►
From YouTube: 2022-10-28-Node.js Node-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
A
I
think
Jack
Jack
told
us
he
wasn't
gonna
be
able
to
make
it.
We
did
did
I.
Think
I
landed
one
of
his
PRS
from
last
week
and
there's
still
one
that
he's
working
on
on
Kavanaugh
but
I.
Don't
think
I
have
any
other
updates
unless
anybody
else
has
on
that.
One
yep
in
terms
of
better
coverage
on
the
Node
AP
on
YouTube
I
I,
don't
know
that
I
don't
have
any
updates.
A
A
C
B
Okay
and
Gabriel
did
this
presentation
so
I
think
at
least
if
we
have
something
existing
on,
the
list
will
be
still
good.
B
A
B
Is
still
still
kind
of
up
to
date,
yeah,
so
let
me
yeah
that
is.
A
B
I
think
it
was
another
one
with
Nicola
and
someone
else.
They
were
doing
some
kind
of
Workshop.
A
Okay!
Now
this
is
still.
A
A
B
B
A
A
Okay,
I
think
that's,
maybe
the
the
the
low-hanging
fruit
anyway.
A
A
Yeah,
no
and
then
we
can
build
on
to
the
yeah
some
some
directly
onto
that.
It
sounds
good.
Okay,
anything
more
on
that
one.
Before
we
move
on.
B
A
A
A
A
I
created
this
PR,
which
listening
to
cast
thanks
for
for
taking
a
look
and
finding
my
Gap
and
your
suggestion
of
adding
a
test
for
the
the
positive
case
is
good.
I
was
thinking.
Maybe
I'll
just
add
that
Define
to
one
of
the
existing
tests
that
doesn't
use
the
external
buffers
that
that
makes
sense
to
you
as
well.
C
A
C
A
Okay,
okay,
so
yeah
I'm
gonna
do
that
there
was
a
suggestion,
so
it
sounds
like
just
what's
in
there
is
enough
and
and
I
think
it'd
be
good
to
keep
it
in
two
separate
PR's.
But
the
next
question
is
there
was
a
suggestion
of
adding
another
method
which
I'm
thinking
seems
like
a
reasonable
next
step
and
I
wondered
what
what
you
two
think
about
that
you
know
adding.
That
is
another
API
which
would
be
like
in
in
the
latest
node
API
version,
foreign.
C
I
think
it
would
be
easier
for
us
to
and
for
the
add-on
Waters
to
having
a
helper
function
which
may
exist
in
their
own
library
or
in
node
I,
don't
API.
Since
we
we
can
achieve
the
same
thing
with
the
existing
node
API
and
I.
Don't
have
to
wait
for
the
API
to
be
released
in
the
full
circle
of
not
a
player
version.
A
A
C
Yeah
right,
if
I
build,
the
help
function
can
be
distributed
for
the
photon,
and
you
know
I
don't
want
to
control
the
right
course.
If
they
didn't
like
the
copying,
then
that's
fine
and.
A
C
B
New
API
to
node
API
itself
right.
C
B
Yeah
I
agree
with
you,
because
I
think
one
of
our
guidance
we're
not
supposed
to
add
API,
which
are
convenience
apis.
We,
if
we
can
achieve
this
functionality
for
existing
appearance,
we
should
not
add
new
right.
A
And
it
really
is
basically
make
one
call
the
yeah
make
one
call
and
then
call
the
other
one.
B
Right,
it
can
be
added
to
our
C
plus
plus
node
API
load,
a
dome
API
right.
A
A
A
Yeah
and
I
think
it
actually
would
be
quite
hard
to
have
them
there
because,
like
what,
if
you
have
two
experimental
ones-
and
you
put
them
in
a
certain
order
and
then
you
actually
land
the
second
one,
but
not
the
first
one
it'll
get
a
new
number.
A
A
You
know
it's
good
to
have
us
all
think
about,
but
I
that's
I
was
thinking
that
that,
wouldn't
necessarily
it
might
be
sember
minor
like
this.
This
PR
could
be
semper
minor,
because
it's
adding
a
new
some
new
options
that
you
can't
use.
C
A
I
guess
I'm
thinking
of
the
case,
so
you
compile
against,
say
you
compile
against
a
newer
version
of
node,
where
you
can
check
that
if
you
run
that
on
an
older
version,
you're
never
going
to
get
that
error
code,
but
it
it's
not
going
to
give
you
an
error.
If
you
compile
with
that
with
you
I,
don't
think
you're
going
to
get
an
error
you're
just
never
going
to
get
that
error
code
back,
so
things
will
still
run.
A
C
Er
I
don't
have
a
strong
opinion.
Okay,
this
yeah.
A
B
Oh
I
I
just
thought
just
a
couple
of
things.
If
we
can
go
to
this
pull
request,
sure
sorry,
I
didn't
see
this
pull
request
before.
B
Okay,
if
you
go
to
files
change
the,
if
you
look
on
the
you
have
added
new
error
code
in
API
types.
B
So
you
see
like
there
is
a
comment
below
says
like
if
we
add
new
suffix
there,
we
need
to
also
update
the
function,
content
and
I,
don't
see
you
change.
This
function,
get
lost.
A
Okay,
yeah
I
didn't
see
that
comment
below
got
it.
Oh
I
will
do
that.
Can
you
just
well
I
can
make
here
I'll.
Add
a
comment.
Yeah
from
team
meeting.
C
B
And
I
was
also
looking
this
I'm,
not
sure
if
macro
is,
is
correct.
Marker
we
severe
compressed
pointers
and
shortcode.
A
I
want
to
take
a
look
and-
and
you
have
some
way
to
be
more
sure
what
I
did
is
like
we
have
a
team
somebody
working
on
our
V8
team,
who's,
implementing
pointer
cage
for
powerpc
and
s390,
and
that's
what
they
told
me
was
the
right
one.
B
Did
you
see
that
the
because
thank
you
today,
electron
team?
They
did
this
patch
inside
of
their
code
and
I
was
at
the
casino
what
they
were
doing
there.
Okay,
I
I
can't
remember
out
of
my
head
right
now
immediately,
but
I
may
actually
look
and
yeah.
A
If
you
can
find
that
one
and
like
comment
in
here
and
say
well,
hey
they
used
this
one.
That
would
be
yeah
because
I
don't
have
a
great
way
to
test
that.
Although
sorry,
what
I
I
did
do
so,
let's
see
I
Define.
That.
A
And
I
ran
some
testing
okay,
but
but
yeah.
If
you,
if
you
can,
if
you
can
dig
up
that
electron
patch
and
we
could
see
what
they
used
and
checked
because,
like
that
would
be
another
another
way
to
validate,
because
yeah
I'm
I
did
what
I
I
could
to
figure
out.
If
this
is
the
right
one,
but
being
more
sure
would
be
better.
It's
kind
of
what
I'm
saying.
A
A
A
Complete
yeah,
it
would
be
nice
if
we
could
see
hey
yeah.
They
used
the
exact
same
one,
because
there
was
that
I
found
three
that
had
cage
in
them.
There
was
one
was
like
caged
Heap.
There
was
this
one,
and
then
there
was
compressed
pointers
in
isolate
cage
and
the
explanation
I
got
was
like
the
isolate
cage
was
the
case
where
you've
got
compressed
pointers
on
but
you're
not
like.
We
wouldn't
run
into
this
case.
It
was
just
the
shared
cage
was
the
one
that
would
cause
it,
but
yeah.
A
If
you
can
find
that
and
tag
that
on
there
I'll
take
a
look
at
it
and
see
if
it
confirms
or
denies
that
we've
got
the
right
things.
I.
B
Was
looking
for
this
V8
and
I
was
seeing
the
changes
marker
at
some
point
yeah
they
were
changing
like
as
they
implemented
as
Cajun
creatures,
I've
seen
that
they
they
have
like
for
the
last
years
of
change,
in
name
of
the
kind
of
macro,
okay,
I'm
not
completely
sure
about,
but
I
I
just
know
like
a
PR
for
an
electron
code.
Has
they
have
a
little
bit
more
changes?
Okay,.
A
Yep
no
yeah,
if
you
can
tag
that
on
there
and
I'll,
take
a
look:
okay,
okay,
yeah
thanks
for
everybody's.
Looking
at
that
anything
else
on
this
one
we
want
to
chat
about
so
like
yeah.
We
won't
necessarily
rush
to
put
to
create
a
new
method.
Maybe
an
example
somewhere
would
be
good
and
I
guess
chensong.
Should
we
do
do
it
in
node
add-on,
API
itself
like
there,
we
could
actually
Implement
something,
although
maybe
that
can't
really
be
implemented
until
we
go
back
until
this
gets
back.
Ported.
C
A
Yeah
I
think
we
may
we
could
work
around
that
in
the
test.
The
question
is:
should
we
or
not,
but
okay-
well,
I'll,
think
more
about
that
I
think
I'll
spend
my
energy
just
getting
this
landed
first
and
then
we
can
think
about
next
steps.
B
B
I'm
sorry
go
ahead,
I
think
it's
exactly
what
the
developer
was
asking
at
least
it's
the
way,
I
read
it
or
they
will
ask
him
to
about
changing
so
I
know
the
adult
API
to
have
this
message
there
right.
C
A
A
Yeah
and
that's
because
the
question
is
like-
does
electron?
C
C
C
A
Okay,
okay
and
in
fact,
for
14
anyway,
if
we
get
a
real,
if
there's
a
real
concern,
we
could
just
back
Port
just
the
definition
of
the
status
code
and
that's
it.
A
A
Working
on
it
but
and
I
think.
A
C
This
is
a
rather
interesting
one
and
we
just
got
a
new
issue
in
the
node
core
2.
you
can
you
can
you
can
go
for
the
link
in
this
one
here:
okay,
yeah!
C
It's
about
the
clean
up
orders
So.
Currently
we
running
the
cleanup
hooks
added
by
the
node
API
first
and
then
we
are
disrupting
the
another
API
environment
and
the
federalization
callbacks
are
invoked
in
the
destruction
of
the
API
environment,
so
the
cleanup
hooks
are
invoked
first
before
the
the
other
add-on
callbacks
like
federalization
callbacks.
C
No
but
I
found
it
can
be
a
solution
to
another
problem
like
say
this:
existing
behaviors.
Okay,
that's
that's
a
Kevin's
linked
another
API
issue.
We
can
go
back
to
that
one
okay,
yeah
and
maybe-
and
not
not
this
one,
but.
A
This
one,
this
one.
C
A
C
C
Basically,
that
decided
functions
are
usually
a
member
of
the
object,
wrap
and
it
can
be
released
once
the
object
drop
is
finalized,
which
means
it
it
can
be
released
in
the
shutdown
children
process
of
the
environment,
of
which
triggers
the
finalization
Callback,
and
the
problem
here
is
that
the
third
shift
function
itself
can
also
be
released
with
the
finalization
go
back
triggered
by
the
shutdown
you
know.
We
have
also
received
complaints
about
this
Behavior.
Since
users,
users
might
still
have
reference
card
on
it,
but
never
mind.
C
The
problem
here
is
that
this
third
cell
function
here
can
be
released
twice.
One
is
from
the
no
that
I
don't
know
not
a
another
API
environment
and
one
from
this
object,
finalization
callback
and
if
the
if
the
Starship
function
is
released
in
the
API
environment.
First
then
the
problem
here
can
be
the
then
that's
the
problem.
What
what
the
problem?
Here
we
got
if
the
object
object,
wraps
finalization
go
back
is
wrong.
First,
then,
there
is
no
problem
at
all.
C
Either
it
it's
just
fine,
since
the
it's,
the
stress
function
is
released
in
the
user
land.
First
then,
no,
then
another
API
environment.
We
are
not
vendorize
it
again,
so
I
I
can
say.
The
core
problem
here
is
that
user
have
no
user
code
have
no
options
on
what
the?
What
is
the
penalized
correct?
What
what
users
expected
correct
corrected.
C
Expected
correct
order
of
this
objects
like
say:
object
of
the
save
function
right
and
but
but
the
fortunate
each
news
is
that
if
users
can
use
this,
if
you're
the
add-on
checks,
this
allocated
object
and
release
them
all
at
once
in
their
own
clean
up
hooks,
then
all
of
these
issues
can
go
can
can
be
gone.
You
know
they
can
decide
which
order
they
release.
This
object
to
this
clean
up,
Hooks
and
the
clamps
hooks
are
invoked
before
the
penalization
Callback
at
the
shutdown
of
the
northern
environment.
C
The
klingamp
hooks
on
the
environment,
okay,
you
see
the
cleanup
works
out
wrong
before
the
day.
No,
the
not
the
API
environment,
so
the
finalization
go
back
are
invoked
after
the
cleanup
hooks.
So
right,
okay
users
can
yeah,
they
can
release
these
objects
first
in
the
pulling
up
hooks
and
they
they.
If,
since
the
cleanup
hooks,
are
running
user
codes,
then
they
can
decide
what
orders
they
would
expect
to
correctly
release
their
own
objects.
C
Yes,
I'm
I'm,
thinking
that,
with
this
clean
up
hook
mechanism,
we
can
help
these
add-ons
to
release
this
objects
correctly.
You
you
know
we
do
this
in
another
Court
who,
since
since
V8,
doesn't
invoke
animal,
it
doesn't
show
any
generalization
callbacks
at
shutdown
for
the
weaker
callbacks
and
now
that
another
core
have
to
keep
track
of
allocated
object
and
release
them
or
otherwise
added
the
shutdown
yeah.
C
So
I
believe
that
might
be
a
a
more
visible
and
robust
approach
for
use
the
land
add-ons
to
do
the
same
thing
to
release
the
release,
their
data
and
the
objects
with
their
own
with
their
own
code.
Oh
go
back.
C
Yeah
no
I
mean
generalizers
are
great,
but
the
problem
is
that
the
GQ
problem
is
that
is
the
short
done.
We
don't
know.
C
The
note
chord
can
don't
know
the
correct
order
for
the
user
land
record
since
user
and
the
code
can
be
very
it's
it's
unpredictable
for
another
core
which
the
order
is,
since
you
can
release
the
testing
function.
First,
all
they
expect
the
object
to
be
released.
First.
C
C
Yeah
I'm
thinking
about
adding
a
new
section
about
this,
how
to
handle
how
to
handle
shutdown
of
the
API
environment
properly
in
another
API
document
section
and
adding
new
examples.
C
Let's
say
how
to
check
these
finalizers
and
the
search
function
actually
about
the
stress:
the
function
generation
at
the
shadow.
Since
we
previously
got
a
issue
about
the
complaints
about
the
function,
generalization
and
the
shot
at
the
moment
of
knowledge,
environment,
shutdown
and
I
can
I
can
find
it
with
a
link.
C
A
A
C
A
C
A
A
A
B
With
succession
not
code,
I
see
a
bit
lost,
it
is
pointing
at
the
back.
We.
C
A
B
A
A
B
It
sounds
like:
are
they
just
doing
the
normal
callback
like
say,
implementing
a
function
and.
A
A
Okay,
let
me
just
quickly
switch
over
to.
A
B
Sorry,
okay,
sorry
I
was
music
about
the
references
and
features
right.
A
Okay,
this,
let
me
see
ref,
you
know
the
number
offhand
or
this
one
here.
B
B
We
can
say
an
API
module
and
a
module
in
it
in
order
to
have
this
init
stuff,
I
have
to
put
them
inside
of
this
node
API,
and
what's
the
interesting
thing
that
I
found
that,
then
we
do
object
only
because
I
want
to
test
both
cases
with
all
behavior
and
new
Behavior
I
found
that
then
we
do
references
for
symbols
and
symbols
does
not
seem
to
have
a
big
reference
semantic.
So
I
have
this
test.
B
If
you
can
you
look,
can
you
click
on
this
last
C
file
like
this
one
reference
object,
only
yeah,
it's
all
the
way
down.
Oh
sorry,
it's
no!
It's!
It's
JavaScript
file,
sorry
test
test.
B
And
if
you
scroll
to
the
bottom
of
this
test.js
file,
see
over
over
here,
look
up
so
what
I'm
doing
like
I
pretty
much
always
reference
is
supposed
to
be,
although
I
have
rev
count
zero
and
supposed
to
be
garbage
collected,
but
what
I
found
that
pretty
much
all
of
them
supposed
to
return
now.
But
if
you
look
on
line,
96
I
have
to
put
special
exception
for
symbols
and
apparently
they're
not
collected,
and
it
means
that.
But
symbols
don't
really
have
in
our
existing
code.
B
The
ref
count
goes
to
zero,
even
we
initiating
all
these
different
things,
but
it
seems
like
inside
of
the
eight
JavaScript
engine.
The
symbols
do
not
have
decryption
cement,
so
it
means
that
even
we're
saying
yes,
it's
supposed
to
be
weak
but
in
fact
they're
not
collected
they.
A
B
I,
don't
know
I'm
running
just
see
a
few
times
like,
let's
trying
to
and
do
it
at
least
twice
over
here.
A
A
A
B
Me,
let
me
check
existing
tests
for
references
to
symbols.
I
want
to
understand
it's.
What
I
I
kind
of
stuff
to
add
this
test
today
morning
and
I
was
surprised
that
I
had
to
do
something
special
about
symbols.
A
B
A
B
No,
it's
not
a
problem,
it's
more
like
if
we
because
ideally
like,
if,
if
we're
saying
we're
going
to
zero
and
we're
saying
Speak
reference
and
big
reference
like
Chang
Jong
sent
before
we,
we
can
actually,
for
example,
we
don't
support
big
reference
functions.
We
can
just
simply
kill
the
object,
make
it
release
it
right,
yeah,
so
I
guess
my
my
question
like
what
is
correct.
Behavior
would
be
for
symbols.
Should
we
instead,
then
we
go
with
ref
count
to
zero.
B
I
know
that
terms
for
sure
does
not
right
and
also
it's
not
it's
not
it's
not
JavaScript
standard,
like
software
that
was
hoping
to
change
with
his
PR,
not
to
have
big
reference.
Symmetrical
symbols,
but
I
actually
see
that
V8
does
not
even
support
it.
So
we
effectively.
C
They
are
trying
to
edit
a
symbol
at
adding
symbols
to
the
to
to
be
able
to
be
added
as
a
weak
map
key.
So
you
know,
if
that's,
if
that's
possible,
then
quick,
quick
semantics
must
be
implemented
for
the
symbols.
B
Okay,
so
you
saved
it
because
it
can
be
used
as
a
key.
C
It's
definitely
not
allowing
them
to
be
a
pre-referenced
with
the
like,
say,
big
ref
or
another,
but
I'm
not
sure
I,
I,
I,
I
right
now,
I'm
not
sure
about
it.
Just
okay.
B
Anyway,
that's
what
it
would
be
great
if
like
Michael,
will,
have
a
look
on
his
PR,
but.