►
From YouTube: Node.js N-API Team meeting - Agu 10 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
So
welcome
to
the
node
napi
team
meeting
for
august
10th
2020
we'll
follow
our
standard
approach
of
looking
at
what
is
in
milestone
11..
So
let
me
just
open
that
up
and
share,
while
I'm
doing
that,
does
anybody
have
any
announcements
that
they'd
like
to
share
nothing
here,
all
right?
A
A
I
guess
we
don't
have
jim
this
week.
So
no
update,
I
don't
know
if
anybody
else
has
any
update
on.
B
B
Other
than
that,
it's
mostly
working
right.
It's
getting
deployed
using
github
actions,
which
is
really
cool.
Okay,
I
don't
know
anything
further.
A
Good
yeah
and
I
turned
I
turned
some
twiddled
some
buttons
before
I
left
and
seemed
to
be
doing
the
right
thing.
So
that's
good
tracking
issue
for
module
is
node
supported.
A
A
A
Okay,
node
conf
proposal
workshop
proposal-
I
don't
know
if
there's
been
anything
news
on
that.
One.
B
I
guess
we'll
wait.
Jim
hasn't
mentioned
anything,
probably
no
news,
yeah
last
time.
Last
time
he
said
that
it's
it's
it's
basically
in
in
in
the
conference
organizers,
hands
now,
so
we've
done
our
part
and
we're
just
waiting
to
hear
from
them.
Okay,.
A
Enable
debug
testing-
I
don't
think,
there's
anything
new
on
that
right.
No.
A
B
A
A
A
B
A
B
Yeah
yeah
this
one,
this
one
gets
rid
of
some
some
of
our
exposed
ex
or
the
some
of
the
symbols
that
we
create,
because
we
we
have
a
lot
of
places
where
we
provide
a
nappy
callback
to
to
core,
and
there
are
just
too
many
like
there's
no
reason
for
there
to
be
a
separate
nappy
callback
for
for
property
descriptors
as
for
functions.
You
know
that
kind
of
thing.
So,
basically.
A
B
This
kind
of
refactors
it
by
you
know
the
the
way
in
which
you
end
up
calling
into
the
module
into
the
know
that
on
api
module,
which
is
you
know,
either
void,
return
value,
return,
void,
return
with
instance,
value
return
with
instance.
So
these
are
the
four
ways
right
and
that
can
be
re.
It
doesn't
matter
if
that's
if
that's
inside
an
object,
wrap
if
it's
inside
a
property
descriptor.
If
it's
inside
a
function,
you
can
always
use
the
same
callback,
and
this
only
applies
to
to
templated
ones.
B
C
B
C
C
B
No,
it's
just
it's
just
for
clarity,
because
it
was
already
it
was
already
using
promises
right.
It's
just
it's
just
a
dot,
then
dot,
then
dot
then
looks
looks
like
okay,
awkward,.
B
Chaining
actually
this
this
had,
I
mean
yeah,
basically
yeah,
so
just
just
to
clean
things
up,
because,
as
I
was
making
that
change,
I
realized
that.
Okay,
while
while
the
threat
say
function,
one
is
correct
in
that
it
does
the
chaining
already.
It
is
kind
of
awkward
in
that
it
uses
promises
when
it
could
use
a
much
cleaner
syntax.
So
that's
why
I
raised
this
issue.
B
A
B
C
C
So
what
I'm
gonna
do
is
just
un
like
remove
that
example
that
I
had
worked
on
and
just
really
port
the
existing
stuff
that
we
have
for
thread
say
function,
because
I
don't
really
have
the
time
to
figure
out.
What's
going
on
with
that
example
test,
I
really
don't
so
I'll.
Just
do
that
and
because
I
know
the
existing
tests
that
we
have
for
the
thread
safe
function
do
work,
because
we
I
did
like
a
implementation
where
tsfn
extended
tsfn
ex.
C
So
those
should
all
work-
and
I
do
have
already
that
test
ported
over.
So
I'm
just
going
to
remove
that
example
and
then
put
a
different
example
on
the
readme
and
then
like
dewey
is
the
correct
approach,
then,
to
put
also
an
example
on
the
node
add-on
examples
github
repo,
because
I
know
there's
been
a
lot
of
updates
now
with
the
github
pages
and
where
we're
hosting
documentation
and
I'm
not
really
sure
kept
up
to
date
with
all
of
those.
So
I'm
just
going
to
put
an
example.
Also
on
note
add-on
examples.
B
That's
that's
always
a
good
idea.
Yeah,
it
never
hurts
to
have
to
have
examples
and
then-
and
then
jim
can
work
them
into
into
the
tutorial
and
and
and
like
basically
link
link
to
the
example
from
the
tutorial
and
have
like
live
code
snapshots
straight
out
of
the
repo.
C
C
B
C
C
In
the
in
the
pr
there
is
that
first
document
there
doc
threadsafe.md,
if
you
can
somehow
open
that
up,
it
does
explain.
Yes,
the
pros
and
cons.
B
B
If
you
want
to
avoid
heap
allocations
and
if
you
are,
if
you
are
unwilling
to
to
handle
the
the
the
terminating
condition
to
make
sure
that
there
are
no
leaks,
then
you
know,
then
you
may
want
to
use
threat,
say
function
ex,
but
if
you're,
if
you're,
okay,
with
handling
the
determination
and
if
you're,
okay,
with
extra
heap
allocations
in
exchange
for
for
the
extra
benefit
of
being
able
to
to
provide
callbacks
for
each
data
item,
then
you'll
want
to
use
the
the
old
thread
safe
function.
A
C
A
Okay
sounds
good,
see
any
other
issues.
People
want
to
talk
about.
B
B
All
right
yeah-
I
guess
this
this-
this
requires
a
deeper.
A
A
A
B
So
none
of
those
look
new,
no
yeah
lucas
is
still
working
on
the
under
the
context
thing.
The
the
one
thing
that
has
landed
is
is
asynchronous
environment,
cleanup
hooks
from
anna
okay.
A
B
So
those
are
now
experimental
and
and
and
trickling
backwards,
and
so
so
because
we
we
are
starting
to
have
like
an
accumulation
of
new
apis.
I
was
thinking
like.
Maybe
we
should
start
thinking
about
napi
seven
right
and
I
I
had
a
thought
which
I
ran
by
the
group
last
week
and
maybe
I'll
run
it
by
you
as
well,
is
what,
if
we
were
to
have
an
api
release,
coincide
with
with
dropping
support
for
for
for,
like
the
oldest
lts,
so
the
one
that
that
has
been
in
maintenance.
B
So
that
means
we
would
release
an
api
in
april
and
and
then
then
we
wouldn't
have
to
worry
so
much
about
pushing
stuff
backwards.
You
know
and
back
porting
and
taking
advantage
of
the
leeway
that
we
have
that
that
you
know
the
rest
of
the
folks
are
willing
to
to
backboard
stuff
even
into
maintenance
releases.
B
You
know
in
april
like,
as
as
the
last
maintenance
release,
we
could
mark
that
as
stable
because
it
had
already
been
around
for
at
least
a
year
right
without
any
changes,
and
it
is
an
api,
so
it
is
pretty
stable
to
begin
with
right.
B
So
you
know
we
we
just
we
just
use
that
opportunity
of
the
last
minor
release
of
the
maintenance
version
to
just
release
a
new
version
of
an
api,
and
unless
we
backport
more
stuff
and
we
push
for
it
during
the
remainder
of
the
of
the
cycle,
an
api
7
would
just
consist
of
of
a
buffer
detaching
stuff,
which
I
think
is
fine.
We
don't
have
to
like
maximize
whatever
it
is,
that
is
in
a
new
release
right.
B
B
A
B
B
So
so
then
the
only
stuff
that
that
wouldn't
make
it
into
seven
would
be,
if
I'm
not
mistaken,
the
the
type
tagging
stuff
that
that
I
just
landed
like
two
weeks
ago
and
and
the
asynchronous
cleanup
hooks
that
anna
landed
last
week
so.
A
B
Well,
if,
if
we
can
backport
them
all
the
way
to
10
right,
and
so
I'm
thinking
like
especially
the
the
asynchronous
cleanup
hooks,
they
they
go
pretty
deep
into
node
right.
So
so
10
may.
A
B
B
A
A
A
Not
only
that
like
we
don't
have
to,
we
could
declare
a
seven
and
not
backport
it
to
one
of
the
earlier
releases
right.
True.
B
A
B
Yeah
and
and
we
we
have
always
sort
of
made
it
a
point
to
say
that
you
know
an
api
x
is
now
supported
on
all
lts
releases.
You
know.
A
A
I
think
that
was
mostly
because
we
thought
those
were
like
important
for
adoption
right.
Okay,
true,
I
think
I
think
if
it's
just
like
it's
now
getting
to
the
point
where
it's
like
here's,
some
incremental
things
which
are
beneficial,
but
you
know
we
don't
need
to
have
them
there
for
people
to
start
adopting.
B
Okay,
yeah
yeah
all
right,
so
so,
okay,
so
along
those
lines,
then
I
guess
the
the
priority
becomes
to
get
these
things
back,
ported,
and
so
so
I
think
I
think,
the
own.
So
the
two
things
being
the
type
tagging
and
the
the
asynchronous
cleanup
hooks.
The
type
tagging
is
pure
js,
so
I
mean
not
pure
js.
It's
pure
engine,
not
not
not
any
node
internals,
so
so
that
would
probably
be
easier.
B
A
B
Right:
okay,
all
right,
okay!
So
so
so,
let's,
let's
keep
an
eye
on
on
how
how
the
back
ports
are
are
coming
along
and
then
yep.
And
then
you
know
when
we
think
they've
baked
long
enough,
then
we
can
just
call
them
an
api.
Seven
right
right.
B
Oh
yes,
yes,
yes,
jim
jim
jim,
pointed
out
during
the
last
meeting
that
that
yeah
the
doc
was
out
of
date.
It
was
out
of
sync
with
the
actual
status
of
these
apis,
and
so
so
he
he
landed,
or
he
submitted
that
pr.
Okay,.
A
Good,
okay
right,
so
anything
else
that
we
want
to
talk
about
on.
It
doesn't
look
like
there's
too
much
else
there
in
core.
B
Thinking,
oh
okay,
the
yeah,
I
I
there's
you
you!
You
submitted
an
issue
I
forget
in
which
repo,
I
think
was
abi
stable
with
with
some
with
like
a
like
kind
of
a
tutorials
page.
That
was
not
gems
but
another
one.
If
I
remember
correctly,.
A
A
We're
trying
to
find
review
external
examples
right.
This
works.
B
Oh
yeah
yeah
yeah,
yeah,
yeah,
okay,
okay,
so
this
is
okay
workshop
we're
going
native
pool
90,
okay,.
B
A
B
Well,
no,
but
in
a
pr
you
can
you
can
only
comment
on
the
changes
right.
You
cannot
comment
on
on
stuff
that
you
don't
change
right
and
if
it's
already
mastered,
then
that's
not
stuff
that
that's
changing
right.
I'd
have
to
write
a
pr
that
changes.
Everything
and
then
just
leave
comments
in
in
the
interesting
things.
Well,.
B
A
A
B
Let's
see
that,
oh,
if
it's
just
a
small
observation,
if,
if
and
when
we
do,
when
you
release
it
for
of
know
that
on
api,
it
would
probably
have
to
be
some
vermin
right
because
we
got
the
add-on
stuff
in
there
now.
So
that's
a
new
api
so
yeah,
unless
we
shuffle
the
commits
to
to
to
just
have
a
another
patch
release
with
just
a
fix.