►
From YouTube: Node.js N-API Team meeting - Sep 11 2020
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
And
actually
before
I
forgot
to
ask,
does
anybody
have
any
announcements
they'd
like
to
share
no
okay?
So
let's
move
on
so
discussing
any
pi
7,
I
don't
think
there's
anything
we
need
to
talk
about
that
gabriel.
Is
there.
B
No,
I
don't
think
so.
I
I'm
gonna
start
doing
the
the
changing
of
the
if-defs
and
and
then
that'll
be
an
api.
Seven.
A
Okay
sounds
good,
reviewing
external
examples.
I
we
just
had
that
on
our
list.
To
do.
I
don't
know
if
anybody's
had
a
chance
to
take
a
look.
A
A
Well,
unless
I
did
something
wrong,
there's
some
sort
of
outage
there.
Anyway,
let's
move
on
the
node
conf
eu
workshop
proposal.
Jim,
have
you
heard
anything
on
that.
A
Okay,
yeah,
I
don't
see
too
many
workshops
on
the
on
the
website.
Yet
so
maybe
that's
not.
They
haven't
figured
that
out
yet
yeah.
I
know
I
haven't
enabled
debug
testing
for
add-ons
yet.
A
And
I
think
that
leaves
us
to
looking
at
issues
which
have
been
raised
in
the
repos.
I
know
nicola
you've
been
working
on
the
next
node
add-on
api
release.
D
Yeah
there
is
a
an
issue
that
I
opened
the
809
so
yeah.
I
think
that
the
this
is
the
paraquest
about
the
documentation
that
gabriel
made.
So
we
need
to
discuss
about
something
because
you
commented
yesterday.
E
A
A
Okay.
Okay,
so
it's
not
that
you
I
I
wasn't
yeah.
That
was
one
thing
I
wasn't
quite
sure
was.
I
thought
that
maybe
before
the
addition,
where
you
s,
you
should
like
you
shifted
the
bass
class
right,
but
okay,
maybe
I'm
confused
and
that
didn't
actually
affect
the
templating.
We
just
hadn't
documented
it
that
way.
A
B
All
those
classes
use
crtp,
so
you
can't
avoid
templating.
B
But
okay,
but
that
the
fact
that
they're
only
available
in
your
versions-
that
is
true.
So
the
your
answer
might
not
be
to
this
comment
because.
B
A
A
A
A
B
D
I
I
don't
know
what
do
you
want
to
with?
You
want
to
do
with
tradeship
function
x.
So
is
it
possible
to
to
to.
A
D
So,
for
now
we
yeah
no
and
in
the
next.
A
C
D
Okay,
when
the
dock
will
be
landed,
I
I
we
start
to
prepare
the
the
release.
D
D
A
A
No
okay,
yeah.
D
The
first
I
won't
discuss
the
first
issue
right:
okay,
so.
D
Yeah,
I
I
just
made
some
experiment
and
when
I
yeah,
I
create
a
class
using
plain
javascript
yeah.
I
got
these
descriptors
for,
for
instance,
middle,
but
if
yeah
I
use
a
simple
example
of
object,
wrap
get
that
the
method
has
the
attributes
set
in
different
ways.
So
I
don't
know
I
I
test
these
on
node,
chrome,
firefox
and
yeah.
We
got
writable
true
configurable,
true
for,
for
instance,
mirot.
D
Instead,
when
we
made
these
on
on
object
wrapped,
we
get
that
all
attributes
are
set
force.
A
D
We
today
we
are
set
to
the
we
are
setting
these.
These
attributes,
like
navi
nabi
default.
B
Yeah,
that's
true.
Let
me
see
here
yeah,
so
we
have
nappy
default,
writable,
enumerable,
configurable
and
static,
so
so
so
yeah.
I
guess
I
mean
where
we
are
consistent
with
core
in,
in
that
we
we
set
mapping
default
if
we
pass
the
default
parameter.
B
So
I
I
don't
know
how
to
answer
this.
I
mean
it's.
It's
it's
just
our
default
and
people
have
been
using
it
this
way
for
how
long
now,
like
two
three
years-
and
so
I
don't
know,
I'm
tending
to
think
that
we
should
just
stick
with
it
like
if
we
change
the
defaults
in
c
plus,
then
that'll
be
inconsistent
with
with
the
core
api
yeah.
B
Well,
okay,
I
see
now
it
says
the
corresponding
non-api
set
prototype
method,
calls
that
are
done
to
my
understanding
of
the
v8
api.
They
use
the
dub,
which
is
defined
inverted
right,
so
so
yeah.
B
I
remember
I
remember
when
we
did
this
so
so
so
v8,
the
the
v8
flags
are
indeed
inverted,
but
we
decided
to
go
with
the
flags
according
to
spec
right,
which
is
which
is
the
same
as
pure
js
right,
which
means
you
know
innumerable
and
and
writable
and
configurable,
rather
than
read,
only
don't
enum
and
don't
delete.
B
So
so
so
that's
how
this
ended
up
happening,
and
so
that's
probably
why
our
default
ended
up
being
the
opposite
of
of
nan's
default,
because
nan
goes
with
the
v8
default,
which
is
open
completely,
whereas.
E
B
C
B
E
B
B
Add-Ons
and
then
there
they
might
get
some
unexpected
behavior
right,
because
then,
all
of
a
sudden
things
are
exposed
to
javascript
that
we're.
A
Another
like
not
be
default
would
still
be
the
same,
but
nappy
open
default
yeah
be
the
one
that
matches
javascript
like
if
that's
what
people
want
99
of
the
time
giving
them
a
shortcut
may
make
sense.
A
B
Would
like
we
would
we
would
we
would
say,
if
def
an
api
less
than
something
then
yeah,
then
or
them
together,
otherwise
use
the
core
value.
I
think.
A
Right,
I
guess
we
could
even
like
in
any
api
itself,
we
could
provide
a
shortcut
which
yeah
right-
that's
was
what
you
just
said,
or
them
together,
ourself.
A
B
But
the
question
is:
do
we
want
to
do
we
want
to
make
this
the
default?
Because
that
would
change
things?
You
know
it
might
even
cause
breakage.
A
D
A
A
B
A
B
On
the
other
hand,
you
you
have
to
you
have
to
add
the
attributes
you
have
to
specify
the
attributes,
whereas
the
nan
you
didn't
so
either
way.
There
is
a
difference,
there's
going
to
be
a
difference
from
right:
either
the
name
is
going
to
be
different
or
the
or
the
signature
is
going
to
be
different,
yeah.
So
yeah
we
we've
already
foregone
the
ability
or
forfeited
the
ability
to
to
to
translate
from
nan
transparency
right.
A
A
I
think
that
one
should
already
have
the
yep
okay,
so
we're
this
one
right.
A
A
Guess
existing
add-ons
right.
D
A
B
B
Yeah
but
yeah
yeah,
but
the
the
result
will
be
if
I'm
not
mistaken,
a
power
of
two
minus
one
right.
If
you
order
these
things
together,
so
so
we
can
still
introduce
power
of
two
values
later,
if
needed.
A
B
It
should
be
yeah
yeah
right.
We
we
could
do.
We
could
do
a
c
plus
preprocessor
directive,
but
that
would
be,
I
think,
a
little
bit
misleading.
I
think
it
would
be
best
if
we
just
added
it
to
the
enum.
C
Michael
maybe
make
the
the
identifier
string
literal
or
code
code
block
the
nappy
open
default.
Okay.
E
A
A
A
And
then
you
would
use
that,
actually
that's
something
we
could
do
even
without
making
it
in
the
c
api
itself.
B
A
A
Okay,
so
that's
that
one
any
others
that
people
think
most
things
look
a
little
older
than
last
meeting
other
ones.
That
people
think
we
should
take
a
look
at.
I
haven't.
A
A
This
one
sorry
just
before
we
get
there
gabriel,
is
this
one.
We
can
close
now
with
your
change.
B
Well,
well,
I
I
would
like
some
some
opinions
on
this.
This.
I
don't
think
we
really
need
this
working
with
javascript
values,
it's
just
a
very
small
document
and
it
just
links
to
yeah.
It's
just
links
to
some
of
these
classes,
but
I
mean
we
we
have.
We.
E
B
B
D
B
Well:
okay,
good
yeah,
all
right!
So
then
I'll
make
it
in
I'll
I'll
make
another
pull
request
for
removing
this.
C
A
Okay,
good
any
others.
We
should
on
the
note
on
api.
D
D
Into
that
now,
if
it's
closed
eight
to
one
zero.
C
D
D
He
asked
why,
okay
with
no
js
12
14
one,
he
don't
have
the
nabi
said
instant
thought,
but
it's
because
I
think
that
that
it's
experimental,
no
soi
is
not
clear
from
the
core
documentation
that
when
we
say
that
these
that
feature
is
added
on
yeah
added
in
it's
not
clear.
So.
D
D
But
in
an
odd
version,
12
a
dot,
a
a
8.0.
We
don't
have
nabi
version,
6
right,
okay,
so
yeah.
This
is
my
problem.
So
right.
That
version
is
the
version
when
we
added
for
the
first
time
that
that
visual
or
that
function,
but
at
that
time
the
function
was
experimental
right.
D
So
these
create
a
conf
confusion
in
my
opinion
and
oh,
maybe
we
we
could
yeah
maintain
added
in
and
then
make
stable
from.
D
I
I
don't
know
we
need
that
information
to
to
specify
these.
B
E
B
B
No
I'm
talking
about
I'm
talking
about
the
the
the
node
add-on
api
documentation
because
because
this
person
started
with
with
with
nf
set
instance,
data
right
and
call
and
call
and
set
instance
data.
So
so,
and
only
then
did
they
look
in
the
core
to
see
the
equivalent
right.
A
A
B
A
B
A
B
Yeah
yeah
I
mean
we
cannot.
We
cannot
fully
track
the
history
for
every
api
entry
right.
In
that
sense,
I
mean
in
most
cases
it's
probably
okay-ish,
but
but
I
think
I
think
what
the
the
kinds
of
headers
we
have
in
core
are
amply
sufficient.
I
think,
is
it
basically
in
core
if
it
doesn't
say
an
api
six
or
an
api,
five
or
whatever,
then
it
is
to
be
assumed
that
it's
experimental,
even
if
we
forgot
to
market
yeah,
but
I
guess
people
don't
necessarily
get
that
right
away.
A
A
B
Have
a
look
at
have
a
look
at
nappy
ad
async
cleanup
hook.
That
thing
has
has
actually
been
broken
right.
B
A
That
actually
is
good
yeah,
so
I
think
that's
good,
but
I
guess
the
point
is
like
you
may
need
to
it.
Tells
you
what
I
think
it's
more,
that
it
tells
you
you
need
nappy
version
six.
However,
it
also
tells
you
it
was
added
in
14.8.0,
and
so
you
might
think
that
you
could
use
it
there
and
you
can
as
long
as
you.
B
A
B
And
the
fact
is,
the
fact
is
that
this
whole
problem
stems
from
the
fact
that
we
forgot
to
update
the
documentation,
so
so
12.8
crystallized
with
a
bug
in
the
documentation.
That's
that's
the
fundamental
problem.
A
I
don't
think
that's
what
confused
them,
though
I
think
in
my
mind
it's
like
they
came
here.
They
looked
at
and
said
it's
12.8,
I'm
using
12.8.
I
should
be
okay,
yes,
but
but
they
didn't
like
the
experimental
bit
didn't
get.
You
know
they
didn't
realize.
Oh
I've
got
to
say
that
it
was
experimental
in
12.8.
A
A
D
C
A
A
You
know
possibly
for
the
version
like
the
12.x
line,
but
it
most
likely
just
the
latest
latest
and
say:
oh
okay.
Well,
that
was
added
in
12.8,
I'm
using
12.8.
Why
isn't
it
working
right?
Yeah,
okay,
okay,
that's
good!
Are
there
others
that
we
should
talk
about
otherwise
I'll
move
on
to
the
node
side
to
see
if
there's
anything.