►
From YouTube: 2020-04-16-Node.js N-API Team meeting
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
B
Yes,
so
our
talk
submission
was
accepted.
Yay
yeah.
I
was
quite
happy
to
see
that
so
are
you
going
to
share
your
screen
michael.
A
Yeah,
sorry,
all
right,
let
me
do
that
there
we
go
sure.
Okay,
there
we
go.
B
Okay,
perfect
so
officially
the
the
two
presenters
are
are
kevin
and
me,
and
so
I
just
wanted
to
just
open
it.
Up.
Nicola
and
gabriel
had
also
indicated
interest
in
participating,
and
I
just
wanted
to
verify
that
if
they,
if
you
guys,
were
still
up
for
for
participating
and
then
I
feel
what
I
should
do
is
contact
the
the
organizers
just
to
let
them
know
what
we're
thinking,
instead
of
just
like
dumping,
a
video
on
them
at
the
last
minute,
but
so
so
nicola
and
gabriel.
A
B
Here
we
go
yep
so
gabriel
and
nicola.
Are
you?
Are
you
guys
up
for
this
sure
yeah?
Okay,
okay,
perfect?
So
I
will
I
will.
I
will
contact
the
the
organizer
just
let's
give
them
a
heads
up
and
apparently
it's
gonna.
We
we
need
to
pre-record
it
michael
if
you
scroll
down
to
the
bottom.
I
I
I
put
a
link
to
the
to
the
notes
for
the
speakers
yeah.
B
So
this
gives
this
is
pretty
much
all
of
the
information
that
I
have
so
essentially
we
need
to
do
a
pre-recorded
session
and,
as
we
see
here,
it's
due
on
may
19th.
B
B
A
And
I
think
yeah
like
you
might
just
send
them.
You
know,
maybe
the
hey,
we're
gonna
have
more
people,
people
from
the
team
participate,
so,
for
example,
if,
like
later
on,
you
find
kevin,
wants
to
do
a
section
or
something.
A
A
C
Sorry,
my
goal
there
is
kevin.
Oh
thanks.
C
C
D
A
A
B
A
B
D
G
It's
it's
it.
It
may
or
may
not
be
that
complicated
like
I
have.
I
have
like
this.
This
program
called,
I
think
it's
called
the
avi
dmox
or
something
like
that.
It's
like
an
open
source
program
where
you
can
just
basically
like
just
copy
and
paste
snippets
and
and
and
you
can
just
basically
concatenate
them-
one
thing
that
we
want
to
make
sure
if
we,
if
we're
going
to
be
concatenating
videos,
is
that
that
they
all
have
like
the
same
resolution.
G
They
have
the
same
kind
of
encoding
and
bitrate,
and
all
that
so
so
as
to
make
sure
that
there's
not
like
like
this,
like
flash
in
the
middle
of
the
presentation,
you
know
when
it
switches
resolutions
or
something
so
so
that
would
be.
That
would
be
something
that
we
need
to
line
up,
which
we
don't
have
to
do.
If
we're
all
streaming
it
into
a
recording
at
a
given.
A
D
D
A
A
A
A
D
B
F
Okay,
the
the
first
step
is
to
prepare
the
slides.
Is
it
right,
correct,
yeah,.
B
F
I
my
idea
is
to
reuse
in
part
the
slides
that
we
already
used
in
montreal
and
we
changed
the
contents.
Okay,
based
on
yeah
all
the
news
that
we
introduced
at
in
last
year,
so
just
to
reuse
all
the.
F
Scheme
that
we
we
have,
if
it's
okay
for
you.
B
Oh
yes,
one
one
thing
the
organizer
said
is
they
will
be
supplying
a
template,
so
I
can
get
that
template
to
you
nicola.
As
soon
as
I
get
it,
okay,
yeah
and
then
we
can
see,
incorporate
that
template
in
into
the
slideshow
that
you
already
have.
B
No,
I
think
we've
got
a
good
start
I'll
get
it
get
in
touch
with
the
organizer
and
then
and
then
we
can
proceed
so
the
one
limitation
I
have
is.
I
am
not
going
to
be
able
to
make
this
meeting
for
the
next
two
weeks,
but
I'll
I'll
post.
Whatever
information
I
have
and
then
I
I
feel
we'll
still
have
enough
time
after
I
get
back
in
early
may,
to
to
get
the
whole
thing
done
by
the
19th.
B
Let's
see,
I'd
have
to
look
at
my
calendar,
so
I'll
be
gone.
Yes,
exactly
the
last
two
fridays
in
april
I'll
be
gone.
Okay,
but
I'll
be
back
on
the
7th.
A
A
D
B
D
But
so
I
tried
to
make
mine
as
some
notes
there.
It's
just
in
markdown
format.
It
probably
it
would
be
better
to
have
it
in
the
document,
but
I'm
having
I
I
just
couldn't
get
it
in
there.
D
Yeah
work
around
right.
The
question
that
I
had
then
mainly
was
we
said
on
the
approach.
What's
changed,
we
had
something
listed
for
test
folder,
I'm
not
sure
what
that
is.
G
That's
the
the
node.js
core
tests.
They
we
had
some
macros
in
there
that
that
had
that
started
with
the
abbreviation
and
api
underscore
it's
usually
most
of
them
are
an
api
underscore
call
which
is
basically
a
macro
that
wraps
handling
the
return
status
and
producing
an
appropriate.
I
think
it's
a
exception.
If,
if
the
status
is
not
okay,
that's
probably
what
it
refers
to,
and
so
we
had
like.
G
We
have
an
api
call,
an
api
assert,
an
api
call
return
void
these
macros,
they
all
change
to
node,
api,
node,
api,
node,
api
call,
node
api
call,
return,
void,
etc.
That's
probably
what
it
refers
to
okay,.
D
Let
me
try
a
different
browser
as
well.
Yeah.
A
A
If
not,
let's
relate
it
to
that
is
remember
the
rename
I
think
jim.
There
was
just
a
few
things.
You
were
still.
B
Yeah,
if
you
okay,
that's
something
yeah.
If
you
scroll
back
a
little
bit
michael,
I
was
working
on
updating
the
documentation
for
the
build
tools
we're
in
good
shape.
Okay,
I've
got
pr,
I
got
four
pr
submitted,
got
three
of
them
landed
and
so
the
last
one,
so
I
feel
yeah,
I
think,
we're
in
good
shape.
I
did
notice.
I
put
a
link
on
this
guy
right.
I
A
B
B
A
A
lot
and
everybody
else
too,
but
that's
that's
great
good,
and
so
really
we
put
out
the
blog
post
and
we
can
put
that
sort
of
as
a
we've
done.
What
we're
gonna
do
on
that
yeah.
A
Okay,
excellent,
okay,
stale
issues,
so
I
didn't
add
any
to
stale.
But
let's
take
a
quick
look.
Nobody
else
is
at
any
so,
okay,
nothing
to
talk
about
there.
A
A
I
know
last
time
we
saw
it's
kind
of
a
little
bit
of
a
downward
thing,
but
that
seemed
to
be
the
same
as
with
other
things.
Now
we
seem
to
see
it
being
going
back
up,
so
I
don't
know,
what's
what's
going
on
that,
I
don't
know
that
that's
necessarily
significant
or
not,
but
nothing
really
new
there.
So,
okay.
A
A
It's
just
been
yeah
it's
getting
on
a
year
so,
like,
I
think,
we'll
just
admit
that
we
would
have
liked
to,
but
we
got
other
things
to
spend
our
note
api
time
on
and
enabled
debug
testing
for
add-ons.
I
think
that
one's
still
worthwhile
as
a
reminder,
but
haven't
done
anything
on
that
crane,
the
matrix
of
tested
methods
so
jack.
I
know
you've
been
kind
of
working
on
that
front.
Anything
you
want
to
talk
about
this
week
need
help
on
anything
how's
that
going.
H
Yeah
so
like
this
week,
I've
been
just
working
on
like
adding
some
tests
to
the
simple
class
and,
like
kevin,
brought
up
like
an
issue
with
like
the
well-known
function
so
like
it
doesn't
like
when
you
register
like
a
symbol
from
like
either
js
side
or
like
the
c
plus
website
like
using
like
symbol.4.
It
doesn't
actually
like
what
you
call.
We
all
know
on
that
symbol.
It
doesn't
actually
return
the
symbol.
Just
returns,
undefined
guess
like
yeah.
So
I
just
wonder
if
I
can
like
get
some
help
with
that.
D
A
A
G
G
Header
for
well
known-
and
it
is
documented
in
the
header
as
get
a
public
symbol,
eg
symbol,
dot,
iterator,
so
so,
okay,
so
symbol,
dot,
iterator,
okay,
so
then,
okay,
so
the
symbol,
class
or
yeah.
The
symbol
class
seems
to
have
some
static
properties
that
are
sort
of
documented
in
the
spec
it
looks
like
and
so
those
proper.
So
then,
yes,
it
is
a
plain
old
property
accessor,
but
you
have
to
choose
the
name
of
the
property
varied.
G
Yeah,
I
don't
even
know
why
we
need
this
because,
like
you
know,
accessing
static
properties
is
pretty
simple.
Like
it's
purely
a
convenience
method,
it
looks
like
and
you're
right
it
may
be.
It
may
be
confusing
that
simple.
Like
this
well-known
you,
you
would
think
that
would
get
like
the
dot
four
behavior,
but
you
don't
so.
A
D
Oh,
what
okay
yeah,
so
I'm
trying
what
is
an
actual?
What's
an
example
of
a
a
well-known
symbol.
G
Yeah
symbol.iterator
is
this
is
the
example
given
in
the
in
the
comment
in
the
an
api.h
file.
G
Okay,
and
and
that's
that's
what
it
does-
that's
that's
what
the
test
does.
Is
it
retrieves
symbol,
dot,
iterator
and
it
checks
that
the
result
checks
out
the
result
that
you
use
it
to
get
like
the
next
thing,
because
it
is
an
iterator.
A
A
G
G
Well,
there
are,
there
are
like
there's
like
there
have
to
be,
for
example,
you
can
you
can
attach
you
can
attach
like
at
a
symbol
called
instance
of
you
can
attach
like
a
function
to
a
class
where
you
can
customize
whether
it
considers
an
instance
whether
it
considers
another
javascript
object
to
be
an
instance
of
this
class
or
not.
So
you
can.
G
You
can
customize
the
behavior
of
the
instance
of
operator
by
by
attaching
a
method
to
class
at
a
certain
symbol,
and
that
symbol
needs
to
be
retrievable,
because
it
has
to
be
the
same
for
every
class
and
so
yeah.
So,
for
example,
dot
has
instance
yeah
they're
there.
That's
an
example.
Yes,
he
has
instance.
So
so
you
so
you
retrieve
the
has
instance:
property
from
the
symbol
object.
G
The
resulting
value
is
a
symbol,
and
then
you
can
use
that
resulting
value
as
the
key
to
set
a
property
on
another
class,
and
the
value
of
that
property
will
be
a
function
which
which
the
engine
will
call
to
determine
whether
any
given
object
is
an
is
an
instance
of
that
class
or
not.
Instead
of
performing
the
default
instance
of
operator
action,
so
right
so.
D
A
H
A
C
G
A
A
So
that
answer
the
question.
A
J
Yeah
heads
up
on
the
nautical
ripple
recently
there
is
a
changes
regarding
the
primordials
in
the
non-core
and
the
weak
reference
and
registry
things,
and
that
caused
a
cost.
A
crash
regarding
regarding
an
issue
we
previously
faced
on
the
second
pass
callback,
and
there
is
a
issue
landed
on
the
master
to
unbreak
dci.
E
J
J
J
We
have
still
to
land
the
fix
to
prevent
use
of
the
free,
but
that
requires
an
another
review
from
gabriel,
but
I've
seen
the
ci
is
all
green.
So
the
the
left
actions
to
the
issue
around
it
is
that,
should
we
revert
the
fixes,
the
temporal
fixes
around
the
finalization
and
land
the
this
one
or
we
just
left
it,
as
is.
J
A
J
Oh
yes,
and
and
further
gabriel's
previous
fix.
I
have
commented
on
it
that
since
the
let
me
find
the.
J
A
J
J
Yeah
the
check
the
persistent
is
weak
will
be
always
false
in
the
situation
of
second
pass.
Go
back
since
we
have
already
reset
the
persistent
in
the
first
pass,
go
back
regarding
the
v8
api
contract,
so
this
check
is
always
filled
and
always
failed,
and
the
fixes
around
the
provided
by
api
is
always
not
not
being
round.
G
You
know
I'm
kind
of
wondering
if
it
might
just
be
simpler
to
to
do
the
fix
that
we
were
going
to
do
originally
when
when
we,
when
we
added
this,
this
removing
the
weak
thing,
and
I
think
I
think
that's
what
that's
what
james
has
done
like
if
you,
if
you
look
at
the
comment
for
for
the
first
pr
regarding
this
issue,
one
of
the
things
that
we
were
going
to
do
as
an
alternative
is
to
to
just
just
have
like
an
extra
flag,
and
I
think
that's
what
james
introduced.
What.
J
J
Yes,
the
flag
will
be
still
be
use
of
the
free
write
since
we
have
already
collected
in
the
environment
teardown,
but
we
are
still
referencing
it
and
use
the
pointer
in
the
second
passcode.
So
it's
still
have
a
problem
on
the
use
of
three
okay,
all
right.
A
C
A
A
And
it's
it's
kind
of
like
adding
the
extra
level
of
indirection
lets
us
still
say:
oh,
this
doesn't
need
to
run
because
we've
done
it,
but
we've
deleted
everything
except
for
this
parameter
and
it
can.
We
can
set
the
param.
We
still
have
the
reference
to
the
parameter
when
we
do
that,
so
we
can
set
it
to
point
to
null,
so
the
callback
itself
can
say,
oh
well,
I
got
nothing
to
do
just
delete
this
thing
that
was
passed
to
me.
G
Okay,
all
right,
okay,
so
like
I'm
still
trying
to
figure
out
why
we
would
be
like.
Why
would
the
flag
alone
we
would
be
using
after
free
like?
Why
would
finalize
get
called
if,
if,
if
the,
if
the
instance
is
destroyed,.
J
G
Right,
yes,
yes,
so,
okay,
so
the
so
the
so
the
environment.
Tear
down
has
already
destroyed
the
object
and
then
the
second
pass
callback
happens.
Correct.
A
G
J
D
G
Oh,
oh,
that's,
okay!
Okay!
This
is
a
good
solution,
yeah,
okay!
So
so,
instead
of
instead
of
passing
the
reference
as
the
as
the
weak
callback
parameter,
you
pass
a
pointer
to
the
reference
as
a
weak
callback
parameter,
and
then
you
set
that
pointer
to
null
when
the
reference
gets
destroyed
and
then
you
delete
it.
Okay,
good
good!
G
J
So
that
causes
a
similar
problem.
So
we
have
to
fix
the
issue
around
the
non-buffer
finalization,
but
I
didn't
see
rising
uprising
crashes
regarding
one,
but
I
I
didn't
find
a
stable
reproduction,
but
I
just
randomly
run
into
the
crash,
while
testing
the
second
pass
go
back.
J
Yeah,
so,
while
thinking
about
the
design
of
the
finalization,
I
I
I'm
thinking
that,
if
the
promoting
the
finalization
and
the
environment
they
are
done
before
the
actual
environment,
another
non-core
environmental
line
is
a
reasonable
holistic
fix
around
the
finalization
issues
and
cannot
call
it
according
to
javascript
issues.
J
A
A
It's
a
couple
weeks
ago,
but
let's
this
one,
we
talked
about
that
this
is
the
back
port
of
eight
right.
A
A
D
A
A
J
J
Non-Add-On
api
api
interface
is
not
designed
for
the
is
not
well
designed
for
the
sea
purpose,
exceptions
disabled,
that
is
to
say,
when
exceptions
is
disabled.
There
is
no
waste,
no
means
to
observe
the
api.
If
the
api
is
successful
or
not
like
say
currently,
the
object
set
in
add-on
api
returning
either.
A
J
The
either
on
the
success
of
failure,
so
when
exceptions
is
disabled,
there
is
no
way
to
observe
if
the
operation
is
a
success
or
not.
So
the
issue
is
also
around
the
bracket
accessing
with
the
purpose
of
successes,
and
the
iterators
will
also
introduce
the
non-observable
issue
around
the
iterating
on
the
objective
keys.
Since
we
can,
we
can
proxy
the
enumerating
on
the
objective
on
keys
with
a
proxy
and
in
the
proxy
we
can
throw
arbitrary
javascript
exception,
so
that
will
cause
the.
J
That
would
make
it
very
easy
to
make
the
program
work
and
not
ex,
not
as
in
the
expectation
since,
when
the
exceptions
is
disabled,
we
have
to
check
every
step
of
the
no
diamond.
Api
call
is
success,
but
there
is
no
compiler
guarantees
that
we
do
the
check
and
that
will
cause
a
crash
since
the
every
node
add-on.
Api
call
is
wrapped
with
the
and
not
be
throwing
field
micro,
and
that
maker
would
cause
a
throw
on
already
pending
exception,
and
that
would
cause
the
program
to
crash.
J
Yeah,
I
I'm
thinking
about
the
issue
that
when
only
enable
those
bracket
accessing
and
iterators,
while
the
exception
is
enabled
so
that
we
can
safely
say
when
the
exception
is
disabled,
you
you
have
to
check
every
operation
on
the
calling
into
javascript
is
tested,
is
checked
that
no
exception
is
strong
or
you
have
to
handle
the.
G
But
like
if
you're
using
like,
if
using
like
a
like
a
square
bracket
operator
for
for
like
setting
a
property,
let's
say
right,
then
then
it
is
your
responsibility
to
check
exceptions
afterwards
anyway,
right
like
if
you,
if
you've
chosen
to
turn
off
exceptions,
then
you
have
to
say,
you
know,
is
pending
exception.
J
Yeah
for
the
iterator
iterator
case,
that
means
for
each
iterating
lexi.
The
first
statement
in
the
fourth
for
loop
will
have.
We
will
have
to
check
the
pending.
J
That
that
that
seems
very
intuitive
and
unreasonable
to
me
and
that
delivers
it
will
be
hard
to
maintain
such
a
addon.
To,
I
think,
to
me.
A
Maybe
we
only
have
a
few
minutes
left
and
I
think
we
should
talk
about
one
other
issue,
maybe
open
an
issue
to
us
to
like
put
some
of
the
discussion
and
suggestions
in,
I
think,
would
be
good
yeah,
because
just
the
one
other
one
I
thought
would
be
good
to
just
quickly
touch
on
is
the
one
we're
trying
to
get
done
before
we
do
a
release
right.
This.
D
A
D
I
I
was
thinking
about
it
and
I
I
don't
know
how,
because
that.
Second,
to
last
step
there,
the
native
thread
finishes,
sleeping
like
the
the
native
thread
is
basically
just
using
the
thread
say
function
at
will
and
we're
expecting
it
to
eventually
get
a
closing
call,
because
it
was
aborted
and
I
don't
know
how
to
make
some
sort
of
synchronization
there
without
relying
on.
D
Because
it
seems
like
the
set,
if
the
set
immediate
call
is
never
executing,
then
we
would
not
be
able
to
do
synchronization
through
javascript
or
something
I
don't
know
like.
The
only
thing
I
can
think
of
is
naively
just
like
keep
trying
to
make
calls,
but
it's
not
a
real
fix.
You
know
like.
Is
there
a
timeout
somewhere
or
it
it's
it's
not
a
time
out.
It's
just
that
the
native
thread
has
10
number
10.
It
makes
10
calls
right.
B
D
G
G
It
is
like
an
intentional
race
condition
right
to
make
sure
that
that
that
you
know
you
can
abort,
because,
obviously
you
you
know
you
there
has
to
be
a
race
condition
in
order
for
for
for
the
abort
to
work,
because
this
threat
just
keeps
doing
what
it's
doing
right,
but
yeah.
There
is
a
possibility
that
the
set
immediate
won't
happen,
which
is
why
the
threat
should
have
plenty
of
time.
G
It
should
be
super
super
busy
for
a
very
long
time
while
at
the
same
time
progressing
because
you
can't
hang
it
for
indefinite
time,
because
then,
if
the
set
immediate
does
happen,
the
threat
is
hung
right
so,
but
that
it
has
to
be
kind
of
a
race,
but
why?
Why
wouldn't
the
set
immediate?
Never
happen.
That's
what
I
don't
understand
like
I.
It
doesn't
happen
in
1.8
seconds.
I
mean
what
is
keeping
the
event
loop.
So
busy
that
it
never
causes
that
immediate,
that's
what
I
don't
get.
A
Sorry
I'll
make
make
kevin.
A
D
Okay,
so,
okay,
let
me
share
my
screen.
I
don't
think
I've
ever
given
zoom
access
to
share,
so
it
may
system
preferences.
But
yes,
so
we
can
kind
of
continue
discussing,
like
I
don't
know,
like
1.8
seconds,
seems
like
a
long
time.
G
Yeah
yeah
plenty
of
time
to
get
to
get
an
immediate
call
back
through
yeah.
So
that's
weird
that
it's
not
happening
but
yeah,
I
I
don't
know
I
mean
you
you'd
have
to
debug
that
I
guess
there's
there's
just
no
other
way.
G
An
alternative
that
you
might
want
to
consider
like
temporarily,
is
to
just
just
keep
producing
numbers
instead
of
just
quitting
after
10..
Just
ignore
the
fact
that
you're
you're
only
supposed
to
produce
10
numbers
and
just
keep
looping
see
if
it
eventually
comes
through.
G
You
know
see
if
it's
like
a
like
a
systemic
issue
or
if
it
really
is
just
a
matter
of
timing,
because
I,
if
I
remember
correctly,
I
said
that
I
I
set
that
number
as
a
number
defined
at
the
top
of
the
file
right
and
I,
and
I
also
export
that
number
to
javascript
right
as
a
as
a
constant
on
the
on
the
on
the
exports
right
just
so
that
so
that
you,
you
don't
have
to
specify
it
once
in
javascript
and
once
in
c,
plus,
plus
right,
and
so
what?
G
If
you
raise
that
number
to
like
100,
you
know
then
yeah.
You
know
that
might
be
you
know,
it
might
be
enough,
you
know,
but
it's
I
know
it's
arbitrary
and
I
I
don't.
I
couldn't
think
of
a
better
solution
than
than
to
just
kind
of
keep
the
thread
busy,
while
while
the
main
thread
does
the
set
immediate
so
as
to
abort.
B
G
I
mean
you
know,
you
know
there
there
is.
There
may
be
another
solution
right,
which
is
to
use
to
use
an
additional
mutex
right
and
it's
kind
of
cheating.
If
you
think
about
it
right,
because
you
you
create
an
additional
mute,
you
basically
create
a
second
synchronization
mechanism
right
and
then
you
you
make
the
thread.
Wait
on
that
new
text
from
the
main
thread
you
set,
the
mutex
after
you've
issued
the
abort
right
and
then
the
thread
wakes
up
and
oh
look.
Lo
and
behold,
it's
closing
right.
G
So
so
then
you
know
what
I
mean
like
you
use
a
second
synchronization
mechanism
right,
because
the
the
idea
is
right.
The
idea
I
mean
the
purpose
of
the
test
is
to
see
whether
whether
the
the
fact
that
the
threat
say
function
is
closing
can
successfully
be
conveyed
to
the
threat
right
if
it
can
well
right-
and
you
have
proven
that
it
has
successfully
been
conveyed
to
the
threat
right.
The
fact
that
you
used
an
additional
synchronization
mechanism
is
arguably
not
part
of
the
test
right.
G
G
Yeah,
because
the
thing
is
the
thing
is,
it
does
in
a
way
test.
The
fact
that
that
triggering
the
abort
has
immediate
effect
right
because
triggering
the
abort
should
be
immediately
like
atomically,
visible
on
every
thread
right.
So,
if
you
trigger
the
abort
and
in
the
next
instruction
the
thread
reads
it.
It
should
be
there
right.
G
G
It
just
feels
it
just
kind
of
feels
like
cheating
like,
like
you
know
the
the
solution
we
have
now
feels
more
natural,
because
that's
how
it
would
happen
like
in
real
life
is
the
threat
just
keeps
running
running
running
all
of
a
sudden.
It
gets
a
closing
and
okay.
What
does
it
do
then
right
exactly,
but,
but
I
don't
think
we
can.
G
We
can
replicate
the
real-life
conditions
for
the
purposes
of
what
is
to
be
a
very
short
and
finite
test
right,
and
so
you
know
where,
where
where
the
timings
are,
are
very
crucial
right,
because
I
mean
the
idea
is
that
in
real
life
the
threat
runs
forever
unless
it
gets
a
closing
right,
because
it's
it's
a
stream,
it
keeps
producing
data
whatever
right.
So
there's
there's
good
reason
why
why
it's
running
forever
right?
But
we
can't
replicate
that
in
the
test,
we
can't
leave
open
the
possibility
that
it's
gonna
run.
G
D
D
Because
I
haven't
been
able
to,
it
was
the
first
time
to
do
it
and
you
have
all
the
permissions
and
what
yeah.
D
All
right
so
then
does
anybody
else
have
anything
they
would
like
to
say
before
we
end.