►
From YouTube: Node.js Node-API team meeting 2021-01-29
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
Hello,
everybody
and
welcome
to
the
friday
january
29th
meeting
of
the
node.js
api
team.
If
nobody
has
any
announcements
for
which
I
will
give
a
brief
pause,
then
we
can
get
started
with
the
agenda.
Here's
that
brief
pause.
A
Okay
agenda:
it
is,
let
me
just
open
things
up:
okay,
api,
I'm
gonna
start
sharing.
As
soon
as
I
got
to
the
agenda,
I
did
not
forget,
though
I
sometimes
do
that,
let's
see
milestones,
milestone,
11
and
start
sharing
the
screen.
That
would
be
this
thing.
A
Yes
share
all
right
here
we
go
so
let's
go
through
the
milestone,
let's
see
rename
primary
branch
to
main
this
would
be
for
know
that
on
api,
I'm
assuming
I
don't
know
if
we've
made
any
progress
on
that
during
the
week.
B
What
we
should
do,
what
we
should
do
is
just
check
the
repo
and
see
what
the
default
branch
is.
That's.
C
A
C
A
A
All
right,
oh
yeah,
of
course,
node
test,
node
add-on,
api
new,
all
right
build,
and
it's
also
set
to
main
oh
yeah.
This.
This
is
a
done
deal
then.
A
B
A
Let's
see
does
this
thing
is
this
thing
as
smart
as
firefox
all
right?
So
then,
so
you
landed
a
pr
correct.
B
Node
add
on
api.
I
thought.
A
B
A
A
This
is
awesome.
Okay,
then
so
just
yeah
we'll
leave
it
open,
as
michael
said,
but
it
looks
like
we've
made
the
switch
so
then
yeah,
I'm
gonna
have
to
pull
the
main
branch
and
and
start
doing
my
local
repo
as
well
yeah,
okay,
renaming
an
api
to
something
more
appropriate
and
descriptive
yeah.
I
got
a
pr
going
with
that.
It's
not
small!
It's
like
plus
eleven
thousand
minus
ten
thousand.
A
Yeah
yeah,
it's
it's
it's
fun!
Let
me
just
see
here:
let's
see
github
node.js,
node,
gabriel
off
and
this
guy
so
yeah
I
mean
it.
It
touches
a
lot
of
things.
I
tried
to
avoid
anything
that
would
make
it
december
major.
Like
you
know,
you
know
we
have
things
like
we
have
like
process,
dot
versions,
dot,
nappy
and
that
that's
just
gotta.
D
A
So
those
things
have
to
stay,
we
can
we
can,
we
may
be
able
to
mirror
them,
but
I
don't
think
we
can
remove
them
anytime
soon.
You
know
they
are
not.
I
would
say
that
they
are
probably
not
part
of
the
abi
stable
guarantee,
because
that
that
refers
to
the
code,
not
the
the
tooling
right.
A
You
know
if
you,
if
you
update
the
tooling,
then
you
then
whatever
binaries
you
have
produced
so
far
will
continue
to
work
right.
So
so
I
think
that's
okay,
to
mirror
and
eventually
remove,
but
I
haven't
touched
it
for
now.
We
can.
We
can
discuss
that
in
the
in
the
pr.
So
I
think
I
I
mentioned
here-
oh
wait.
No!
I
didn't
because,
like
I
actually
had
to
close
a
previous
pr,
so
that's
a
good.
It's
a
good
point.
A
A
And
I
made
a
comment
here:
I
think
this,
oh,
and
this
is
another
thing
we
got
to
talk
about
so
if,
if
you
can,
if
you
can
think
of
some
names
for
the
macro
that
defines
the
add-on
that
you
know
evaluates
to,
you
know
having
the
you
know
having
the
library
constructor
and
all
that
you
know,
we
have
that
macro
going,
okay,
so
copied
from
the
pr.
A
I
think
yeah
there,
that's
how
I
want
that
to
look
okay,
so
so,
basically,
basically
the
problem
is
that
we
had
we.
We
had
these
these
macros
right,
an
api
module
and
an
api
module
in
it
right.
A
B
B
B
A
So
please
start
bike
shedding
some
names.
D
A
A
Let's
write
them
all
down
and-
and
just
just
please
comment
on
on
this
pr
and
then
I
might
even
put
a
mention
into
to
all
the
collaborators
so
that
we
can
get
some.
We
can
get
some
input
here.
You
know
once
we
have
a
few
suggestions
and
then
let
them
chime
in
and
then
we'll
pick
one
from
there
right
that
doesn't
conflict.
D
A
Have
an
opportunity
to
to
to
do
this
in
a
non-breaking
way,
because
I
mean
there
are
many
names
out
there.
It's
just
hard
to
pick
one
anywho,
so
so
yeah,
that's
that's
where
that's
at
the
the
tests
are
finally
passing
so
I
I
didn't
miss
anything,
but
I'm
I'm
still
sort
of
grabbing
the
repo
to
see.
Where
else
I
can.
I
can
make
changes
that
that
are
not
like
december
major.
A
So
it's
it's
still
kind
of
a.
I
think
yeah,
it's
still
called
whip.
So
so
there's
that
all
right,
so
then
the
talk
submission
I
saw.
I
saw
your
abstract
gym.
I
thought
it
was
great.
Thank
you.
So
here
it
is
yep.
So
everybody
else,
please
read
it.
A
Oh
yeah!
I
should
I
should
add
my
thumbs
up
there,
because
I
made
a
comment,
but
that's
a
nice
way
to
aggregate
responses.
So
how
how?
Where?
Where
does
this
go
like?
Do
you
submit
this
on
our
behalf
and
and
then
we
just
wait
and
see
or
are
we
still
are
we
still
gonna
copy
edit
this
or
how?
What
what's?
What's
the
way
forward
here?
Okay,.
B
I
think
what
we
ought
to
do
is
we
have
two
more
weeks:
it's
it's
due
on
the
15th,
so
we
still
have
two
more
weeks.
I
think
it
would
be
good
to
have
a
discussion
about
who
would
like
to
participate,
because,
when
I
put
in
the
proposal
what
I'll
do
is
I'll
list
the
people
who
will
be
presenting.
A
Oh,
I
see
okay,
okay,
and
I
mean
given
that
by
my
understanding
this
is
not
a
physical
conference
right.
There
should
probably
be
no
limit
right
on
on
the
actual
number
of
presenters,
because
I
think.
A
B
So
I
think
we
could,
I
think
we
have
another
week
to
to
let
people
have
a
chance
to
review
the
abstract
and
then
maybe,
when
we
have
more
of
a
quorum,
we
can
talk
about
who
would
like
to
talk
to
to
different
points,
and
what
we
might
might
want
to
do
for
our
next
step
is
to
create
an
outline
of
what
the
presentation
would
be
like
and
then
who
would
be
presenting
those
major
sections.
A
Yeah
yeah
yeah
yeah.
I
think
I
think
we
did
we
did
well
in
in
in
in
the
way
we
like
put
put
like
people's
like
heads
on
each
slide
and
and
then.
B
Yes,
yes,
and
I
think
the
article
that
the
foundation
published,
I
think,
is
a
really
good
basis
for
the
material
that
we
would
want
to
include
in
this.
In
this
talk,.
A
A
This
can
take
us
to
the
end
of
the
meeting
and
fill
up
whatever
time
we
have
left,
because
I
think
we
have
a
few
of
them.
So,
let's
see
tracking
issues
for
modules,
oh
yeah,
yeah,
yeah,
okay,
so
reporting
news-
I
don't
I
haven't-
had
a
chance
to
to
look
at
this,
but
nicola
would
probably
know
more,
and
so
we
are
left
with
his
comment
from
10
days
ago
about
a
new
module,
but
this
is
also
the
the
canonical
time
to
check
our
status
on
npm
js
right.
A
A
A
Okay,
well,
it
got
merged.
So
there's
that
I
can
always
just
follow
up
again
if
we
need
and
just
do
some
copy
editing,
but
that's
good.
That's
made
progress
all
right,
debug
testing
for
add-ons!
Oh,
I
I
don't
know
this
is
to
michael.
I
I
don't
know.
If
he's
had
any
time
to
look
at
this
all
right,
let's
see,
enabled
twisting
okay
yeah.
So
I
guess
yeah.
Okay!
So
that's
that's
the
ci!
That's
in
his
hands!
Okay,
debug
tests
for
node
add-on.
I
don't
remember,
do
you?
A
A
A
We
don't
really
have
you
know
like
in
node.js
there
they're
like
there
are
these
checks
right
which
which,
which
will
abort,
if
you're,
building
and
debug
right
it's
it's
slightly
more
sort
of
pedantic
when
you
build
in
debug
mode,
but
I
don't
think
we
have
any
any
such
distinction
between
debug
and
release
mode,
so
debug
mode
for
us,
if
I
remember
correctly,
is
exactly
equivalent
to
just
adding
debug
symbols
rather
than
adding
additional
checks,
so,
but
it's
still
good
to
have
a
debug
build
right.
A
Okay,
so
I
think
that's
that's
the
update
on
this
issue
and
then
okay
yeah
a
list
of
issues.
Okay,
so
then
right
so,
let's,
let's
just
let's
just
start
start
looking
at
issues.
I
think
we
should
probably
start
with
the
core
repo,
because
that
one,
if
I'm
not
mistaken,
is
a
little
bit
quieter.
A
So
let's
go
here
and
node
issues,
labels
and
api
yeah.
That's
another
thing:
shall
we
change
this
label
because
you
know
it's,
you
know
we're
gonna,
we're
gonna
lose
we're.
Gonna
lose
some
track
if
we,
if
we
rename
this
label
to
at
least
it
has
a
dash
in
it.
So
there's
that
oh,
this
is
new.
A
D
Right-
and
it
was
pointed
out
that
the
serializations
have
to
inherit
from
this
base
object
or
transferable,
and
then
there
was
like
no
way
to
get
it
to
be
inherited
from
the
javascript
side
or
something
from
my
brief
looking
at
this,
but
I
the
the
request
is
actually
kind
of
nice.
I
mean
it's
like
the
ability
to
serialize
and
deserialize
your
own
object.
Oh.
A
I
see
I
see
what's
going
on
here.
Okay,
I
I
see
the
api
says
it
all
and
that
be
defined:
transferable
class,
okay
or
that's
going
to
be
node
api,
because
you
know
one
thing
that
came
to
mind
when
I
was
doing
this
rename
is
that,
yes,
we
have
symbols
that
start
with
with
an
api
underscore
right,
but
and
we
have
no
choice
but
to
define
macros
that
that
alias
those
symbols,
but
that
do
not
result
in
any
code
that
that
that
has
such
symbol
names
right.
A
It
will
still
be
an
api
underscore
when,
when
it's
all
said
and
done
and
in
the
binary,
but
there
is
nothing
preventing
us
from
having
new
symbols
which
have
the
correct
prefix
of
node
underscore
api
so
going
forward.
We
we
don't
need
to
do
this
macro
thing
anymore
right,
because
there
there's
nothing
that
says
that
symbols
have
to
start
with
an
api
underscore.
So
so
this
is
just
sort
of
you
know
going
backwards
without
being
api
breaking,
but
going
forward.
There's
we
we
should
like
new
symbols.
A
B
A
Yeah,
so,
okay,
so
so,
then,
all
right,
I'll
I'll
go
I'll,
go
over
the
the
docs
again
in
the
rename
and
because
we
do
have
this
guide
to
adding
a
new
node
api
and-
and
I
did
already
rename
everything
in
there
so
I
might
have
just
I
might
have
just
hit
it
by
mistake.
You
know,
or
or
just
as
a
side
effect
and
and
just
address
this
issue,
but
but
maybe
we
should
maybe
we
should
make
it
explicit.
A
Like
maybe
add
a
blurb
to
to
that
that
says
you
know,
new
symbols
shall
have
the
prefix
node
underscore
api
underscore.
Do
something
useful.
A
So,
okay,
so
we
need
more
stuff
than
that,
because
we
okay
new
and
then
what
is
null
pointer
there
all
right,
because
you
know
we
need
like
we're.
Gonna-
have
to
have
c
functions.
The
pointers
to
which
you
pass
into
an
api
defined
transferable
class,
which
probably
call
which
probably
do
some
some
serialization,
but
okay,
all
right.
Okay.
So
this
is.
This
is
a
december
minor
proposal
that
that
we
need
to
work
on
all
right
see
what
else
is
going
on
in
the
core
repo?
A
C
Yeah
I
have
added
the
flag
to
disable
to
disable
the
new
throwing
the
letters
users
can
go
back
to
the
previous
behavior
good
on
breaking
their
code
immediately
either.
As
we
have
discussed
last
time,
the
flag
will
be
eventually
removed
in
the
next
and
next
major
version
right.
So
the
guy
everything
is
already
passed.
So,
okay.
A
Okay,
all
right,
okay
and
and
yeah
okay.
So
then,
when
when
michael
gets
around
to
it,
he
can
probably
dismiss
his
his
his
objection
left
after
he
does
his
review.
Okay,
good
good,
go
ahead!.
A
Yeah,
I
I
can
I
when,
when
I
review
this,
I
can
probably
give
him
a
nudge
like,
like
you
know,
I
can
say
like
it.
It
looks
to
me
like
like
it's
it.
It
now
conforms
to
your
to
your
requested
change,
and
then
you
know
I'll
give
him
some
time
to
to
confirm
that
so
yeah.
I
I
don't
think,
there's
anything
blocking
anymore.
If
we
have
that
flat.
C
One
thing
is
that
I've
I've
been
and
the
flag
is
being
added.
I
have
been
reviewing
what
what
conditions
the
exceptions
were
still
loaded
so
as
for
now
it
seems
that
we
only
solving
the
exceptions
in
the
thread.
Save
function
go
back
so
and
and
it's
vandalizer,
so
the
black
names
currently
is
called
no
force.
Nappy
uncle
exception
policy,
so,
alternatively,
we
can
call
calling
it
a
no
for
snappy,
no
first
there's
a
function,
uncle
extraction
plus.
Oh.
A
I
see
yeah
yeah
yeah,
that's
that's
good!
That's
that's
good
right,
because
threat
safe
function
is
a
is
a
node
api,
specific
thing:
it's
not!
It's
not
present
anywhere
else
right,
so
you
would
have
to
be
using
node
api
for
you
to
be
using
the
threatsafe
function.
Yeah!
That's.
C
That's
a
good
point
and
the
policy
is
enforced.
Luckily,
we
will
not.
We
will
ensure
future
and
addings
will
be
insured
to
not
swallow
the
exceptions
right
so
yeah.
A
Oh,
I
see
I
see
I
see
so
you
want
to
like
future
proof.
It
yeah
yeah,
yeah,
yeah,
yeah,
you're
right
you're
right,
maybe
maybe
it
should
say
like
yeah
yeah.
Maybe
it
should
say
node
api,
then
so
you
know
the
the.
The
thing
is
that
if
I
remember
correctly,
the
only
reason
that
there
are
exceptions
in
there
is
that
the
stack
is
such
that
there
is
no
way
to
to
return
any
kind
of
error.
Indication
right
like
this.
Could
this
could
not
have
been
done
without
with
the
node
api
status.
A
It
could
only
have
been
done
with
the
with
with
an
exception
right,
because
there
is
no
place
where
to
return
a
node
api
status
right,
but
going
forward.
I
think
it's
very
important
for
any
sort
of
node
api
edition
to
to
try
to
avoid
throwing
exceptions
if
at
all
possible
right.
I
I
don't.
I
don't
think
it's
it's
it's
clear
enough
in
the
documentation
that
that
adding
new
node
api
status
is
is
always
preferable
to
to
to
throwing
exceptions
right.
A
There
should
always
be
a
node
api
status,
and
then
we
don't
need
to
worry
about
exceptions,
but
since
we
have
these
exceptions
and
since
there
might
be
more
situations,
as
you
said,
lucas
that,
where
where
these
exceptions
might
be
necessary,
then
then
you're
probably
right.
You
know
here
in
the
in
the
pr
you
might
want
to
comment
on
the
name
right.
So
just.
C
C
I
may
not
be
clear
that
I
was
mentioning
the
command
line.
Flags
name,
namings-
that
you
can
see
the
namings
in
the
review
file
changed
in
the
first
first
file
changes
in
the
top
of
the
page
on
the
top
of
it
on
the
top
of
the
page
and.
C
C
A
Honestly,
honestly,
I
think
I
think
the
the
the
more
generic
name
is
better
because
it
will
catch
future
issues,
although
yeah,
so
we
don't
know
what
the
timing
is
right.
Currently
you're
right
that
the
threat
say
function
is
the
only
one
that
has
this
right
sooner
or
later,
we're
going
to
remove
this
right
so
yeah,
so
so
the
default
will
be
to
to
just
throw
uncaught
exceptions
right,
but
how
soon
will
that
be?
Will
that
be
before
we
drop
this
flag?
A
You
know
like,
let's
say
the
next
situation
that
comes
up
that
that
that
requires
exceptions.
Will
the
next
situation
come
up
before
we
drop
this
flag,
or
will
it
come
up
after
we
drop
this
like?
We
don't
know
right,
so
so
it's
better
to
call
it
node
api!
Well,
okay,
that's!
Instead
of
an
api,
we
can
just
call
it
node
api
to
go
to
go
with
the
new
name,
but
that's
that's
a
minor
thing
right,
but
it's
better
to
to
to
have
it
refer
to
node
api
in
general
and
then
possibly
catch
new
situations.
A
Before
we
drop
the
flag,
then
calling
it
threat
say
function
and
then
have
it
be
misnamed
right,
because
then
you
know
who
knows,
let's,
let's
say
just
absurdly
like
async
work,
all
of
a
sudden
starts
throwing
right
and
then
and
then
you
know
it's
it's
threats
a
function,
but
then
you
have
to
explain
that
it's
not
only
threat
safe
function,
so
so
I
think
node
api
is
is
is
good.
It's
it's
generic
enough
without
being
too
generic.
A
That's
just
my
opinion.
Yeah
great.
A
Okay,
that's
good!
So,
let's
see
we
were
here
so
then,
okay,
I
think
the
rest
is
pretty
okay
yeah
we
this!
This
is
still
in
draft.
A
What
is
this
pass?
Exceptions
to
js
that
are
thrown
by
js
function
in
async
context,.
A
A
A
There
we
go
it's
not
that
important
either
oh
and
close
oops
there.
We
go
all
right.
So
then
issues
here
we
go
okay,
so
I
think
the
rest
are
pretty
much
status
quo
and-
and
we
can
talk
about
them,
but
we
also
need
to
talk
about
the
the
know
that
on
api
repo,
so
let
me
let
me
move
there.
A
Okay,
let's
see
well
grind
detected
memory
leak.
We
have
a
leak,
that's
not
good.
D
D
I
think
if
you
go
scroll
down
here,
I
think
it's
better.
When
we
have
the
new,
I
think
it's
that
instance
accessor
callback,
data,
yeah.
D
A
Yes,
okay,
so
we
create
that
and
we
add
it
to
a
property
descriptor.
That's
right!
That's
right!
That's
right!
That's
oh,
wait!
A
minute!
Is
this
yeah?
Okay?
I
I
remember
having
this
issue
before
this.
This
was
happening
everywhere
before
it
was
just
allocating
and
not
deallocating,
and
then
I
added
this
thing
called
attach
data
that
would
do
that.
A
But
then
it
ended
up
being
such
that
when
you
create
these
things,
you
need
to
pass
in
the
object
onto
which
they
are
attached,
so
that
you
can
attach
like,
like
a
finalizer
to
the
object
right.
So
this
is
an
access
right.
It's
gonna
be
an
accessor
on
on
on
on
a
class
right,
or
it's
going
to
be
an
accessor
on
on
an
instance,
either
way
you're
going
to
have
a
an
api
object
that
you
can
attach
data
to
and
attach
a
weak
callback.
A
Now
the
problem
was
that
that
you
needed
to
pass
the
object
into
the
function
in
order
for
this
attaching
to
happen
right,
whereas
in
this
case
this
is
instance
wrap.
Therefore,
this
stuff
needs
to
be
attached
to
the
this
needs
to
be
attached
to
the
resulting
constructor.
If
I'm
not
mistaken,
because
you
you
yeah
this
instance,
accessor
gets
called
with,
with
with
a
javascript
instance,
but
but
the
accessory
itself.
The
data
for
the
accessor
is
not
instant
specific.
It's
constructor,
okay,
yeah.
A
So
so
then,
basically,
these
these
things
get
called
as
parameters
to
define
class
or
define
add-on
right
in
the
case
of
define
class.
We
need
to.
We
need
to
attach
this
data
to
the
data
that
is
being
stored
inside
the
javascript
class
constructor,
so
that
when
that
thing
goes
down,
this
data
is
released
in
the
case
of
the
add-on.
A
A
D
I'm
sorry
can
we
go
back
to
the
issue
just
so
that
I
can
understand
the
the
fix
that
that
yeah.
D
D
To
the
to
the
class
destructor,
something
that
de-allocates
all
of
those
instances:
okay
and
then.
D
A
A
A
Okay,
so
here
you
go
property,
descriptor
access,
all
right,
so
see
how
you
see
how
you're
passing
in
the
object
here,
because
normally
you
just
return
a
property,
descriptor
right
and
then
this
property
descriptor
will
get
attached
to
this
object
right.
But
but
this
method,
this
method
used
to
be
completely
declarative.
It
would
just
it
would
just
allocate
this
and
then
set
it
as
a
as
an
item
here
right
and
then
and
then
you
know,
the
problem
was
that
this
object
wasn't
getting
passed
in.
A
So
then,
when
do
you
free
this
data
right
like
when?
When
do
you
free
this?
You
don't
have
a
chance
to
to
free
it
at
any
given
time,
because
all
you're
given
is
just
some
primitive
types
here,
like
the
name
and
and
properties,
and
all
that
you
can't
attach
anything
to
those
right,
and
so
the
idea
is
you
pass
in
the
object
here
and
then
and
then
you
have
something
onto
which
you
can
attach
data,
and
so
the
data
that
you
put
that
you
put
here
is
now
attached
to
the
object.
A
So
it's
going
to
go
away
with
the
object
itself
right,
but
evidently
I
missed
a
spot
right
where,
where
you
have
the
the
the
class
or
add-on
property
descriptors,
sorry,
the
instance
property
descriptors,
either
attached
either
belonging
to
javascript
class
or
to
to
the
instance
data
where
you
also
need
to
do
this
step
right
except
there,
you
cannot.
You
have
no
object
that
you
can
pass
in
because
well,
it
hasn't
been
created.
Yet
right,
like
the
javascript
class,
hasn't
been
created
when
you're,
defining
when
you're
calling
this
method
to
define
the
accessor.
A
So
then,
probably
what
what
I
need
to
do-
and
I'm
surprised
I
haven't
done
that
is-
is
to
is
to
iterate
over
the
instances
after
I
have
created
the
class
so
as
to
attach
all
the
data
to
the
class
and-
and
let
me
just
double
check
that
I
haven't
done
that
so
then
we
want
define
class
right.
A
So
then
yeah
see
here
this
should
be
catching
it
actually,
oh
wait!
No,
because
it's
static.
A
B
A
I
see
okay,
so
so
for
the
functions
you
can
just
create
function,
which
is
a
helper
method.
You
know
which,
which
does
the
attaching
for
you
so
for
static,
void,
method,
callback,
wrapper
and
static
method,
callback,
wrapper,
you're,
taking
care
of
right
because
in
in
either
case,
you're
using
create
function
which
does
attach
data
internally
right.
So
so
those
pieces
of
data
are
fine,
but
the
accessor
data.
You
don't
right
so
because
the
accessor,
the
accessors,
you
just
pass
them
in
here
right.
You
have
props
data
and
props
count.
A
Therefore,
after
the
class
is
defined,
you
have
to
once
more
go
over
the
properties
and
look
only
at
the
data
and
make
sure
that
you
attach
the
prop
data
to
to
what
what
is
essentially
the
the
constructor
right,
see
you
you
create
you
create
the
class
here
and
then
you
attach.
A
A
A
A
A
A
A
Because
you
know
there
are
a
myriad
ways
of
calling
this
instance
access
all
right.
Does
it
does
it,
you
know,
does
it
work
with
just
a
plain
object,
wrap
and
then-
and
then
you
know,
this
is
happening
with
every
single
object
rep
or
was
there
a
very
specific
way
that
they
called
it?
You
know
the
the
code
may
have
a
little
may
have
a
little
morsel
in
there
that
that
is
causing
this.
You
know,
whereas
otherwise
it
doesn't
happen.
So
that's
why
I'm
asking
this
question.
A
Okay,
so
then,
where
were
we.
B
I
noticed
the
call
stacks
are
implicate
sqlite3.
A
Okay,
so
so,
okay,
so
it's
probably
it's
probably
that's
that's
a
good
observation.
It
looks
like
this
might
be.
This
might
be
like
a
yeah
instance
rap.
Why
is
it
calling
instant
trap
directly
shouldn't
it
be
calling
object,
wrap?
A
Oh
no,
I
see
I
see.
Okay
in
probably
has
has
like
a
defined
class
statement
inside
it,
because
init
is
the
subclass
of
object,
wrap
that
is
being
created,
and
this
is
the
constructor
for
it
right,
because
it's
n
object
and
it
it.
It
returns
the
constructor,
and
so
so.
This
thing
probably
has
a
defined
class
statement
in
there
and
then
okay
yeah,
so
so
so
kevin.
I
don't
know
if
you
want
to
if
you
want
to
have
a
look
at
at
this
function.
A
B
And
that,
and
just
as
an
aside
that
that's
the
work,
I
did
to
migrate
that
code
from
nan
to
to
an
api.
Okay.
A
All
right,
I'm
just
gonna,
go
backwards
until
I
find
the
list
of
issues
there
we
go
now.
Is
that
where
we
go?
No
all
right.
I
got
lost
in
my
tabs,
all
right,
yeah.
I
think
I'm
just
going
to
start
a
new
one.
There
we
go
all
right,
so
we
had
a
leak.
That's
not
good
to
forward
user
callback!
Okay,
that's
next!
One.
D
Yeah,
I
saw
this
one
too,
and
he
asked
how
to
create
a
pr
so
gave
him
the
steps
to
that.
But
I
haven't
seen
his
pool
yet.
D
No,
he
didn't
create.
He
didn't
link
any
code
that
he
created,
but
this
is
what
he
was
requesting
here
is
to
move
the
callback
instead
of
copying.
It.
A
Okay,
what
are
the
implications
of
that?
Because
it
looks
like.
A
D
Yeah,
so
it's
the
callback
that
you
pass
into
thread
say
function
blocking
call
right
now,
I'm
not
too
familiar
with
copies
and
forwards,
and
autos
and
things
like
that
but
like
if
you
were
to
pass.
D
He
wants
to
add
this
as
a
new
method,
where
you
were
to.
D
D
D
D
A
That's
yeah,
I
don't
even
know
if
this
isn't
just
a
simple
patch
like
we
don't
need
an
another
overload
right,
because
what
what
code
could
this
possibly
break
right,
you
know,
would?
Would
anyone
use
the
use
this
value
after
it
was
already
passed
into
into?
Well,
I
guess
they
might
eh
it's.
A
D
D
A
Yeah
yeah
yeah
yeah
a
small
improvement
that
that
would
probably
address
most
people's
use
case
because
I,
I
suspect,
very
few
people
use
use
the
callback
afterwards,
especially
since
it's
most
often
a
lambda
right
like
a
lambda,
is
just
kind
of
like
it's.
It's
like
it's
basically
a
throwaway
thing
right
I
mean
unless
you
store
it
in
a
variable
and
then
call
it
it's
it's
pretty
much
throwaway.
D
D
He
apparently
made
a
a
a
local
branch
and
yeah.
He
gave
him.
D
A
Okay,
you
made
the
local
branch,
you
know
in
the
worst
case.
In
the
worst
case,
they
could
probably
just
post
like
like,
like
just
just
in
line
just
put
a
put
a
diff
here.
Like
you
know,
you
could
just
do
this
right
right
and
just
patch
in
between
or
just
upload
it
as
a
gist.
A
You
know
if
all
else
fails
and
then
and
then
let's
see,
do
we
have
okay,
so
we
have
their
full
name
and
then
we
can
offer
that
we'll
create
a
patch
with
them
as
the
author
and
then
we'll
do
the
ppr
just
just
to
get
the
code
right.
So
it's
quite
possible.
You
know
you
like
you,
you
apply
the
patch
and
then
you
like
git
commit
minus
minus
author,
and
you
do
this,
though
they
would
have
to
give
you
an
email
address.
A
You
can
you
can
you
can
probably
share
your
email
with
them
and
then
and
then,
because
you
know
your
email
is
your
email
is
already
in
the
github
blogs,
so
you
can
probably
just
share
that
with
them.
You
know.
A
A
A
That's
right,
yes,
wait!
Why
is
this
not
static?
That's
a
good
question
yeah.
Maybe
this
should
be
static.
Okay,
all
right!
This
seems
like
a
simple
fix.
I
I
I'm
trying
to
remember
if
we
had
some
issues
with
with
resolving
the
symbol
if
it
is
static,
but
I
think
I
think
we're
just
passing
this
as
a
function
pointer,
so
yeah.
This
should
probably
be
static.
A
A
D
Thread
and
then
somebody
posted
pointed
them
to
thread
say
functions
as
a
comment,
so
waiting
to
hear
back
about
that.
A
A
All
right
so
no
build
to
compare
stretchy
function
and
oh
yeah,
that's
a
good
one.
That's
a
good
one
yeah!
I
remember
we.
We
had
this
one
last
week
as
well
right,
yeah
yeah
this.
This
would
have
to
be
done
in
core
as
well,
because
you
would
have
to
you
would
have
to
have
a
way
to
expose
the
the
value.
A
There
is
one
place
where
you
can
compare
them
ish
and
that's
in
the
in
the
in
the
marshaller
right
in
the
js
or
call
js
function
right.
You
can
just
pass
in
the
function
and
see
if
that's
the
function,
but
that's
not
too
convenient.
A
All
right,
wrapped
object
generates
seems
to
have
memory
issue.
I
wonder
if
this
is
a
duplicate.
A
A
D
Yeah,
so
I
I
vaguely
remember
this
one
I
sounded
like
he
wasn't
able
to
locate
the
no.
I
didn't
comment
on
this.
D
Yeah,
so
I
have
no
idea
how
he's
trying
to
build
this
at
all.
I
did
try
to
do
the
translate
for
those
symbols
and
I
think
it
was
like
unresolved
symbol
or
something
like
that.
Oh.
A
Oh
yeah,
okay,
yeah
yeah,
yeah,
okay,
so
so
that's
the
yeah.
I
figured
it's
a
linker
error
just
from
the
looks
of
it,
and
this
correct
me
if
I'm
wrong
also
looks
like
a
windows,
because
the
windows
has
these
like.
A
A
That
that
this
was
this
was
ported
to
an
api
yeah,
and
this
was
also
ported
to
an
api
ref.
I
remember:
ref
was
like
a
native
add-on
for
creating
like
like
weak
references
on
the
javascript
side.
So
so
this
is
this,
is
you
know,
as
a
side
effect,
new
information
about
stuff
that
got
ported
to
an
api
yeah,
but
yeah
you're
right
I
mean
I'm
not
sure
how
I
can
help
here.
It
just
seems
like
a
get
clean
and
then
rebuild
might
might
do
the
trick.
A
D
And
then,
with
the
last
couple
of
minutes,
do
we
want
to
look
at
nicola's
node,
api,
headers,
repo.
D
So
I
see
the
we
have
the
include
directory
as
an
export.
Why
do
we
need
to
have
the
symbols
as
an
export.
A
Oh
wait,
let
me
see,
let
me
see
which
file
are
you
looking
at?
You
mean
like
like
an
index,
it's.
D
A
Yeah,
you
know,
and
and
this
list
should
probably
be
generated.
D
Right
that
was
another
thing
that
I
wanted
to
discuss
with.
How
are
the
the
actual
include
files
supposed
to
be
updated
like
should
we
try
to
investigate
in
some
sort
of
automation,
of
updating
those
files
on
every
node
release.
B
A
D
A
B
A
A
A
All
right
so
here
we
go
I'll,
just
add
a
single
comment
and
and
and
we'll
we'll
give
nikola
a
chance
to
enlighten
us
with
with
the
use
case
for
this,
because
I'm
sure
there
is
one
otherwise
he
wouldn't
have
done
it.
I'm
just
curious
what
it
is
honestly
right.
A
Okay,
so
all
right,
so
I
think
we're
at
time.
So
is
there
anything
else
we
should
be
talking
about
before
we
close
up.
A
All
right,
if
not,
then
thank
you,
everybody
for
joining
and
we'll
see
each
other
next
week.