
be
done
in
this
release
out
communiques
pontil
is
to
me
if
it
is
already
existed.
Previous
releases
should
not
be
a
blogger
term.
If
it's
a
regression
from
something
that
we
used
to
work
and
it
suddenly
doesn't
work,
that's
the
only
thing
that
strip
your
blocker
in
there
are
any
API
changes
that
are
mid-flight
and
if
we
just
release
it
that
might
cause
some
issue
in
the
future.
E
E
C
C
I,
just
like
the
fact
of
the
risk
like
involved,
like
maybe
like
it's
not
maybe
good,
some
people
don't
notice.
For
example,
some
bug
my
resulting
data
loss
that
people
might
not
be
noticing
that
or
due
to
some
other
reasons,
race
condition.
Somehow
it
doesn't
happen
all
the
time
it
might
happen.
Sequence
is
severe.
E
F
I
feel
like
the
tension.
Is
that,
like
what
geez
she's
saying
this,
because
if
we
don't
treat
the
release
as
a
forcing
function
to
fix
that
issue,
we
might
not
actually
fix
it
right,
which
is
a
separate
problem.
Yeah
totally
separate
problem,
but
I
feel,
like
that's
kind
of
been
the
reality.
A
little.
D
E
C
E
C
C
G
I
think
it's
probably
a
issue
would
require.
You
know
the
judgment
on
the
base
case
scenario,
but
I
think
there's
a
bottom
line
here
right,
but
if
something
if
the
next
release
is
so,
if
is
this
is
a
critical
enough,
you
determine
whether
it
can
go
into
the
next
minor
release
instead
of
patch
release,
because
if
you're
you're
saying
well,
this
can
go
into
the
next
pass
hide
release
when
you
have
an
IVA
released
the
current
release.
E
F
Guess
there's
a
question
here
to
me,
which
is
like
was
it
that
wasn't
known
for
a
long
time?
We
didn't
fix
it
like
that's
a
process
issue
that
we
have
where
we
don't
actually
focus
on
fixing
server
bugs
and
the
review
should
the
release
should
not
be
a
forcing
a
forcing
function
for
that
there's
a
separate
concern
of
like
the
books
discovered
great
before
the
release.
I
think.
Maybe
that's
the.
What
thing
would
considering.
E
F
Agree,
I
think
if
we
were
really
diligent
about
fixing
issues
and
we
knew
that
it
was
gonna
make
it
into
Apaches,
then
it
wouldn't
be.
As
concerning
to
say,
here's
the
release,
there's
some
critical
issues.
This
is
where
they
are,
keep
that
in
mind
and
the
patch
releases
start
to
fix
some
of
that.
That's
right!
That's
what.
E
They've
been
trying
to
do
I
guess
you
know
Milligan,
but
we're
calling
on
critical
issues
now.
I
know
the
diesels
like
what
are
the
non
critical
issues.
At
least
people
have
to
know
it's
me,
but
to
me
like
just
because
if
you
found
something
one
day
before
versus
two
days
later,
doesn't
change
the
fact
of
the
severity
of
the
bodies.
E
H
To
chime
and
I
mean
I
never
have
seen
release
process
generally.
The
only
blocker
I
would
consider
a
release,
and
this
goes
to
when
our
argument
is.
If
there
is
a
regression,
we
should
do
a
visual
treat
it
as
a
blocker.
It's
not
a
regression.
We
I
mean
it
shouldn't
be
treated
as
a
blocker.
That's
it
it's
it's
that
simple!
As
that,
and
if
something
severe
as
found,
we
should
release,
noted
and
inform
everybody
that
we
have.
This
bug
and
people
are
working
and
they'll
come
up
in
a
patch
release.
E
I
think
one
other
thing
that
I
found
where
we
have
to
probably
block
release
was
cases
where
someone
was
doing
a
new
feature.
It
was
doing
some
changes
in
how
many
checkpoints
stuff
where,
if
it
didn't
fix
it
like
that,
then
upgrade
would
have
been
an
issue
down
the
line
in
the
next
release.
I
think
yeah,
I,
think
backwards.
Compatibility
to
me
would
probably
be
one
of
the
things
that
where
we
would
make
it
a
blocker
if
it's
somehow
breaks
backwards,
compatibility
that
is
where
I
would
say.
A
E
C
I
think
that
sounds
fine
prescients
spend
too
much
time.
We
have
a
bunch
of
so
I
think
the
Action
Army
is
I.
Think
as
long
as
we
stick
to
the
current
release
cycle
and
schedule
like
one
month
for
each
patch
release
and
the
three
month
of
remember
for
the
the
minor
release,
I
think
that
should
be
fine.
Let's
try
that
first
and
only
Walker
release.
If
it
is
regression
or
oh-
maybe
this
is
pretty
sure
that's
severe
enough
to
block
anything
else,
should
not
be
blocking
relief
and
stick
to
the
current
schedule.
C
C
E
C
I
I
The
company
is
supporting
open
source
project
with
CI
infrastructures
to
building
new
projects
on
arm.
They
provide
quite
powerful
machines
and,
from
my
perspective,
it's
like
a
side
project
to
make
messes
runnable
as
a
whole
like
master
and
agent
unarmed,
but
I
saw
there
is
a
interesting
community.
For
example.
I
Horizon
is
using
arm
servers,
so
I
just
want
to
propose
running
mesos
like
supporting
our
bases
architecture
for
masses,
so
this
is
required
to
get
resources
like
arm
service
from
act
and
plug
it
into
junkies
and
start
to
fix
mugs
and
making
like
proper
support
for
our
machine.
So
this
is
the
story
and,
like
my
question
is:
who
is
the
person
responsible
for
this
decision?
Like?
Is
there
any
interest
in
somebody
want
to
shepherd
this
topic
or
who
do
I
need
to
talk
to
from
messes
or
Apache
community
site
to
make
it
happen?.
C
Yeah
so
I
think
as
far
as
I
know
that
the
folks
from
Verizon
told
me
that
they
make
that
part.
The
missiles
are
armed
working
like
several
month
ago,
I'm
not
sure.
What's
the
current
status,
they
made
a
lot
of
patches
to
fix
some
of
the
issues.
Many
regarding
upgrade
leveldb
and
like
some
like
sick
bus
issues
due
to
alignment
I.
Think
yeah.
Can
you
imagine
those
patches,
I,
think
those
are
landed
already
so
I'm
not
sure
like
I,
think
one
thing
missing
is
the
CI.
C
I
From
my
perspective,
it's
like
building
messes
on
arm
is
currently
I.
Do
not
have
access
to
like
machines
that
could
be
delivered
by
packs
like
96
cars,
for
me
is
like
a
side
project
to
just
build
a
messes
caster
on
Raspberry
Pi,
and
that
was
what
I
proposed
I
think
two
years
ago
for
with
summer
of
cult
for
sure
she
was
not
accepted,
and
now
I
can
work
on
this
like
provide
a
patches
like
I,
do
before
with
leveldb
and
some
signals
and
caramel
yes
calls.
I
Fortunately,
it
looks
like
this.
Some
of
these
patches
also
fixes
problems
for
other
architectures
from
four
guys
from
IBM
I
thought.
The
Cisco
issue
and
leveldb
was
also
as
I
noted
as
free
night.
X
was
fixed
by
this,
and
some
I'm
searching
for
a
shepherd
who
will
be
like
can
give
some
time
to
review,
fixes
and
was
more
important
for
me
to
gather
resources
from
works
on
arm
initiative
like
Apache
foundation
or
I
know.
Some
commuters
need
to
commit
that.
Yes,
MSS
project
will
participate
in
this
program
to
get
this
hardware.
C
Result
yeah
because
I
wasn't
under
the
impression
that
at
a
time
the
Verizon
folks
to
me
like
months
ago
that
they
make
that
meso
saw
arm
working
so
I
assumed
that
most
of
the
issue
has
been
fixed
and
I
hope
to
upstream
some
the
patches
that
you
need,
because
I
was
under
dying.
Freshmen
I.
Don't
realize
that
there's
so
many
yeah.
I
Verizon
guys
are
running
only
agent
and
I
successfully
run
MSS
agent
with
Prayer
1.0
version,
but
in
a
current
master
there
are
some
dependencies
that
are
not
working
yeah
without
continuous
integration.
We
just
can't
assume
that
the
latest
release
of
messes
is
working
on
arm
to
the
Desmond
problem.
C
Cis
machine
to
support
our
moving
forward
so
that
we
can.
We
know
that.
Oh
my
patch
breaks
the
arm
and
I
need
to
fix
that
right
now,
like
everyone
is
in
the
dark
yeah,
so
that's
kind
of
almost
a
we
should
have
an
arm
machine,
ideally
like
it
should
be.
A
server
machine
can
be
a
raspberry
5
because
it's
otherwise
it'd
be
amazing,
most
likely
so
before
we
do
cross-linking
so
guys.
K
So
just
a
reminder
for
the
base
is
called
paging
this
year.
There's
two
guys
from
home.
They
keep
a
top
two
running
mates
as
a
marathon,
including
the
master
and
agents
on
the
armored
machine
which
I
just
paste
the
link
in
the
Google
Doc.
So
fiercely.
We
have
heard
a
couple.
Other
cases
use
cases,
people
in
committee
running
basis
on
arm
and.
K
That's
this
unreasonable,
just
like
starter,
like
mixer
song,
I'm
working
and
then
have
these
folks
working
together
to
discuss
the
right
solution
and
then
to
to
figure
out
like
what
do.
We
still
need
to
to
merge
into
measures
to
support
arm
instead
of
like
each
of
you
guys
doing
the
increment
your
own
patches
and
then
doing
it
privately.
L
H
C
If
you
have
any
patches
that
you
think
that's
gonna
well,
if
that
touches
that
I
could
container,
elated
and
stuff
and
feel
free
to
ping
me
I'll
be
happy
to
Shepherd
those
badges
and
for
the
lab,
ODB
I
I
used
to
remember
in
the
Verizon
has
a
patch
that
we
hasn't
learned
but
I
thought
like
we
once
we
upgrade
to
one
night
which
worried
it
that
just
solving
issue
but
I,
don't
realized
it
sooner
issue.
There.
E
So,
regarding
the
CI
I
think
someone
needs
to
donate
our
machine
to
the
ASF
Jenkins
pool,
and
once
someone
donates
it,
we
could
have
our
stuff
land
on
there
and
that's
what
we
try
to
do
with
IBM,
but
actually
they
currently
just
decommission
those
for
whatever
reason.
But
that's
the
only
way
for
us
to
have
a
CI
is
essentially
that's.
Probably
the
most
easiest
way
is
to
have
like
someone,
maybe
the
own
company
itself,
the
medical
machines
to
ASF
Jenkins
pool,
and
then
we
can
make
sure
we
run
our
build
on
that
machine
as
well.
E
I
E
I
E
We
need
someone
to
donate
an
arm,
one
or
two
boxes
to
SF.
E
E
E
C
So
most
the
case
is
not
like
the
understanding
like
arm
architecture,
because
it's
hidden
by
the
Linux
kernel
on
some
like
there
might
be
some
real
issues,
but
I
would
imagine
mostly
issues
like
operating
hours
like
dependency
and
I
level
DB
or
some
the
low
level,
because
it's
essentially
Linux
so
most
of
the
functionality
should
be
similar.
So
I
don't
expect
too
many.
C
C
You
have
acne
patched,
feel
free
to
sing
us
I
think
can
ping
me
or
ping
builder
or
giving
any
other
the
committers,
and
they
will
properly
for
or
just
paying
us
a
like
damn
channel,
and
if
we
have,
how
would
have
to
look
into
that.
I
will
be
happy
because
I
don't
need
much
work
to
make
it
work
and
that
there's
some
some
benefit
of
doing
that.
Okay,.
C
I
C
Tommy,
okay,
so
we
have
27
minutes
of
that
pants.
Are
you
on.
N
J
C
Do
you
want
to
do
you
want
me
to
finish
that
first
I
think,
like
I,
think
we
haven't
finished
the
world
by
release
yet
okay?
Why
don't
you
talk
through
that
and
I'll
just
use
the
last
funny
imbalances:
okay,
yeah!
So,
okay,
let's
do
the
release
panty,
so
once
I
release
so
been
no
event,
we
have
a
release
manager.
Four
one
five
already
know:
okay,
so
is
there
any
new
come
in
here
that
wants
to
volunteer
one
two
five
release
management?
C
C
N
C
So
another
thing
on
the
green
office
like
okay,
let's
go
over
to
you
I
gotta,
take
a
note,
say:
hey
we're
charge
for
end
of
November.
C
C
Like
or
something
one
thing
is
like,
maybe
like
every
time
he
eats
releases,
there's
a
major
feature
being
richer
best
blog
post,
Google
Doc
first,
like
very
early
and
people
start
filling
those
information
in
mm
about
the
the
new
features
and
new
features.
Rather
than
asking
releasing
the
ingrid's
to
do
all
these
release
blog
post,
which
he
might
not
have
any
contacts
for
I
kind
of
just
rip
the
workload
and
and
then
have
some
person,
do
the
editing
in
the
end,
to
make
it
coherent.
C
E
That's
in
progress
and
stuff,
like
that
and
I
think
the
first
edition
hasn't
come
out
yet,
but
that's
one
of
the
things
that
we
thought
they
would
publish
on
the
blog
is
the
newsletter
kind
of
thing
slightly
different
than
a
featured
blog
post
more
about
work
in
progress,
kind
of
thing
in
stuff.
Like
that,
so
that's
one
thing
we
could
do
in
the
blog's
section
but
yeah,
the
least
blog
post
would
definitely
be
useful.
Is
people
never
seem
to
have
time
for
that
stuff?
I
mean
yeah.
E
I
need
a
rock
well,
if
we
don't
have
time
to
write
user
dogs,
so
I
would,
if
you're
adding
anything
I
would
prioritize
writing
user
dogs
and
then
blog
posts
that
use
it.
We
have
not
even
any
user
Docs
right
now
for
big
features.
It's
very
limited
user
dogs
I
would
prioritize
that
writing
that
as
part
of
big
features,
even
before
talking
about
blog
posts.
So
this
is
related
through
the
doctor
down
intensity.
E
K
J
C
Yeah
I
think
so
yeah
I
think
so
I
think
Mike
doesn't
seem.
We
should
do
I.
Think
like
her,
like
I
worked
with
Gilbert,
take
initiative
for
the
once
I
release
blog
post
to
offer
like
a
P,
you
guys
on
any
new
features
that
you
get
getting
once
I
release,
and
then
we
start
a
blog
post
early.
So
that's
really
all.
Let's
start
this
discussion
already
like
regarding
releases
and
I
first
started
talk
right
now
and
asking
people
to
fill
in
the
blog
post
where
they
go,
I
mean
I
mean
didn't
some.
C
The
content
of
the
blog
post
has
Spanish
and
I
can
be
copied
from
dog.
That
should
be
fine,
but
we
should
start
early
rather
than
just
jamming
to
the
end
of
the
release
like
the
last
day
of
the
release,
because
they're
gonna
be
so
much
burden
to
write
that.
But
if
you
start
early
should
be
easier,
I
would
say.
K
B
K
C
E
C
K
Think
I
think
we
should
standard
standardized
the
release
process.
Every
time
we
release
a
new
version.
We
have
Apache
to
do
their
best
to
help
us
to
do
the
marketing
to
notify
the
Macy's
user.
Like
oh,
we
have
the
new
release
and
this
is
the
other
features
we
have
and
like
so
it
can
improve.
The
situation
like
right
now
may
be
many
other
misses
use
it.
They
are
still
using
like
0:20
a
new
27,
so
we
can
like
help
them
to
upgrade
which
we
adjust.
B
O
L
F
L
L
O
F
O
So
get
the
other
thing
I
want
to
talk
through
was
the
plank
or
mat
so
we've
talked
about
this.
A
bunch
and
I
would
really
love
to
see
us.
Do
this,
but
there's
just
a
bunch
of
things
to
talk
through
so
I
tried
to
put
together,
I
guess
a
plane
format
state-of-the-art
where
we're
at
today
and
then
we'll
just
talk
through
this
in
the
last
15
minutes
and
take
notes.
O
O
O
So
there's
no
way
to
do
that
with
my
informant.
So
we
have
a
choice.
We
can
either
set
continuation
indent
with
two
before,
in
which
case
we'll
get
this
more
space
indents
here
and
some
space
indent
here,
however,
we'll
also
get
for
space,
indent
other
places
which
I
should
I
even
segment.
So
the
one
of
the
main
reasons
why
to
do
the
consideration?
Sorry,
that
for
space
here,
I'm
guessing.
O
The
main
main
difference,
I
think,
was
to
differentiate
the
space
here
effectively
versus
the
space
than
in
the
continuation
here,
probably
because
they
do
that.
Did
everyone
follow
that
when
I
did
yeah,
so
so
it's
less
of
an
issue
for
us
I'm,
however,
because
they
probably
did
it
because
they
do
that.
My
guess
is-
is
that
it
could
possibly
be
an
issue
Landon's
since
we're
Atlanta's
you.
O
So
the
first
thing
that
you
start
to
get
is
the
continuation.
After
an
assignment
this
means,
rather
than
plus
two
here
you're
going
to
get
a
plus
for
the
other
place
you
get.
It
is
something
that's
like
this,
where
the
continuation
of
the
less
than
less
than
by
four
here's
another
Corky's
one.
When
you
were
super
long
line,
and
you
have
a
variable
name
afterwards,
that
will
be
now
be
contented
by
four
instead
of
two
by
here's
another
one.
You
have
something
like
a
binary
operator
in
dense.
O
O
O
O
O
B
O
Also
end
up
getting
some
funky
stuff
like
this,
where
basically
it
always
tries
to
align
the
question
mark
in
the
poem
aligning
them
here,
the
lines
in
there,
and
so
so.
Sometimes
you
just
have
now
a
truck
I
mean
trying
to
tweak
it
to
see
if
I
can
get
it
to
do
like
like
I,
don't
love
this
one.
This
is
where
it
actually
is
like.
Oh
I'm
gonna
try
this
so
so.
There's
that
one
okay,
okay,
let's
keep
going
bring
constructor
initializes.
So
this
one
is
an
interesting
work
where
basically.
O
O
All
of
the
circumstances,
with
these
two
declarations,
you
can
give
two
flying
format.
It
does
what
you
want,
where
it
actually
makes
them,
be
one
per
line,
which
is
nice
within
certain
tests,
work
and
basically
flatten
it
down
to
single,
oh,
no
way,
two
bases
take
always
do
with
do
one
for
line
small
complete
the
purpose
a
as
well.
O
Okay,
alright,
so
now
I'll
get
the
calls
so
in
the
latest
are
not
the
latest.
In
the
version
with
my
apartment,
amazing
images
from
Albion
Gordo,
you
can
actually
get
pretty
close
to
what
we
do
with
our
cars
so
get
so
that
the
order
order
of
preference
looks
something
like
this.
It's
on
the
same
line
it'll
try
to
do
that.
If,
however,
it
can
get
everything
on
the
second
line,
then
it
will
try
that
next
and
finally,
when
that
doesn't
fit,
it
will
go
to
this
corner.
O
O
O
It's
and
it'll
move
it
to
all
be
on
the
same
lines
again
wants
to
try
to
get
stuff
on
the
same
line
as
much
as
possible,
technically
two
lines,
whereas
in
other
places,
but
it
will
fall
back
on
place
where
it
pulls.
The
function
call
argument
there's
a
little
bit
inconsistent
in
that
way,
which
is
in
some
cases
it's
not
inconsistent
from
its
perspective.
But
from
our
perspective.
O
D
O
O
There's
one
weirdness
here
when
it
comes
to
the
calls
which
is
aggregate
initialization,
which
it
doesn't
treat
as
a
call
it
changes
all
the
code
from,
but
would
be
this,
which
is
what
we'd
expect
this
one's
saying
constructors
halt
to
this,
and
my
guess.
It's
just
because
plain
format
is
not
not
up
to.
O
O
O
O
So,
what's
interesting
is
when
you
do
chain,
it's
actually
smart
about
chaining.
So
even
if
you
could
have
fits
some
of
these
bends
on
single
line
here,
it'll
still
it'll
still
do
the
chain.
Now,
if
you
can
fit
that
first
lambda
and
everything
all
on
the
same
line
and
we'll
do
it
like
this.
If
ever
you
put
anything
on
the
body,
it
is
lambda
and
it
will
also
change
the
first
manner
which
it
being
trying
to
be
pretty
smart,
and
it's
recognizing
that
there's
other
chaining
down
here.
O
So
to
do
that,
if
you
don't
chain-
and
you
still
can't
fit
the
first
lamb
that
all
in
the
same
line,
then
then
this
is
kind
of
how
it
falls
down.
It
says:
okay,
well,
it'll
still
try
to
get
something
all
together
on
one
line
which
is
this
lambda,
so
we
can
do
that
and
get
this
whole
thing
on
one
line.
So
we'll
do
that
now,
although,
when
you
can't
fit
this
whole
lambda.
O
B
O
O
O
Okay,
all
right
and
then
here's
some
spots
where
you
sort
of
think
it
kind
of
gets
to
something,
slightly
consistencies,
which
again
consistency,
it's
more
just
for
just
following
these
rules,
but
places
where
say
if
somebody
woulda
grabbed
it
like
this
explicitly,
it
will
change
the
wrapping
here
now
it
doesn't
do
this
everywhere
again,
as
I
showed
you
there's
some
places
where
it
actually
is
keeping
the
wrapping
the
way.
Some
of
the
original.
O
L
O
O
Chip
you
decide
how
it
everyone's
to
see
asset
in
places
where
it's
doing
it.
This
way
pal
casino
case
we
have
wolf
as
well.
Unfortunately,
a
rather
different
system
I'll
be
doing
it
one
way
which,
like
all
right
and
then
the
other
place,
where
kind
of
world
where,
where
we
might
have
done
hates,
wasn't
formatting,
and
it's
now
like
doing
some
reformatting
cuz
things
like
this,
where
we
might
have
chosen
to
break
up
to
Falls.
O
Here
again,
it's
gonna
try
to
fit
as
many
things
as
one
line
as
we
can,
and
so
it's
gonna
pull
it
in
this
one's
not
actually
as
bad
as
I.
Think
when
someone's
trying
to
do
something
like
this
they're
trying
to
say
something
like
you
know,
what
we
have
is
a
numerator
in
the
denominator:
here's
the
numerator.
O
Okay,
alright,
and
then
here
are
some
of
the
unexpected
things
and
I.
Just
I
just
would
not
have
expected
at
all.
I,
don't
see
very
much
what
my
formatting
agent
it
does
things
like
this,
where,
again,
you
have
something
in
the
code
base
that
it's
actually
changing
a
lot
of
things
to
you'll,
actually
see
plan
format,
change
a
lot
of
things,
the
return
type
function,
name
break
plus
for
grandmas
and
it'll
change
it
to
return
type
and
then
pull
everything
onto
a
single
line.
O
O
C
F
O
So
so
what
what
I
was
thinking
that
the
protocol
I
was
thinking
is
that
we
could
we
could
start
using
it
and
if
people,
so,
the
idea
is
that
if
they
start
to
work
on
a
file,
my
suggestion
is:
they
run
five
Farman
on
the
file,
their
first
patch
in
the
review
chain.
It's
just
a
client
format
and
then
they
they
work
off
of
that.
O
O
O
F
O
J
O
F
F
O
F
And
I'm
for
it
I
think
it's
probably
operated
in
your
book
too.
To
me,
it's
the
yeah,
the
automatic
loop,
the
tooling,
in
the
project
that
gets
automatically
generated,
should
tell
people
if
they're
violating
the
client
format
stuff.
And
if
me,
if
you
want
to
make
that
on
a
file
basis,
then
that
could
force
them
that
change
the
file.
But
it's
kind
of
heavy
weight
for
new
contributors
and
stuff
to
have
to
do
that.
Maybe
or
we
can
just
do
it,
one
shot
yeah.