►
From YouTube: Kubernetes SIG CLI 20210526 - bug scrub
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
The
host
today
will
be
eddie
and
I
forgot
to
eddie
to
make
a
co-host
which
I'm
just
fixing
so
eddie.
Take
it
away
from
here.
B
Okay,
so
I'm
going
to
share
my
screen
and
we'll
start
with
the
triage
party
board,
which
has
a
new
config
array.
Did
anyone
have
anything
they
wanted
to
specifically
look
at?
We
can
jump
in
first
anything
on
the
customize
list
or.
B
B
Okay,
909.
A
A
Overall,
the
future
it
sounds
reasonable.
I
would
just
follow
the
the
guidelines
that
we
set
if
it
gets
sufficient
traction
and
sufficient
uploads.
C
I
got
up,
I
got
up
early
and
started
working
on
it
before
the
meeting
now.
Do
you
filter
out
things
that
don't
have
triage
accepted?
Is
that
how
this
works.
B
Yes,
that's
the
filter
right
now.
Let
me
pull
that
up
real
quick.
D
I
had
a
quick
question
eddie
on
those
two
that
we
just
looked
at.
There
was
a
triage
needs
information
and
there
was
still
the
label
that
said,
needs
triage
or
something
like
that.
I
think,
is
that
that
seems
weird,
that
there's
both
of
them
say
needs
triage
when
they
actually
have
some
kind
of
triage
label
is
mrmi
off
on
that.
B
I
think
that
needs
information
will
not
remove,
needs
triage,
okay
got
it
because
that's
like
we're
waiting
for
them
to
get
back
to
us,
and
I
think
this
one
when
it
yeah,
so
this
got
reopened
and
that's
when
it
got
needs
triage
applied.
So
I
think
this
was
an
early
issue
before
we
even
had
that,
so
all
right
quirks
good
catch,
though
sean.
B
Yes,
there
is
as
for
now
we
can
change
the
filters
to
be
whatever
we
want,
but
right
now,
it's
just
filtering
on
these
two
rules
in
here.
I'll
just
highlight
these.
B
C
C
A
B
Right,
well,
we
can
tweak
whatever
you
guys
want.
Whatever
will
be
helpful,
really.
C
Yeah,
let's
go
for
it:
okay,.
C
C
What
is
the
kind
on
this?
Can
you
scroll
down
a
little
bit
more
kind
of
the
api
version
v1
and
then
like?
C
E
The
reporter
seems
to
be
just
saying
that
it's
a
diff,
but
this
is
supporting
our
conversation
jeff
about
how
it
would
be
disruptive
to
people
to
bring
in
that
change.
Coming
from
the
mlb3.
C
So
this
is
the
thing
where
I
suspect
this
doesn't
change
the
the
way
kubernetes
will
respond
to
these
objects.
If
you
apply
them
to
the
api
server,
it's
just
a
different
formatting,
that's
screwing
up
some
sort
of
test.
C
I
guess
so
why
don't
you
put
triage
accepted
and
why
don't
you
assign
it
to
me?
I
could
look
at
it
a
little
bit
more
detail
later.
C
I
may
end
up
just
closing
it
because
we
can't
we're
not
go
back
and
patch
the
old
310
version
of
this
so,
but
I
can
take
a
look
at
it
and
just
confirm
that
it's
a
formatting.
B
Problem
cool
yaml
anchor
support.
E
There
are
a
bunch
of
issues
like
this
one
for
what
it's
worth,
I
that
that
may
or
may
not
be
duplicates.
I
think
we
need
right.
C
Yeah
angering
is
it
fights
it
can
if
it
can
it's
another
technique
for
sort
of
a
weird
form
of
templating
that
you
can
do
in
in
yaml
straight
up
yaml
that
can
sometimes
fight
what
you're
trying
to
do
with
customize.
C
I
think
there's
five:
why
don't
you
try?
Let's
put
a
needs,
help
wanted
on
this
and
somebody
who
can
find
a
specific
reproduction
for
it.
That'd
be
great.
B
C
C
F
C
Right,
I
mean
another
another
way
to
do
this.
There's
so
there's
usually
another
way
to
do
these
things
when
you,
when
you
try
to
use
your
ink
anchors,
you
can
use
like
a
a
name
prefix
or
a
label
directive
and
customize.
So
I'd
want
to
dig
into
why
I
want
to
ask
if
there's
another
way
they
could
do
this.
C
E
I
guess
part
of
the
issue
here
is
that
before
kml
we
were
just
reading
the
yaml
using
standard,
yaml
reading
methods,
so
before
customize
even
processed
this,
it
was
already
like
the
anchors
were
already
resolved,
so
we
didn't
care
if
they
used
anchors
or
not.
But
now
kml
is
trying
to
be
super
fancy
about
how
it
manipulates
the
data
structure
and
preserves
everything.
So
it
doesn't
actually
get
expanded.
E
The
way
you'd
expect-
and
we
end
up
having
to
know
about
these
features
of
yaml
that
otherwise
we
could
just
be
agnostic
too,
so
like
it
worked
before,
and
it
doesn't
work
now
I
guess
is,
is
part
of
what
all
these
issues
are
pointing
out.
E
F
E
Well,
I'm
not
sure
what
we've
decided.
What
we
want
to
say
about
it.
E
Do
do
we
want
this
to
work?
Do
we
not
want
it
to
work
like
we
could?
I
think
we
understand
why
it
doesn't
work.
It's
that
it's
the
switch
to
kml
and
and
therefore
like
we'd,
have
to
re-implement
all
of
the
yaml.
All
the
ammo
features
ourselves
if
we
wanted
them
to
work
like
it's
just
not
handled
by
kml
right
now,
but
do
we
want
it
to
be
or
not
like
to
jeff's
point?
E
There
is
always
another
way
to
do
this,
customize
isn't
designed
to
work
this
way
specifically
and
has
its
own
mechanisms
for
handling
deduplication
and
and
string
sharing.
I
guess
is
what
they're
doing,
but
you
can
also
see
the
argument
that
this
is
just
yaml.
Why
are
we
not
supporting
standard
yaml
features
when
we
parse
our
input?
C
Yeah,
so
I'm
on
the
side
of
not
not
spending
too
much
time
on
trying
to
get
anchors
to
work.
In
all
cases
I
mean
it's
standard.
Yaml
is
a
little
bit
debatable.
What
standard
gamma
is
so.
C
Well,
yeah,
but
the
spike
is
also
like
got
some
inconsistencies
in
it
right.
So
what
we
could
try
is
see
if
there's
a
family
of
these
kinds
of
issues
that
involve
anchors
and
maybe
there's
somebody
who's
who
can
own
it?
Who
cares
about
anchors
and
wants
to
use
them
in
concert
with
customize,
because
there's
it's
gonna
be
complicated.
There's
gonna
be
lots
of
tests.
You're
gonna
have
to
write.
C
I
suspect
that
we
would
encourage
them
in
this
particular
case.
I'd
say:
what
are
you
trying
to
accomplish
and
can
we
accomplish
it
some
other
way
without
using
anchors?
E
C
Mean
the
thing
the
thing
is:
if
you
like
yeah,
it
will
also
get
complicated
with
overlays
right
if
you
expect
them
to
work
across
anchors
to
be
recognized
across
overlays.
That's
also,
there's
gonna
be
a
problem.
There
too,
I
mean
it's
theoretically
possible,
but
it
would
be
tricky
and
customized
as
other
mechanisms
for
achieving
the
same.
G
B
C
B
B
B
C
C
So
we
have,
it
may
fall
under
the
category
of
caching
stuff.
That's
recovered
from
like
remote
urls.
There's
another
person,
that's
working
on
this
with
git
repository.
So
every
time
you
try
to
use
a
remote
url
as
a
base,
it
downloads
it
because
customized
traditionally
hasn't
been
in
the
caching
business.
E
C
No,
I
don't.
I
don't
think,
there's
a
if
that's
what
it
is.
I
don't
think
there's
a
any
code
for
looking
for
a
local
repository.
I
think
it's
trying
to
get
it
from
I'm
trying
to
treat
the
urls
as
getting
the
getting
charts
from
a
re
remote
chart,
repo.
C
C
Maybe
maybe
maybe
repo,
if
that
might
also
be
a.
C
I
think
that
stable
is
a
default
value
for
a
variable
for
a
place
to
look
for
charts,
and
I
think
that
error
looks
like
a
customized
error,
but
I
might
be
wrong
in
any
case.
C
C
D
A
C
E
C
I
think
it's
just
an
attempt
to
use
a
to
use
a
field
that
expects
to
go
to
a
remote
repo
for
in
an
attempt
to
open
a
local
file.
C
G
C
I
mean
the
so
customize
started
three
years
ago
and
wrote
it.
It
evolved
a
a
file
system,
abstraction
back
when
you
know,
I
think
several
other
people
started
file
system
extractions
about
abstractions
about
the
same
time,
to
support
writing
unit
tests
and
pulling
all
that
out
and
replacing
it
there'd
be
a
really
good
reason
why
you
wanted
to
do
that.
C
It's
not
it's
it's
throughout
the
whole,
throughout
all
the
code,
so
I'd
have
to
like
have
a
somebody
would
have
to
give
a
good
concrete
reason
why
this
would
be
better
and
how
flow
would
be
improved.
It's
not
obvious
to
me
right
now
why
that
would
be
the
case.
A
A
E
C
C
B
B
C
A
I
think
one
of
the
commenters
is
saying
that
it
might
be
a
it's
possible.
I'd,
be
curious
to.
A
To
split
the
output
run,
customize,
build
and
type
that
into
the
file
and
then
try
to
apply
that
separately
so
that
it's
clear
where
the
problem
is
and
most
likely
it
might
be
a
problem
with
actually
applied
yeah.
It
looks
like
I
agree.
It
looked
like
an
apply
problem
and,
secondly,
I'm
missing
information
about
customize
and
keep
cuddle
expressions.
B
C
Well,
I
mean
in
this
in
this
case
they
should
just
do
the
build
customize
build
and
see.
If
that's
a
problem,
I
don't
think
the
error
was
from
the
build
the
build
step.
D
A
And
versions
of
customizing
out
there
running
this
against
and
the
cluster,
it
seems
weird
that,
if
keep
cutting
apply
would
not
work
with
crds,
we
would
be
getting
autonomous
about.
E
It
silly
question:
as
a
non-admin:
how
do
I
get
proud
to
remove
the
needs
triage
label?
I
just
tried
a
bunch
of
things
and
it's
not
answering
me.
E
A
A
A
In
that
case,
just
manually,
if
you
haven't
access
it,
not
I
think
jeff
has
the
admin
access
to
do
it
in
a
customized
repo.
C
I
I
can
do
it.
Katrina
at
the
moment
is
not
on
the
customize
admins
team
and
I
think
that
may
be
what
allows
you
get
to
get
the
ux
for
changing
labels.
Yeah.
E
A
I
think
that's
a
new
new
thing
for
oh
for
github
just
introduced
recently.
B
C
B
C
Right,
I
believe
this
is
a
legit
bug.
The
question
is
given
that
we
want
to
deprecate
vars.
We
want
to
work
on
it,
we're
replacing
bars
with
this
replacement
transformer
which
is
going
to
you
know
it
yeah.
E
C
I'm
pretty
sure,
but,
like
it's
been,
it's
been
talked
about
as
a
deprecated
issue
for
like
a
couple
years
now,
so
what
we?
The
problem,
is
it's
work,
so
the
attitude
that
phil
and
I
took
was
basically
look.
It
does
what
it
says
in
the
box
right
now
and
we
we
just
haven't.
We
have
been
lazily
deprecating
this,
since
we
haven't
been
working
on
it
for
api
version.
C
1.0
though
we
do
want
to
actively
remove
it
because
it
fights
other
solutions,
we're
trying
to
come
up
with
it's
very
it's
a
very
invasive
feature
and
it
also
kind
of
violates
one
of
the
things
customize
attempts
to
do,
which
is
to
work
with
bases
that
have
nothing
special
in
them.
No
template
variables,
no,
no
things
that
need
to
be
processed
right,
so
vars
currently
require
you
know
a
dollar
variable
kind
of
notation.
C
A
I
don't
know
basically,
whenever
someone
is
using
a
var,
throw
a
warning
that
in
some
time
will
be
deprecating
use
and
the
plug-in
you
mentioned
instead
and
let
it
be
like
that-
and
you
know
be
a
little
bit
more
vocal
about
the
deprecation,
oh
to
users,
and
then
you
know
just
start
the
process
and
in
a
couple
of
months
actually
do
it.
A
C
Right,
the
other
thing
we've
been
doing
is
working
on
customized
as
a
fixed
command,
and
the
purpose
of
the
fix
command
is
to
read
in
a
customization
file,
look
for
things
that
have
changed
and
write
out
a
new
customization
file.
That
does
things
in
a
new
way,
and
that
is
something
we
want
to
provide
with
vars.
C
To
the
extent
that
is
possible,
so
yeah,
so
the
message
so
error
message
on
anybody's
using
vars:
yes,
that
we
can
do
email
announcement.
Yes,
we
can
do
anything
else.
A
No,
I
think,
that's
sufficient,
you
know
you
put
it
in
release,
notes
and
you
will
be
throwing
those
warnings
and
people
will
be
processing
and
then
a
couple
releases
just
drop
the
functionality
and
it
and
and
that's
it
I
think,
that's
the
issue
that
is
talking
about
enhancement,
replies
bars,
maybe
just
point
to.
C
A
A
Okay,
probably
we
can
probably
extend
the
deprecation
period
for
as
long
as
we
wish
to
or
for
as
long
as
people
will
be
complaining
and
eventually,
if
needed,
we
can
always
provide
a
fallback
flag
for
some
time
for
the
for
everyone
that
will
be
super
late
to
the
game
and
then
eventually
just
get
rid
of
it.
I
think
I
mean,
as
long
as
we'll
be
vocal
about
this
change,
I'm
all
in.
C
Oh,
so
so
back
to
that,
are
you
back
on
that
initial
one
yeah?
So
I
would
say
there
that
comment
that
if
you'd
like,
I
could
comment
on
this
too,
but
we
could
comment
that
we're
not
gonna
fix
this,
that
it's
just
that
bars
are
gonna,
be.
B
C
C
By
the
way
we
now
have
14
issues
with
hell
pointed
labels
on
them.
So
I
want
to
build
out
like
about
30
or
so.
C
And
get
some
more
basically
parallelize
the
work
get
more
people
involved.
B
E
Are
labeled
one
time
less
first
time,
issues
that
help
help
wanted
is
kind
of
a
not
quite
as
low
bar.
If
I
understand
the
distinction
correctly
like
true
first
time
issues,
I'm
not
convinced
we
have
a
lot
of
those.
C
Me
me:
I
agree
with
katrina
the
help
point
it
is
for
somebody
who
really
wants
to
dig
in
and
understand
the
code
so
also
to
facilitate
that
process.
I'm
adding
an
architecture
document
that
explains
how
customize
works,
kind
of
a
code
walk
to
help
people
get
up
to
speed
on
on
fixing
things
that
are
more
complicated,
but
I
think
the
ones
that
are
labeled
help
wanted
are,
relatively,
I
wouldn't
say,
they're
super
easy,
but
they're
also
localized
they're,
not
like.
Oh,
let's
replace
the
file
system,
abstraction
or
anything
really
huge
or
like.
A
E
Customize
config
command,
though
there's
some.
F
C
In
the
customized
repo,
so
if
somebody
wants
to
tackle
it,
that'd
be
great.
B
Okay,
all
right,
so
we
have
10
minutes
left
I'd
like
to
talk
through
something
that
I've
been
putting
together
I'll
share
it
here.
We
don't
all
have
to
read
it
right
now,
but
the
pldr
is
we
keep
getting
all
these
issues
opened
about
people
accidentally
deleting
all
of
their
cluster,
with
mis
commands
and
misunderstanding,
how
deleting
namespaces
works
and
throwing
all
flags
and
all
spaces.
B
So
the
most
recent
one
was
like.
I
think
this
was
even
like
yesterday
yeah,
so
this
person
opened
this
issue
yesterday
and
so
the
there's,
no
confirmation
when
you
go
to
delete
your
namespaces
and
throw
all
on
there
and
it
just
blows
everything
away
and
deletes
all
your
resources,
and
we've
historically
said
that
there
is
it's
a
breaking
change
which
it
is
it'll
break,
anyone's
script,
who's
using
that
or
relying
on
that,
so
introducing
a
confirmation,
prompt
or
a
new
flag
would
require
people
to
modify
their
existing
scripts.
B
We
talked
it
over
slack
yesterday
with
the
other
leads,
and
I
think
we
all
came
to
agreements
that
we
should
probably
do
something
about
this
as
the
rate
of
people
running
into
it
is
going
up.
So
I'm
going
to
send
this
out
to
kdev
at
some
point,
so
folks
have
time
today
to
give
it
a
review,
give
it
thoughts,
kind
of
laid
out
a
plan.
It's
basically
the
first
pass
at
a
cap
really
about
a
process
to
move
towards
confirmation.
Add
some
flags
that
make
it
easier.
B
Other
bits
in
here
are,
if
you're
using
dash
n
while
operating
on
a
global
cluster
scoped
resource,
it's
a
no
op
and
has
no
effect.
So
we
consider
that
to
be
an
implementation
bug
and
it
should
error
and
warn
you
there's
a
few
changes
in
here
that
want
to
suggest,
as
kubernetes
is
growing
in
popularity
and
new
people
are
coming
in
and
accidentally
deleting
all
this
stuff,
so
give
it
a
read.
B
Please
give
feedback
and
I'd
like
to
send
this
out
by
end
of
day
today
or
tomorrow,
so
any
thoughts
or
questions
on
that
early.
C
A
So
when
we're
talking
with
eddie,
because
that's
not
the
first
approach,
I've
been
looking
at
something
similar
in
the
past
and
it
was
shut
down
back
then.
B
Yeah,
so
that's
the
plan
I
laid
out
here
in
this
alpha
beta
ga
thing
over
the
course
of
three
releases
for
one
year
we
can
start
by
introducing
a
ask
for
confirmation
flag,
and
so
I
don't
realize
that
rm
has
this,
but
if
you
do
rm
rfi,
it
will
always
override
the
force
and
always
ask
for
confirmation.
So
there's
already
prior
work
out
there
of
always
enforcing
some
sort
of
confirmation
flag.
B
B
That
sounds
like
a
good
idea
and
then
so
part
of
the
first
release.
I
also
said
we
should
add
a
confirmed
force
flag.
People
can
just
start
using
this.
B
It
really
it
would
be
a
no
up,
but
you
could
start
putting
it
in
your
scripts
and
then
it
would
also
start
warning
that,
hey
by
version
x,
we
are
going
to
be
doing
this.
You
know
we
can
roll
out
blog
posts
and
release,
notes
and
always
call
this
out
and
try
to
educate
me
as
much
as
we
can
and
then
pretty
much.
I
said
within
one
year,
so
three
releases
is
when
it
would
require
either
conform
firm
to
be
passed.
B
I
not
married
on
any
of
these
flag
names
either
so
but
yeah,
that's
the
tl,
dr
and
there's
tons
of
issues.
It's
like
it's
getting
really
bad
people
are
blowing
their
up.
G
B
G
B
Yeah,
so
I
liked
looking
at
the
rm1.
So
if
I
like
go
into
temp
real,
quick
anything
so
like
touch
foo
right,
so
if
I
rm
rfi
foo
and
it
starts
saying,
would
you
like
to
examine
the
files
and
foo
remove
this
file?
Remove
this
file
right,
so
we
don't
necessarily
have
to
show
them
everything
we
think
individually,
but
like
hey.
These
are
the
files
that
are
going
to
be
removed.
Are
you
sure-
and
this
would
only
be
for
using
like
a
dash
all
flag-
that's
my
fault!
B
E
B
Yes,
so
that
is
what
I
put
in
here,
I
did
say
so,
since
namespaces
are
cluster
resource
using
the
namespace
flag,
when
deleting
them
should
error,
this
flag
has
no
effect
on
cluster
resources
and
confuses
users.
We
believe
this
to
be
an
implementation
bug
that
should
be
fixed
now
and
then
I
had
an
open
question.
Should
this
apply
to
all
cluster
scoped
resources
or
just
name
spaces.
A
And
I
did
just
that
for
all
cluster
resources
and
it
got
shut
down
a
couple
years
back
with
a
backwards
compatibility
issue.
That's
why
we're
trying
to
rephrase
it
differently,
although
I
was
actually
thinking
about
throwing
an
error
at
a
user
that
he's
passing
both
all
and
a
namespace
against
a
non-mainspace
resource.
B
I
think
that
I
spent
five
minutes
gathering
examples
of
this
in
the
wild,
and
you
know
many
years
ago
there
weren't
as
many
people
using
kubernetes
as
there
are
now,
so
I
think
that
we
can
at
least
make
a
case
for
it
in
terms
of
data,
be
like
hey
like
yes,
this,
I
even
called
this
out
somewhere.
I
said
these
are
breaking
changes
that
shouldn't
be
taken
lightly.
Strip
stocks
and
applications
will
all
need
to
be
modified.
Putting
our
empathy,
hats
on,
though,
we
believe
that
the
benefits
and
protection
to
users
are
worth
the
hassle.
B
D
And
and
so
there'll
be
no
interaction
with,
say
finalizers,
so,
like
you
know,
finalizers
are
used
quite
often
to
ensure
that
you
know
something
doesn't
get
deleted
accidentally
or
or
actually
even
right.
Now,
there's,
I
think,
maybe
manche
knows
about
how
there's
trying
to
create
an
annotation
for
like
do
do
not
remove
accidentally
or
something,
and
so
so
I
won't
deal
with
any
of
that
right,
it'll,
just
yeah.
The
final
results
will
keep
working
the
way
that
they
work.
A
A
They
will
explicitly
ask
you
to
type
yes
or
the
resource
name
you're,
trying
to
remove
that's
the
case,
for
if
you
want
to
delete
a
repo
in
docker
or
quay,
they
will
ask
you
to
explicitly
type
the
name
of
the
reaper
you're
trying
to
remove.
A
A
So
I
was
originally
thinking
that
we
would
go
the
path
of
if
you
try
to
pass
all
and
the
dash
n
against
the
name,
space
resource
cluster
scope,
resource.
Sorry,
we
would
be
putting
a
warning
of
you're
combining
something
that
we
know
is
a
foot
gun.
E
A
D
B
Okay,
I'll
check,
if
that's
the
issue,
there
are
two
others
we're
out
of
time,
so
feel
free
to
jump.
If
folks
have
to
go,
there
are
two
other
things
I
wanted
to
call
out.
So
I
think
that
all
flag,
I
think,
the
all
flag.
We
should
rename
to
all
resources.
I
think
that's
really
confusing
to
users,
so
we
can
do
that
over
three
releases
as
well
and
then
force
right
now
is
already
used
by
ignore
finalizers
delete
stuck
resources,
so
we
can't
use
force
unless
we
change
what
force
does
as
well.
B
A
Maybe
just
confirm
I've
seen
commands
where,
if
you
want
to
redo
them
and
they
have
some
resources
created
and
in
the
fact
they
would
say
that
something
happened
or
we
run
before
if
you
want
to
override
whatever
you're
doing
you
pass
confirm.
B
A
B
Think,
if
we're
good,
we
we
should
match
these
all
together.
So
if
we're
going
to
make
these
changes,
we
should
be
like
hey
like
this.
Is
the
time
we're
going
to
make
the
breaking
changes
around
delete?
Everyone
needs
to
read
these
change
logs
blog
posts
to
understand
them,
so
I
I
feel
like
we
can
get
away
with
a
batch
of
changes.
A
E
I
have
seen
over
the
years
a
ton
of
confusion
about
the
word
force,
meaning
that
we
just
delete
the
representation
in
the
ncd,
essentially
like
everyone.
First
approaching
that
thinks
that
it's
going
to
somehow
forcefully
remove
the
running
container.
They
don't
think
that
they're
leaving
garbage
around
in
their
cluster.
A
B
A
Have
a
look
at
the
at
the
dock
I'll
commend
there.
We
can
it's
good
to
start
the
discussion,
it's
not
like
something
that
will
be
very
easily
taken.
This
will
be
breaking
people
a
lot
of
people
and
it
has
to
go
through
a
proper
due
diligence.
I'm
inclined
to
even
send
it
over
to
architecture.
B
A
Yeah
or
maybe
just
pop
up
during
the
next,
their
next
call
and
bring
this
topic
over
to
their
attention
and
then
just
outline
the
issues
and
what
what
the
plan
is
and
how
we're
we're
hoping
to
roll
this
out,
because
that's
basically,
what
is
it
about
as
long
as
the
role
procedure
we
come
up
with,
will
be
long
enough
and
we
will
be
sufficiently
vocal
enough
about
those
changes.
B
Yeah
cool
well
thanks
everyone
for
staying
late
and
talking
through
this,
give
the
doc
a
read,
please
and
feel
free
to
comment
and
all
that
jazz.
I
was
still
I'd
like
to
send
this
out
an
email
today
or
tomorrow
to
get
started.
I'd
really
like
to
get
some
of
these
into
122,
whether
or
not
we
can
get
all
of
them,
but
the
sooner
we
even
like
the
force
confirm
flag.
If
we
can
squeeze
that
in
but
yeah
I
I
just
I
see
so
many
issues
of
people
just
deleting
their
and
there's
stuff.
A
I
mean
even
if
we
would
just
throw
a
warning
right
away,
even
without
doing
anything
it
might.
Actually.
I
don't
know
somehow
they
could
react
faster
than
only
after
somebody
will
pick
up
the
call
and
pick
up
the
phone
and
call
them.
Oh,
you
actually
removed
everything
because
they
are.
They
were
not
even
aware
that
this
got
deleted,
or
maybe
maybe
we
could
put
a
sleeve
there
for
10
seconds
so
that
they
could
quickly
press
ctrl,
ctrl
c.
E
A
B
A
I
just
it
just
popped
up
in
my
head.
I
think
one
of
the
email
providers
has
something
like
that
that
you
can
send
an
email
and
then
hit
stop
or
cancel
for
some
period
of
time.
I
think
gmail
did
that
for
someone
else
for
some
time.