►
From YouTube: 10-12-2021-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
Okay,
yeah,
I'm
sorry,
I
didn't
notice
that
got
it.
Okay,
okay,
so
great.
Welcome
to
the
node
api
team
meeting
for
december
10th
2021.
A
We
will
follow
our
standard
agenda
of
looking
at
the
things
which
are
tagged
for
most
down
11..
Anybody
have
any
announcements
they'd
like
to
make
before
we
get
started.
A
Okay,
I
will
share
the
screen
with
the
which
screen
is
that
here
we
go
okay.
I
think
this
is
the
right
one.
A
B
Yeah,
I
don't
think
I've
gotten
any
more
comments
other
than
yours.
For
now,.
A
A
D
Is
there
like
an
a
link
or
something
I
remember
seeing
this
blog
post,
but
I
don't.
A
Yeah
we
dug
up
the
link
jack.
Is
there
any
chance
you
could
post
the
link
into
the.
B
A
D
A
A
A
Okay
sounds
good:
the
the
serial
port
review
and
stuff
anything
to
update
on
that
front.
D
They
are
still
trucking,
along
from
what
I
heard
last
is
they
were
able
to
complete
some
testing,
but
I'm
still
not
sure
at
what
point
they
are
for
having
it
be:
a
production
ready,
release.
Okay,
I
think
they're.
Still
on
the
last
I
heard
on
monday,
they
are
hoping
to
have
a
beta
release
soon,.
A
A
E
E
I'm
not
sure
if
people
have
any
concern
with
the
example,
maybe
you
can
slide
upwards.
E
That
they
are
saying
that
this
example
is
misleading,
because
we
can
assign
a
value
to
a
another
np
object,
so
we
can
change
the
other
number
object
and
this
one
can
be
can
be,
that
is
to
say,
when
we
changing
the
objective
to
the
opportunity.
One
can
be
changed
too.
So
I
I
believe
this
problem
is
more
about
the
design
and
the
copy
semantics,
rather
than
the
set
constantness
right.
E
C
A
A
Can't
because
I'm
like
that
seems
like
to
me,
that's
independent
of
whether
the
method
is
const
or
not.
Right,
like
like
the
a
const
method,
doesn't
mean
to
me
that
there
isn't
some
other
way.
The
object
can
change
underneath
you
right,
it
just
means.
If
you
call
the
method
on
the
object,
you
expect
it
not
to
change
the
object.
You
called
it
off.
E
Yeah
right
so,
but
it
is
changing
the
project
and
so
effectively
I'm
finding
that
these
methods
have
side
effects
and
I'm
finding
the
side
effects
is
very
another
thing
about
the
constantness
of
a
function.
So
if
a
function
can
cause
side
effects,
how
can
we
make
it
marketed
as
constant
so
that
that's
the
problem
and
not
the
less
than
understanding
of
the
author.
A
C
D
F
Yeah
to
me,
I
I've
seen
such
situation
quite
often
like
then
we
implemented
this
smart
pointers
or
something
like
this
and
people
capture
it
in
lyandas,
like
a
lambda
capture
by
defaultly,
understood
all
captures
as
a
constants.
F
So
it
means
that
it's
in
people
writing
kind
of
lambdas.
For
them,
it's
very
difficult
to
use
such
objects.
If
methods
not
const,
they
either
have
to
mark
the
whole
lambda
to
be
immutable,
which
often
they
don't
really
mean
to
do
because
mutable
lambdas
may
be
not
good
with
some
scenarios
special
threading
scenarios
and
think
it's
either.
People
have
to
do
this
technique,
copying
objects
and
it
makes
this
code
bloat
if
you
like.
F
So
if
we
don't
really,
if
we
really,
it
depends
on
semantic,
if
we're
just
wrapping
up
some
kind
of
reference
and
don't
it
all
these
methods
not
really
actually
changing
pointer
itself,
but
only
doing
some
actions
on
this
kind
of
wrapped
up
pointer
being
all
these
methods
to
be
constant
is
really
kind
of
beneficial.
In
many
scenarios.
A
A
So
if
I
call
you
know
if
I
call
set
and
then
if
I
call
get
set,
get
I'm
going
to
get
a
different
answer
on
the
second
get
right
and
you'd
think
that
the
compiler
one
of
the
optimizations
it
might
do
is
if
it
sees
you
doing
too
const
like
if
you
see,
if
you
did
a
a
a
get
and
then
you
do
a
set
which
is
a
const
and
then
you
do
a
get,
which
is
the
exact
same
thing?
Would
the
compiler
not
be
fair
to
say,
okay?
A
F
It
was
like
a
very
interesting
talk
from
herb
sater.
Then
he
kind
of
offers
to
treat
constance.
It's
more
like
a
promise
of
multi-threaded
safety.
If
you
like,
it
was.
D
D
Well,
I
just
want
to
point
out
another
instance
of
a
pr
that
we
had
where
we
were
discussing
removing
const
or
keeping
const
or
whatever,
and
we
actually
left
it
there,
and
I
think
the
reasoning
is
on
the
bottom
of
this.
This
pr.
D
E
Actually,
I
want
to
point
out
that
the
the
very
strong
opinion
that
convinced
me
not
to
make
the
changes
that
even
this
pr
is
going
to
break
existing
codes
and
I'm
not
very
strong
of
opinion
or
landing
this
pr.
So
sometimes
it
could
say
when
we
end
adding
new
cons
to
methods.
I'm
very
concerned
about
the
semantics
of
this
method,
but
this
kind
of
thing
is
already
in
the
wild
and
we
when
we
remove
them
the
euro
column
might
might
break.
So
I'm
not
strong.
E
I
I
don't
have
a
strong
opinion
on
lending
this
pr,
but
I
have
a
concerned
opinion
on
landing
on
adding
new
const
marks,
to
like
say,
method
like
stat
and
delete
or
any
other
method
that
is
seemingly
causing
that
effect
and
very
obviously
their
concentrating
effect.
A
Right
and
they're,
like
they're
they're
intended
so
they're
they're,
not
even
side
effects.
They're,
like
you,
call
a
set
to
change
the
state
of
the
object
and
and
okay
that
if
that
state
was
in
c
plus
land,
so
that
you
know
we
could
have
a
value
in
the
c
plus
object,
then
you
would
clearly
say
that
is
not
const
right.
A
A
I
I
I
guess
it's
the
I
we
haven't
at
least
personally,
I
haven't
seen
enough
to
convince
me.
It
should
be
just
the
pure
state
of
the
c
plus
object
so,
and
you
know
I'm
certainly
open
to
that.
I
just
haven't
seen
where,
where
there's
something
that
makes
that
case
is
to
like
know,
ignore
all
the
other
anything
else.
F
I
I
think
that
I'll
I'll
kind
of
look
up
what
in
standard
library
for
c,
plus
plus
the
way
to
implement
a
student
share,
ptr
or
or
stood
big
ptr?
F
I
guess
we
can
follow,
maybe
their
pattern
somehow,
but
at
least
then
inside
of
company,
I
had
to
implement
some
kind
of
smart
pointers
effectively.
I
have
some
kind
of
class
which
wrapping
up
pointer
and
this
pointer
points
to
some
other
state.
F
My
personal
experience,
being
it's
always
better
to
allow
all
these
cons
methods
to
change
your
state,
which
is
somewhere
else
like
as
long
as
they're,
not
changing.
What's
inside
of
this,
don't
change
in
this
pointer,
which
will
ripen
up.
They
should
be
all
fine
to
do
their
job,
always
conspiracy,
because.
C
F
Become
so
much
easier
to
use,
it's
just.
A
E
E
E
D
E
D
A
It
definitely
is,
I
think,
like
knowing
that
makes
me
think
we're
more
likely
on
the
right
track.
I
still,
I
still
think
you
know
if,
if,
if
you
know
like
vladimir
can
lay
out
the
here's,
the
reasons
why
it
makes
sense
or
whatever
from
experience
and
what
the
sort
of
a
deeper
c
plus
plus
plus
look
we
you
know,
that's
that's,
definitely
something
we
should
factor
in
too,
but
it
does
it
did
if
they
were
all
const.
It
would
sway
me
the
other
way
right,
but
since
they're
not
it
does
sway
me.
D
F
F
I
I
want
to
kind
of
draw
a
kind
of
a
parallel
with
javascript
say
in
javascript,
and
using
const
is
pretty
much.
You
can
call
any
method
on
this
const
object
which
can
modify
a
state
of
the
object
if
you
like
in
javascript
right
and
as
long
as
you're,
not
changing
the
variable
itself,
like
exactly
internal
guts,
I
mean,
like
variable
itself
can
be
replaced,
but
any
methods
and
stuff
like
can
be
called
in
any
state.
It
can
be
changed
and
stuff
like
that.
F
A
A
A
So
I
think
there
was
some
discussion
a
little
while
ago,
the
meeting
I
wasn't
that
I
missed
and
but
we've
got
a
comment
back
to
the
originator,
so
I
think
we're
just
waiting
to
see
if
there's
any
response
to
that.
A
E
Sorry,
I
don't
have
many
times
this
week.
I
I'm
not
sure
I
can
take
it.
Look
at
me
attending
my
manner.
A
Okay,
but
we'll
we'll
leave
it
open
if
anybody
else
has
time
to
take
a
look
too,
but
otherwise,
we'll
that's
sort
of
our
next
step
is
to
take
another
another
look
and
see
like
it's
kind
of
sounded
like
we.
We
thought
we
were
doing
this,
the
the
right
thing
by
telling
v8
that
the
method
had
side
effects,
so
it
shouldn't
try
calling
it
a
preview
mode,
but
it
was
still
happening.
A
Right
we
were
throwing,
we
were,
we
were,
we
couldn't
throw
primitive
errors
before
and
this
is
sort
of
fix
that
up
and
that
we've
wrapped
them
and
then
you
can
unwrap
them,
and
but
it
didn't
actually
solve
the
problem
right
in
the
in
the
console,
which
sounds
to
be
like
the
console's
trying
to
it's
trying
to
execute
the
method
to
preview
the
results.
A
B
Also
sorry,
this
bug,
I
think,
disappears
in
their
newer.
Newer
source
builds
like
when
I
pulled
like
near
the
newest
node.js,
build
and
build
it
from
source.
This
file
disappears.
But
when
I
reverted
to
the
one
I
was
using,
which
is
version
14.17
and
like
it
just
happened
again,
I'm
not
sure.
If
there's
any
changes
in
there.
B
B
C
C
A
A
A
Right
or
where
it
was
fixed
in
one
of
the
things
yeah
yeah,
it's
more
like
fixed
at
some,
it's
like
okay,
it
was
fixed
when
we
did
this
and
if
that
included,
a
v8
update,
that's
quite
possible
that
v8
wasn't
respecting
that
flag
or
something
for
some
reason
or.
D
B
Yeah
and
the
master
branch
was
I
think
a
month
ago,
or
something
like
that,
because
there
was
that's
remember
when
I
had
that
issue
with
the
ci
passing.
So
I
do
a
rebase
on
the
master.
So
I
think
it
was
around.
C
B
A
A
E
A
G
D
A
E
D
Yeah,
it
has
to
do
with
some
sort
of
not
generics.
What's
it
called
in
c
plus
plus
he
says
that
cvvec
class
requires
type
arguments,
so
he
needs
to
figure
out
how
to
add
that.
A
But
that
cv
vac
doesn't
is
that
actually
a
node
at
on
api?
No,
it's.
D
D
What
that
is,
I'm
familiar
with
with
it,
but
I've
never.
I
haven't
used
it
recently.
I
did
like
some
some
academic
work
with
it,
but
we
can
see
that
the
first
sentence
where
he
says
after
this
code
block
of
course
the
example
won't
work,
and
he
said
so
something
like
this
cvvec
inch.
Three,
you
see.
D
Has
let's
just
take
some
thinking.
E
E
D
A
D
Exactly
yeah,
that's
the
the
complexity
you
know
like
I'm
assuming
he
wants
the
constructor
of
when
you
construct
this
object
in
javascript.
You
pass
to
it
the
size
that
you
want,
and
then
he
wants
that
size
to
be
reflected
in
the
senior
vec
call
type.
It's
just
a
guess,
because
that
would
be
the
only
reason.
D
But
I
guess
we
can
just
do
some
some
some
c
plus
research
on
that,
because
that
seems
just
to
be
a
c
plus
thing,
not
necessarily
a
no
doubt
on
api
thing.
A
A
D
Friday,
the
17th-
I
should,
I
believe
yes,
I
should
be
there.
I
do
start
vacation,
though,
on
the
18th,
so
I'll
be
gone
basically
a
month
after
the
18th.
So
one.
D
Oh
yeah
it
because
gabe
was
the
one
who
requested
changes,
so
he
has
to
re-review
it.
Okay,
that's
fine!.
D
D
G
Is
one
thing
one
other
thing
I
wanted
to
ask
you?
You
have
mentioned
the
add-on
with
the
filter.
The
add-on.cc
did
not
work.
The
test
on
add-on.cc
did
not
work,
but
it
it
is
the
most
simplest
one.
It
should.
I
I
even
tested
it
today
morning
for
making
it
not
to
work.
I
made
a
slight
change
in
the
code
like
a
coding
bug
or
something,
and
then
I
made
a
syntax
error
to
make
it
fail.
So
that's
puzzles
me
like
which
version
of
node
did
you
test
on.
D
I
think
I'm
using
a
12.
12.
G
D
Yeah
that
like
because
add-on
work
add-on,
does
not
work
but
object,
wrap
works,
because
I
remember
in
your
example
that
you
sewed
for
the
pr
it
uses,
object,
wrap
and
then
I
tried
that
one
and
it
succeeded-
and
I
was
like
okay
well,
let
me
make
just
a
bash
script
that
runs
the
unit
test
for
all
of
the
files
and
then
from
there
I'm
seeing
a
lot
of
failures.
Okay,
I
don't
know
if
it's
do
you,
what
type
of
operating
system
do
you
use.
G
Mac
are
you
I'm
on
a
mac?
Okay,
I'm.
G
Okay:
okay,
let
me
take
a
fresh
package
because
I
have
been
using
the
same
package
for
like
the
same
repo
for
two
months.
So
so
maybe
let
me
clone
it
refresh
and
then
try
it.