►
From YouTube: 2022-08-05-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
B
Yes,
I
I
tried
to
identify
which
of
these
bindings
has
been
used
to
to
implement
real
nativity
today
and
for
now
yeah.
Only
the
rust
flavor
here
has
been
used.
B
B
Is
in
in
in
neon,
if,
if
I
were
a
member
like
a
low
level,
binding
and
the
rest
are
only
experiment
until
until
now,
so.
B
The
best
approach,
so
I
can
open
a
pr
and.
B
We
can
have
a
an
issue
or
a
page
on
abyssable
node
label,
where
we
take
the
responsibility
to
update
the
the
list
and
in
the
core
documentation.
We
link
that
that
documentation
that
we
have
in
abyssable
node.
A
B
I
I
okay,
I
I
have
a
preference,
so
I
I
prefer
to
have
a
a
page
or
a
markdown
page
on,
for
example,
disabled
repo
and
then
update
the
the
list,
because
if
we
have
because
the
process
is
too
simple,
no
if
we
won't
update
the
list
on
abysmal
node
and
instead,
if
we
add
this
list
directly
on
the
official
node
documentation
sometime,
these
change
this
we
need
to
change.
B
These
leads
for
every,
not
the
js
version,
because
because
maybe
I
don't
know
some
bindings
that
today
is
used
a
lot,
maybe
in
one
two
years
will
be
replaced
from
another.
So
it's
not
avoid
optional
if
you
want
to
to
implement
a
native
add-on.
B
So
this
is
my
my
reason
to
to
avoid
to
put
the
list
directly
on
the
node
core
docs.
A
Paragraph
which
is
like
it
is
possible
to
develop
add-ons
in
other
languages
when
used
because
of
you
know:
node
ap,
node
api.
It
still
relies
on
the
same
api
as
in
here
and
then
here's
another
link
to
you
know
in
the
add-on
api
repo.
We
maintain
a
list
of
some
of
those,
those
things
right
and
it
might.
A
Rest,
one
might
be
worth
I
sort
of
kind
of
like
a
page
that
you
would
go
to
and,
like,
I
think,
even
mentioning
the
rust
one,
because
that
seems
like
a
very
active
one,
like
you
know,
saying
like,
for
example,
you
can
actually
develop
your
add-ons
and
rust
and
they
have
a
pretty
good
ecosystem,
but
go
off
to
this
other
page
to
get
the
details
of
all
the
other
options.
How
does
that
sound.
A
A
Sounds
good
thanks
for
researching
all
those
because
yeah.
I
think
it's
really
interesting
that
that
you
can
and
having
that
that
be
obvious
is
a
good
thing
better
coverage
for
node
api
on
youtube.
I
don't
know
if,
as
anybody
has
any
comments,
I
know
kevin
was
gonna,
take
was
gonna,
spend
some
time,
but
he's
been
out
for
a
while.
I
know-
and
I
think
nick
you
were
going
to
take
a
look
at
that
too.
So.
B
Yeah,
you
opened
the
an
issue
about.
D
A
B
Okay,
okay,
one!
So
I
I
try
to
produce
these
videos
and.
B
B
I
will
share
before
with
all
of
you,
maybe
in
issues,
and
then
we
probably
we
we
publish
on
the
social.
So
I
think
that
this
is
the
right
approach.
A
B
A
B
Oh
I
am
in
this
week.
I
will
think
at
what
kind
of
video
I
could
prepare
and
then
share
the
some
ideas
we
in
the
issue
that
we
have
and
then,
if
I
will
receive
the
okay
by
you
and
others,
I
I
start
working
on
these
ideas.
B
Okay,
so
yeah,
and
then
we
can
plan
the
the
release
of
the
video.
So
maybe
I
can
produce
more
than
one
video,
but
we
can
plan
to
release
maybe
one
a
week.
I
I
I
don't
know
so
like
you
would
just
yeah
yeah
that'd
be
great.
A
Okay
sounds
good:
okay,
drive
forward
towards
full
coverage.
I
think
jack.
You
were
excited,
yeah,
go
ahead.
D
Working
on
adding
some
test
coverage
for
type
thresh
functions,
I'm
just.
B
But
I
should
have
something
concrete
by
next
week.
A
Oh
great
sounds
good,
okay,
stale
issues,
so
let's
flip
over
to
these.
A
B
B
D
A
A
C
John
started
last
last
time
and
there
is
also
a
pull
request.
He
created.
C
D
D
D
It's
about
the
viet
api
contract
that
the
finalizer
or
gets
back
in
store
urban
capacity
store
can
be
invoked
in
another
style.
So
you
just
can't
have
to
create,
create
wrappers
around
the
original
fundamentals
so,
but
we
still
have
spaces
to
improve
the
condition
that
we
can
still
move
away
from
the
buffer
constructor
and
just
leave
the
wrappers
around
the
v8
already
first
packing
store
of
another,
but
I
didn't
have
into
the
details
of
this
problem.
So
it's
just
a
quick
overview
of
this
issue.
D
A
A
D
Okay
yeah,
this
pi
doesn't
change
any
api
contract.
It's
just
and
called
tiding
of
the
code
in
another
api
implementation
around
the
panel
like
this
second
plus
callback.
D
So
it's
basically
make
the
cancellation
of
the
second
pass
callback
possible.
So
we
can
greatly
simplify
the
the
greatly
simplifies
the
process
processing
around
the
in
cancelable.
Second,
pascal
pack,
right.
C
A
D
So,
basically,
not
no
tests
need
to
be
changed
and
all
tests
has
passed
and
there
is
a
little
problem
I
will
fix
it
later,
but
the
sale
has
been
passed
for
all
the
tests,
so
I
hope
people
if
people
have
time
can
you
take
a
look
at
this?
It's
about.
Hopefully
I
can.
The
new
information
can
help
people
to
understand
the
the
the
quick
the
quirks
around
the
finalizer
better,
since
it's
way
more
straightforward.
D
Yeah,
it's
and
we
used
to
have
many
tricks
around
the
second
pass
callback,
since
it's
not
cancelable
and
we
have
to
create
temporary
pointers
just
for
the
second
pass
callback
to
be
not
crashed
to
not
question
the
process.
D
Let's
see
yeah
it's
the
last
post,
the
point
the
this
person
made
is
that
not
prs
and
the
neon
completely
completely
prevent
the
user
from
calling
into
javascript
from
vending
analyzers.
D
So
and
in
that
sense
I
do
people
when
people
are
using
rpr's
and
the
new
in
this
language
wrappers
they
are
not
going
to
call
into
javascript
in
the
finalizers.
So
I
believe
this
is
one
possible
alternatives
for
vladimir's
work
around
the
analyzers
and
new
financial
types
to
prevent
users
from
going
to
javascript,
but
it
has
a
eager
implication
to
allow
people
to
get
their
pointers
freed
earthly
earlier.
D
So
that
made
me
to
think
that
the
alternative
and
a
new
alternative
to
the
current
approach
is
that
we
can
introduce
a
new
flag
so
that
people
can
still
use
current
api,
but
their
behavior
is
changed
with
the
flag
set
and
like
say
not.
Bias
and
in
young
can
just
settle
the
flag
and
they
don't
have
to
change
any
code
and
their
panelists
can
be
invoked
eagerly
and
in
a
more
timing
manner.
D
So
they
don't
have
to
change
any
code
just
to
set
a
flag
to
the
environment,
and
I
I
thought
this
would
be
more
adoptable
more
easier
to
adopt
for
this.
C
Are
you
talking
about
introducing
with
flag
for
the
whole
node.js
process.
D
C
So
I
remember
at
some
point
one
of
pr's:
I
was
trying
to
introduce
this
concept
of
features
like
remember,
like
a
different
things
which
we
can
set
up
pure
add-on.
C
It
was
not
applicable
to
that
case
because
I
was
trying
to
fix
a
factory
bug
around
get
all
properties.
We
didn't
have
this
configurable
filter
working
correctly,
but
maybe
yeah.
D
D
A
Then
the
feature
flag
makes
a
lot
of
sense
and
it's
probably
simpler
in
the
end
right
like
to
basically
say
it's
always
going
to
be
one
way
or
always
the
other
way
versus
individually.
Have
you
know,
be
able
to
have
some
that
are
one
way
or
some
that
are
the
other.
C
And
I
bet
like
in
99.9,
people
really
want
just
to
call
c
plus
plus
distractors,
like
they
don't
really
want
to
call
javascript
in
the
first
place.
It
would
be
nice
if
we
can
somehow,
like
maybe
like
a
newer
version
of
node
api,
so
say
if
you
have
a
certain
node
api
level,
nine
or
ten.
We
just
simply
use
this
non-javascript
coordinate
finalizers
as
default.
A
D
I'm
not
sure
how
that
would
work,
since
not
the
api
version
is
just
a
flag
to
hide
some,
not
api
functions.
It
just
doesn't
have
any
effects
on
how
the
api
behaves.
Since
that's
not
that's,
none
of
the
core
is
not.
You
cannot
get
that
version
from
the
bedroom.
C
The
way
I
think
it
might
work,
but
much
every
time
we
register
a
new
module.
In
fact
it's
a
macra
right
and
so
potentially
we
can,
depending
on
the
version
of
our
node
api,
we
can
decide
what
this
macro
expands
to
so
for
the
new,
depending
on
the
order
new,
because
thankfully
this
kind
of
feature
registration
should
happen
during
module
registration
as
a
part
of
this
macro
expansion
and
stuff.
So
we
can
introduce
defaults
there.
A
Yeah,
I
think
I
I
bet
you
we
could
figure
out
how
to
do
it.
The
question
is
like
from
the
abi
compatibility
story,
but
I
don't
think
like
I'd
wait
till
we
actually
have
the
feature
there
and
we're
you
know
and
like
I
wouldn't
want
us,
I
don't
think
we
need
to
spend
a
whole
bunch
of
time
figuring
out
how
we're
going
to
make
the
if
and
if
we
should
make
it
the
default
or
not.
C
C
A
A
And
that's
where,
like
an
option
that
applies
to
everything,
will
potentially
encourage
not
using
javascript
in
the
finalizers,
because
it's
sort
of
like
all
or
nothing
versus
the
individual.
Like
a
new
api
that
lets
you
choose,
you
can
sort
of
have
your
cake
and
eat
it
too.
Right
right,
I
mean
vladimir
since
you've
looked
at
it
a
lot
I
mean,
do
you
have
any?
C
Well,
I
I
don't
have
concerns
because
I
believe,
like
whoever
wants
to
call
javascript
from
from
this
known,
javascript
calling
finalizers.
They
can
always
do
it
in
their
code,
so
think
in
their
kind
of
kind
of
pure
c,
plus
plus
analyzers.
They
can
just
simply
schedule
like
set
immediate
or
something
which
will
do
actual,
maybe
javascript
calling
if
they
ever
want
to
do
it.
So
thank
you.
Maybe
it's
worth
for
us
to.
C
You
know
one
of
our
unit
tests
to
show
even
you
choose
like
for
node
add-on
to
not
to
call
javascript
finalizers,
but
sometimes
you
do
want
to
do
it
and
we
can
show
how
to
do
it
so
right
you
can
do
it.
I
think
we
can
address
all
consortium.
People
might
have.
A
C
So
I
I
I
agree,
it's
it's
much
better
solution
that
we
were
thinking
before
because
choosing
for
another
type,
ph
finalizer,
it's
it's
a
pain
right.
C
D
C
Okay,
I
I'll
need
somehow
to
build
it
on
top
of
your
current
pr
right.
C
Which
we
we
can,
we
can
iterate
upon,
and
we
can
potentially
merge
it
to
into
your
existing
pr,
but
it
might
be
too
big.
Your
pr
is
already
too
big.
Maybe
it's
probably
still
worth
to
do
it
to
the
next
step.
A
A
C
D
No,
I
I
just
saying
that-
and
this
is
just
a
single
point
of
this
pr-
and
we
have
another
next
step
for
this
pr-
is
to
move
away
the
not
be
great
external
buffer
from
another
puzzle.
So
that's
another
part
of
the
job
fnw
that
that
that
can
be
differed
until
vladimir's
work
on
the
new
finalized
types
has
landed.
C
Okay
I'll
work
on
it
kind
of
this
coming
week.
Hopefully
we'll
we'll
get
it.
We
can
discuss
the
like
next
meeting.
D
A
D
A
This
kind
of
makes
sense,
though
I
mean
this
is
this
is
a
little
related
to
what
we've
been
talking
about
in
that
they're
saying
it's
it's!
You
can
get
into
a
point
where
you're
trying
to
run,
I
think
a
finalizer,
but
you
can't
actually
call
back
into
javascript,
and
so
this
is
a
method
that
actually
will
let
you
tell
before
you
try
and
do
that.
D
It
can
be
useful,
but
I'm
I'm
thinking
if
it
written
for
us
what
else
you?
What
else
users
can
do
for
that
part
for
that,
since,
if
it's
false,
the
best
users
can
really
just
maybe
just
crash,
or
they
have
to
go
back
to
another
solution,
if
they
can
go
back
to
another
solution,
why
not
just
use
the
solution
that
it
doesn't
need
to
go
into
gas
at
the
first
place,.
A
C
D
D
So
in
the
shutdown
are
made
mainly
about
the
finalizers,
it's
it's
the
finalizers,
let's
say
when
they
are
not
allowed
to
when
they
are
using
the
new
parallel
straps,
for
example,
they
are
not
allowed
to
call
you
into
javascript,
then
then
they
should
not
to
call
into
javascript
in
the
panelists
at
all
at
all
conditions
right,
so
that
matters
doesn't
matter
if
the
environment
is
shutting
down
or
it's
not
shutting
down.
D
But
I
do
I'm
thinking,
we
need
a
use
case
for
this.
C
D
A
A
B
Yeah
and
even
coming,
and
he
added
the
the
example
on
an
example.
Okay,
so.
A
A
Okay,
so
otherwise
I
don't
see
anything
sort
of
new
in
that.
Let's
go
to
this
we'll
move
this
issue.
D
B
D
B
B
A
B
A
And
otherwise
I
don't,
I
don't
see
anything
too
new,
so
anything
else
that
we
think
we
should
talk
about
before
we
close
out
for
today.
A
Okay,
well,
if
not
thanks
for
everybody's
time,
I'll
see
you
all
in
a
couple
weeks
and
thanks
again
to
chenzong
for
sharing
next
week
and
have
a
good
week
and
see
you
in
get
up
thanks,
bye,.