►
From YouTube: 2021-09-03-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
So
welcome
to
the
node
api
team
meeting
for
september
3rd
2021
we'll
follow
our
standard
approach
of
looking
at
the
issues.
We've
tagged
for
the
milestone.
Anybody
have
any
announcements
before
we
do
that.
A
B
A
Okay,
great
so
still
moving
ahead,
that's
good!
Let's
see
tracking
issue
or
stale
issues.
Okay,
so
let's
look
at
stale
issues.
Okay,
we
don't
see
anything
new.
D
Last,
I
guess
this
is
one
that
we've
left
right,
waiting
to
hear
back
from
the
poster
to
get
some
any
sort
of
clarification
or
whatnot.
A
A
A
A
A
A
D
One
thing
that
I
just
see
right
now
on
line
44..
Okay,
you
get
a
space
between
before
that
parenthesis
and
symbol
description.
Do
you
see
that.
C
A
A
A
E
A
D
F
I
have
a
question:
is
that
things
we
are
only
like
say
losing
the
restrict
in
the
core
to
symbols?
So
I
believe
the
best
solution
in
here
is
that
we
don't
extend
the
not
ap,
not
p
reference
in
order
to
appear
to
other
primitive
values,
but
rather
than
just
extend
to
simple
values
too,
but
in
the
error
like
say,
nappy
error
where
create
create
native
reference
to
those
primitives
and
loosen
it
to
be
allowed
to
reference
to
arbitrary
javascript
values,
so
that
the
semantics
of
nappy
reference
will
not
be
like,
say
different
with
the.
A
C
A
I
guess
the
usage
is
still
sort
of
trending
upwards,
which
is
good,
but
we
don't
have
any
specific
ones
we
know
about.
I
guess
in
terms
of
dependence,
it's
we
could
remember
that,
but
I
don't
remember
what
it's
been
before
so
anyway.
A
So
I
guess
nothing
to
discuss
on
that
one:
creating
a
matrix
of
tested
methods
so
jack.
Do
you
have
things
you'd
like
to
discuss
on
that
front.
E
Nothing
at
the
moment
I
just
working
through,
like
the
three
essay
functions,
it's
quite
a
bit
of
an
api,
so
I
do
like
go
through
create
requests
early
but
should
be
done
like
next
week.
I
think.
A
D
F
D
F
The
linker's
configuration
in
node
core
is
just
way
too
much
tight
bonded
to
the
core
project
so
and
it's
hard
to
migrate
it
to
be
used
in
order
to
appear.
C
F
F
A
D
A
A
few
I
think
standard
might
be
along
those
lines,
just
in
the
like
it,
it
uses
pretty
sure.
That's
the
one
that
uses
semicolons,
which
I
believe
we
have.
F
B
It
would
be
a
problem
if
you
have
a
big
oracle
supplement.
B
A
A
That's
part
of
as
part
of
the
test
of
part
of
adding,
like
I
know
like
in
node
core.
We
wouldn't
do
that
simply
because
it's
a
pain
to
to
backboard
and
everything,
but
it
we
very
rarely
back
port
to
any
other
thing,
so
it
might
be
just
like.
Let's
just
run
it
there's
a
big
change,
we'll
assume
that,
because
it's
run
by
the
linter,
if
it's
all
okay
and
then
just
leave
it
does
that
make
any
sense.
D
A
Right,
like
it's
a
it's,
a
mechanical
enters
done.
The
change.
The
the
challenge
I
see
is
where
we
defer
it
to
when
you
actually
have
another
pr,
and
then
you
have
a
whole
bunch
of
changes
mixed
in
and
you
don't
know
which
are
changes
and
which
are
so
if
we
just
sort
of
run
it
all
as
part
of
it,
it
might
get
around
that.
F
The
point
is
that
we,
the
the
git
ceiling
format
we
are
using
now
it's
just
formatting
the
lines
you
are
changing,
so
it's
fairly
low
overhead
when
you're
reviewing
the
new
changes.
Since
it's
it's
only
formatting
the
line.
You
are
changed,
yeah,
so
not
the
whole
file.
You
are
changing,
so
it
will
now
format
the
other
other
non-relevant
lines
to
your
change.
A
F
A
Yeah
yeah
and
it's
a
little-
you
know
one,
it's
mechanical,
we
don't
think
there'll
be
any
problems.
We
don't
need
to
review
it.
Two
it's
test
code,
so
risk
is
a
little
bit
lower
as
well,
and
we
don't
do
back
ports
regularly.
So
the
impact
of
the
code
being
changed
from
previous
version,
especially
in
the
test
stuff.
A
F
I
find
those
two
up
both
fines,
since
we
are
just
formatting
our
test
cases,
but
I
will
prefer
the
incremental
one,
since
we
are
more
the
work
to
reveal
that
the
change
will
be
much
more
enlightened.
We
don't
have
to
reach.
F
It's
hardly
the
lingerie
make
mistakes,
but
you
know
we
are
changing
a
lot
of
files
who
can
tell
that
there
is
nothing
else.
We
are
accidentally
adding
to
it
or
any
other
things,
and
I'm
not
saying
that
it's
intended,
but
it
may
be,
since
we
are
hardly
reviewing
the
changes.
So
it
will
be
hard
for
us
to
identify
any
problems,
but.
D
D
Is
like
on
they
have
local
changes
on
their
branch,
and
then
they
run
the
linter,
and
then
they
commit
all
of
those
changes
as
part
of
the
linting
process,
but
it
actually
includes
their
local
changes.
That
would
be
something
that
would
be
hard
to
find
because
we
would
be
assuming
that
everything
is
fine
is.
Is
that
one
of
your
concerns
then
yeah
yeah
they're
an
example
of
a
concern
right.
A
The
other,
the
other
thing
we
could
do
like.
If
that's
the
main
concern,
we
could
actually
have
two
people.
Do
it
separately
right,
like
whoever's
got,
the
pr
could
run
the
linter
boom
here
it
is
that
should
come
up
with
a
very
fixed
diff
right,
except
for
some
files,
and
then,
if
somebody
else
does
the
same,
okay,
I'm
just
going
to
run
the
checkout
run
the
linter.
D
Yeah
I
mean
that's
also
makes
sense
like
if
somebody's
doing
the
the
person
who's
introducing
the
linter
changes
can
put
in
their
pr
the
steps
that
they
did
in
order
to
run
the
linter,
and
then
somebody
that's
reviewing
the
pr
just
needs
to
perform.
I
guess
those
same
the
same
steps
and
then
see
if
they
can
get
the
same
output.
D
D
A
Okay,
yeah,
that's
good
I'll.
Just
put
you
know,
I
think
this
is
also
like
a
good
thing
that
anybody
can
pick
up.
So,
let's
put
a.
A
Okay,
that
sounds
good,
okay,
so
jack
that
that
was
on
the
test
front.
There
wasn't
anything
else
you
wanted
to
chat
about.
E
No
that's
different.
Yes,
okay,.
A
B
I
won't
only
remember
that
I
have
paragraphs
on
unload
api
leaders
headers
and
I
made
the
changes
that
you
requested.
So
please
take
a
look
when
you
have
time.
B
D
A
C
B
Yeah,
and
so
I
think
that
he
is
asking
us
if
it's
a
good
idea
use
worker
threads
instead
of,
for
example,
yeah,
I
already
uv
or
essentially
async
worker.
No,
if
you
use
on
api
or
node
a
api,
the
the,
if
you
see
at
the
lincoln
issue,
the
problem
is
that
libya
will
handle
a
pool
of
four
treads
by
the
fourth
yep
and
it's
a.
B
Yeah,
you
need
to
wait
one
that
one
of
this
connection
is
will
be
recent,
no
right,
okay.
So
this
is
a
problem
and
the
the
so
this
is
the
main
problem.
Essentially.
B
But
I
I
think
that
it's
the
same,
if
you
use
the
worker
thread,
because
you
have
a
limited
number
of
cpu
available
on
the
system,
and
so
you
need
to
implement
a
pool
of
worker
threats
to
handle
all
the
things
and
then
yeah
and
do
the
same
thing
that
libya
we.
B
Do
under
the
hood
for
you
no,
but
with
the
worker
thread,
so
you
you
need
how
to
implement
these
these
this
logic
but
yeah
you,
you
can
do
the
same
thing
using
the
the
the
worker
child.
For
me.
A
B
Yeah
yeah,
it's
it's
a
limited
asia,
but
yeah.
B
H
B
Yeah
using
the
drugs,
it's
it's,
it's
yeah!
It's
it's
all
about,
and.
B
This
could
discourage
this.
This
approach.
A
D
Yeah
right,
but
just
for
my
clear
questioning
the
lib
uv
worker
pool
of
four
that
default
pool
that
doesn't
does
that
interact
with
the
number
of
worker
threads.
You
can
have
no
right.
D
B
Right
yeah,
because
at
one
point
you
know
if
you
have
so,
for
example,
for
eight
cpus
yeah
you
can
it's
not
a
good
idea
in
terms
of
performance.
Oh
continue
to
create
a
new
worker
treadmill.
So
if
you
have,
for
example,
20
worker
threat,
it's
it's
not
ideal
for
me.
So.
B
You
you
need
to
to
have
a
right
number
of
worker
threat
and
have
a
policy
for
doing
this.
No
a
good
example
of
a
library
that
do
this,
for
me
is
bishop,
for
example,
yeah.
So
yeah.
You
need
to
improvise
something
like
like
like
that
and
then
use
use
user
user
use
it.
It's
yeah!
You
are
adding
some
ovary
in
this
case,
but
you
can
remove
the
limitation
of
have
only
four
four
thread.
D
Okay,
so
they
be
able
to
remove
the
the
limitation,
but
then
it
basically
puts
the
architect
the
job
of
architecting,
the
correct
thread,
context,
handling
and
whatnot
to
the
programmer,
because
now
they'd
be
very
making
worker
threads.
Instead
of
leaving
everything
to
live
uv
right.
B
A
B
C
C
B
Yeah,
a
control
controller.
Yes,
yes,.
A
B
Have
your
tried
model
so
inside
the
native
code
you
can
do
you
can
create
any
number
of
threads
that
you
you
you
want.
I
I'm
not
yeah.
D
D
D
Actually
I
don't
know
if
the
thread
safe
function
actually
allows
you
to
to
do
multiple
things
at
the
same
time,
because
I
think
the
thread
say
function
just
allows
you
to
queue
a
call
into
to
node
and
then
it'll
run
you
that
your
your
c
your
c
function
on
the
main
thread
at
a
specific
time.
I
don't
know
if
you
could
be
able
to
spawn
multiple
of
those.
D
I
think
the
main
thing
is
like
their
use
case.
If
they're
doing
a
lot
of
heavy
work
on
that.
That
thread.
A
I
think
the
threads
say
functional
like
on
the
main
thread.
The
the
issue
in
this
case,
I
don't
think,
is
the
work
on
the
main
thread.
Necessarily
right,
like
the
issue
is
they're,
creating
a
lib
uv
thread,
they're
making
a
native
call
and
they
must
be.
Are
they
like
they're
using
libya
v
to
create
the
threads?
Now,
I'm
not
sure
thread
say
function
may
be
silent
on
how
you
create
your
thread.
D
D
If
they're
just
doing
something
where
they're
creating
a
bunch
of
threads
and
doing
something
very
simple
and
small,
then
yeah
they
could
go,
I
think
it
would
make
sense
to
go
to
thread
safe
functions.
I
think
we
may
need
to
take
a
little
more
look
into
what
their
their
code
does
right.
D
Because,
like
the
the,
if
they're,
if
they
just
have
some
code
that
you
every
time
you
call
a
function,
creates
a
thread
and
if
they
have
that
in
some
loop,
then
of
course
it's
gonna
only
allow
four
simultaneous
ones
if
you're
using
lib
uv
and
if
what
that
thread
is
doing
internally
is
really
small,
then
yes,
it
could
easily
be
a
thread,
say
function
design.
I
would
end
up
having
to
figure
out
how
to
handle
the
the
native
threading
themselves.
A
D
A
So
maybe
that's
another
another
disadvantage.
You
know.
B
Yeah,
because
you
you
made
the
extra
calls
to
from
javascript
to
to
united
code,
I
I
think
is
it
right.
B
Yeah,
I
I'm,
I
think
it's
it's
good.
A
A
Okay,
thanks
any
other
ones
that
we
should.
B
No
for
now,
no
okay
on
the
main
or
not
not
draper.
If
there
is
some
issues.
A
This
one
there's
been
a
bit
of
back
and
forth
these
ones.
It's
always
gc.
This
one
is
interesting
in
that
he
seems
to
say
just
by
wrapping
a
buffer
and
another
object.
The
issue
goes
away,
so
I
think
it's
interesting.
A
A
That's
that
one
happy
promise
then.
C
A
D
A
D
H
Bottom
or
it's
or
it's
duplicated,
or
you
can
find
that
issues
that
one
there.
Oh.
A
C
A
D
Yeah-
and
this
was
his
first
one
and
then
he
closed
it,
but
this
is
where
even
gabe
said
it
above
too.
E
D
C
D
C
C
G
D
B
Or
maybe,
to
avoid
these,
we
can
wait
one
week
and
we
will
close
next
week.
A
B
Yeah
yeah
you
just
just
for
waiting
some
objection
to
our
decision.
B
H
A
A
B
Of
yeah
go
ahead,
calling
navi
underscore
function
from
mingw.
I
think
that
I
answered.
A
Any
other
ones
that
people
want
to
discuss.
A
Okay,
okay,
anything
else
we
should
talk
about
before
we
close
out
for
today,
then.
B
If
gabriel
approved
my
pr
on
load,
abby
headers,
I
think
that
I
can
authorize
it
to
publish
npm.
Is
it
right,
yep,
okay,.
A
And
I
think
you
know
if
let
me
just
take
a
quick
look:
yeah,
no
node
add-on,
sorry
node
api.
A
B
A
The
screen-
oh
yeah,
okay,
sorry,
yeah,
good
point,
yeah,
I'm
looking
at
it
myself,
but
that
doesn't
help.
A
Right
so
like,
if
we
look
at
the
collaborators,
we
always
want
to
have
the
node.js
foundation.
Now
the
name's
not
right,
we
may
want
to
change
that,
but
anyway,
like
we
have
this
account,
which
the
build
working
group
has
access
to.
So
if
you
can
add
that
classic
that
account
as
a
collaborator
as
well,
that
would
be
good
for
the
new
package.
H
B
A
A
Okay,
if
not
well,
thanks
for
everybody's
time
this
week,
great
meeting
as
always-
and
we
will
see
everybody
in
github
in
next
week-.