►
From YouTube: 2020-10-02-Node.js N-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
B
A
Okay,
so
the
first
one
on
the
list
is
the
free
seal.
Apis
and
shelley
had
raised
the
pr
in
node
core
or
was
just
looking
for
some
feedback
from
the
napi
team.
A
C
Sure
yeah,
it
basically
just
implements
the
equivalent
of
object.freeze
and
object.co
in
an
api.
Essentially,
it
is
something
that
is
potentially
useful
to
some
modules
in
electron
for
security
purposes,
just
like,
for
example,
being
able
to
pass
objects
across
process
boundaries,
depending
on
how
you're
like
dealing
with
that
kind
of
information.
C
The
one
thing
that
did
come
up
in
here
that
I
would
be
interested
in
y'all's
feedback
on
would
be
the
fact
that,
michael
you
brought
up
the
fact
that
it
would
be
something
or
it
would
be
potentially
interesting
to
try
to
implement
like
is
frozen
and
is
sealed
as
well.
The
reason
I
didn't
initially
do
that
for
context
is
because
v
it
doesn't
directly
expose
a
way
to
test
that.
D
C
Example,
the
way
that
you
freeze
something
or
seal
something
in
v8
is
by
calling
set
integrity
level
with
either
k,
freeze
or
k,
frozen
or
kcl,
or
something
like
that,
but
there's
actually
no
equivalent
get
integrity
novel,
which
is
kind
of
unfortunate.
The
way
that
v8
tests
that
themselves
is
they.
Basically,
I
like
to
test
there,
but
like
the
way
the
v8
tested
themselves.
Is
they
basically
like
compile
like
objects?
C
Dot
is
frozen
and
then
run
it
against
the
results
of
their
call
to
set
at
deputy
level.
So,
okay
yeah.
I
know
if
you
look
at
the
test,
it's
a
bit
unfortunate,
but
short
of
that
is
to
say
that,
like
theoretically,
it's
possible
like,
I
could
probably
figure
out
a
way
to
do
it,
it's
more
if
we
feel
or
if
you
all
feel
that
that
is
like
something
worth
pursuing,
given
the
unfortunate
constraints
that
exist
upon
it.
A
Yeah
I
mean
that
my
the
reason
I
asked
is
kind
of
like
when
I
looked
at
the
specs.
It
was
like
okay,
there's
a
there's,
a
freeze
and
seal,
and
then-
and
it
is,
is
that
and
I
I
can
see
us
getting
questions
like.
Why
did
you
add
one
versus
the
other
and
not
the
other?
And
I
guess
like
you
might,
I
might
have
just
missed
it.
When
you
were
talking
there,
you
you
actually
are
using
frozen
and
sealed
in
an
electron
or
somewhere
else.
C
Yeah
actually
directly
is
it,
and
I
think
that
there
would
be
people
that
want
to
use
it
in
a
similar
way
that
we
do
when
dealing
with
process
boundaries
like
we
do
we
do
we
do
it
or
we
basically
like
perform
deep
freezing
of
a
lot
of
things
in
this
very
similar
way.
When
safely
like,
we
basically
like
safely
proxy
objects
across
process
boundaries.
A
A
A
Right,
okay,
okay,
just
because,
like
one
of
the
things
that
you
know
in
any
api,
is
we're
looking
to
provide
what
people
need
and
use
versus
everything.
Fair
point:
yeah,
okay,
so
I
I
mean
I
think
that
answers
the
questions
I
had
do.
Other
people
have
questions
concerns.
E
Oh
this
is
this
is
part
of
like
ecma
script
right
this
this
this
static
function
on
object
right,
so
then
it
makes
sense
that
we
should
have
it.
C
Yeah
I
like
in
the
documentation
that
I
added
I
like
all
the
locations
that
it
appears
with
spectex.
E
C
E
Right
right,
right,
yeah,
yeah,
okay,
so
so
so
the
interfacing.
Well
I
mean
the
interface
exists
at
the
javascript
level,
but
not
at
the
v8
level.
So
I
mean
you
could
implement
it.
You
could
implement
it
just
by
you
know,
retrieving
the
global,
then
retrieving
the
object
of
the
global
and
then
retrieving
this
method
of
the
object
and
then
calling
this
method.
And
then
you
know
returning
the
result
right.
That
would
be.
That
would
be
the
implementation
of
of
is
frozen
on
the
javascripts
or
on
the
native
side
right.
E
That
would
be
a
massively
sort
of
inefficient
implementation,
because
you'd
have
to
make
a
whole
bunch
of
calls
into
the
engine
just
to
establish
that
exactly.
C
C
E
Because
you
know
anybody
could
do
it
because
there,
if
there
is
a
way
to
do
it
in
javascript,
there
is
a
way
to
do
it
in
in
native,
because
you
just
call
you
just
pretend
to
be
javascript.
So
if
people
really
really
need
that
on
the
native
side,
we
could
potentially
point
them
to
it,
and
if
the
need
is
great
enough,
then
we
could
do
that
whole
inconvenient
thing
for
them,
eventually
on
some
later
version
of
an
api,
so
that
you
know
our
argument
is
not
that
we
cannot
do
it.
C
E
A
E
A
C
Here
for
more
context,
if
anyone's
curious,
I
put
a
link
to
the
way
that
v8
tested
in
this
zoom
chat.
D
Yeah,
I
have
some
questions,
but
my
question
is
not
blocked
for
this
one.
Just
to
generalize
the
question
to
how
how
do
we
define
the
a
guaranteed
on
those
ecommerce
script?
Apis,
like
like
say
we
have.
There
are
no
api
guarantees
on
ecosquid,
so
the
api
regarding
the
javascript
objects
like
object,
javascript
api-
they
were
probably
might
make
it
upgrade
like,
say,
adding
new
options
and
adding
new
api
shapes
like
so
those
javascript
we
have
corresponding
in
mk.
We
do
have
to
provide
api
guarantees.
D
So
I'm
thinking
about
that,
if
we
are,
how
do
we
handle
this
api
changes
regarding
to
the
ecmascript
specification
another
broadcast
for
this
one?
I'm
just
curious
about
the
this
topic.
A
D
Yeah,
I
I
I
recently
have
a
proposal
to
ecoscope.
Let's
say
we
add
a
new
parameter
to
the
error:
constructor
the
parameter
just
adding
adding
a
cost
to
the
other
prototype
and
there
are
instance
itself.
So
next
it
is
very
safe
to
add,
adding
a
parameter,
optional
parameter
to
the
javascript
api.
But
we
don't
have
this
ability
to
add
those
parameters
in
api,
but
they
are
still
ready.
The
ecmascript
portfolios.
D
E
D
Yeah
just
curious
about
your
topic
burn
not
specific
to
this
playoffs,
so.
E
Yeah,
I
I
I
think
I
think
you
know
the
the
only
concern
that
I
that
I
can
imagine
us
having
is
if,
if
ecmascript
themselves
break
the
spec
somehow
and
make
it
backwards
incompatible,
but
like
that's,
I
I'd
rather
play
the
lottery.
Then
you
know
then
do
that
because
I
I,
I
suspect
very
strongly
that
that
you
know
breaking
the
entire
javascript
ecosystem
is,
is
not
something
they
want
to
do
very
lightly
unless
there's
a
super
super
good
reason
for
it.
D
Yeah
this
not
broken
now
break
the
web
is
the,
I
think,
is
the
core
of
the
t-39,
but
there
are
conditions
that
there
might
be
a
small.
D
I,
I
think,
a
proposal
that
will
change
the
changes.
Behavior
some
minor
javascript
part
like
for
security
for
security
reasons,
yeah
okay,
there.
E
F
E
Yeah
I
put
a
big
fat
warning
in
the
in
the
documentation
saying
that
you
know
refer
to
this
proceeding
of
of
ecmascript
and
and
and
avoid
using
this
api,
but
I
mean
it's
still
there
because
I
mean
people
had
been
using
it
right,
so
yeah
yeah
I
mean
the
deprecation
is
not
is
not
out
of
the
question.
I
don't
think
we've
had
any
deprecation
so
far,
but
but
yeah
it's
it's
not
we
can.
We
can
definitely
add
deprecation
if
needed.
I
think
right.
A
Okay,
so
it
sounds,
it
sounds
like
there's
no
concerns
everybody
agrees.
It
makes
sense
to
add
this,
given
that
it's
it's
actually
being
used
and
is
needed,
I
can
see,
there's
maybe
a
question
or
two
in
the
pr,
but
otherwise
I
think
the
feedback
you're
looking
for
shelly,
I
think,
is
yet
sounds
like
the
team's
on
board
and
just
needs
to
go
through
the
regular
process.
C
Well,
thanks
y'all,
actually
last
question
for
the
question
that
I
see
above
there
is,
if
it
fails,
my
original
presumption
was
that
it
should
just
return.
The
original
object
like
it
should
just
return
the
object
unchanged
with
the
status,
noting
that
it
failed.
A
C
A
C
That's
a
good
question:
let's
see,
I
don't
recall
reading
that
in
the
spec
when
I
was
looking
at
this,
but
I
can
double
check
because
it
is
likely
that
these
that's
fine.
A
A
C
Api,
it's
just
the
status
right
now,
yeah.
E
So
so,
if,
if
okay,
so
if
it
may
throw
an
exception,
let
me
see
your
implementation.
Okay,
you've
got
nappy
preamble,
so
that's
good
okay.
So
then
let
me
see
check
to
object
return
status
if
false,
okay,
just
it's
sort
of
a
as
an
as
an
implementation
thing.
If
you're
using
nappy
preamble,
then
then
please
have
a
look
at
the
the
macros
you're
using
like
check
to
object
and
return
status
if
false,
because
we've
slowly
been
introducing
equivalence
of
most
of
our
macros,
like
that.
E
That
say
with
preamble
right,
so
we
have
like
a
macro
x
and
then
we
introduce
a
macro
x
with
preamble,
because
in
the
cases
where
you
do
have
preamble,
the
preamble
consists
of
of
basically
a
try
catch.
E
So
so
then,
if
you,
you
know,
if
you
can
return
at
multiple
places
from
from
the
from
the
api,
then
at
each
of
those
places
you
have
to
check
whether
an
the
try
catch
has
caught
something
first,
because
if
you
don't,
you
will
still
return
an
error
status,
but
but
people
may
want
to
handle
the
the
exception,
and
so,
if
you,
if,
instead
of
returning
the
error
status,
you
return
that
be
pending
exception,
because
your
your
try
catch
has
caught
then
then
that's
a
that
better
informs
them
that
that
there
is
an
exception
and
they
don't
need
to
call
is
exception.
E
Pending
anymore,
they
can
just
go
straight
for
getting
clear
exception.
So
there
are
examples
where,
where
all
these
macros
are
defined
for
for
like
something
underscore
with
preamble,
so
I
don't
think
we
have
a
check
to
object
yet.
But
if
check
to
object,
returns
early,
then
then
well
or
actually.
E
Okay-
okay,
that's
good!
That's
good!
Yeah!
Okay,
though,
although
I
think
in
this
case
it
is
not
possible
to
throw
an
exception
before
you
actually
call
set
integrity
level,
so
basically
use
your
judgment,
but
but
after
any
point
in
the
body
of
the
function
where
an
exception
may
have
been
thrown,
please
use
the
with
preamble
things
to
bail
out,
because
then
that
will
return
nappy
pending
exception,
rather
than
the
nappy
status
that
you
wish
to
return.
E
Okay,
good,
so
so,
then,
then,
please,
please
change
that
to
return
status
it
falls
with
preamble
and
and
then
that
should
be
good,
because
the
the
the
initial
check
to
object,
I
think,
is
fine
because
no
exception
there
was
no
opportunity
to
throw
an
exception
at
that
point,
so
it
should
be
fine
to
just
check
to
object
and
then
yeah
just
change
that
to
with
preamble,
and
I
think
the
other
I
think
for
the
other
api.
The
same
thing
applies
because
it
it's
very
similar.
E
E
A
Okay,
so
let's
move
on
to
the
next
issue:
this
is
new
version
of
v8,
causing
failures
and
test
objects.
I've
been
spending
some
time
this
week.
None
of
us
could
recreate
it
on
our
own.
A
I
ended
up
clearing
some
workspaces
and
I
think
the
last
one
I
cleared
may
have
done
the
trick
because
the
run
the
job
that
I
ran
after
that
seems
to
have
not
failed.
I
was
hoping
that
we'd
have
a
regular
run
today,
which
didn't
run
so
I've
kicked
off
another
another
run
myself.
Let's
just
see
here.
If
it's
still
running
last
time,
I
looked
but.
A
Looks
like
it's
still,
maybe
still
compiling
away
anyway,
I'm
hoping
that
that
will
be
resolved.
Having
done
some
cleanup
versus
it
actually
being
a
problem.
E
A
A
E
A
A
Yeah,
so
that
I
mean
that's
the
only
kind
of
thing
I
can
think
of.
If
provided
that
you
know
this
really
has
fixed
it,
but
it
it.
You
know
having
basically
taken
the
exact
same
job
and
said:
okay,
let's,
let's
run
it
and
I
tried
doing
things
like:
let's
run
the
the
non-canary
version
like
the
regular
version,
because
the
jobs
are
normally
interspersed
when
they
run.
A
You
know
we
run
mostly
from
you
know,
say:
master
regular
node,
and
this
runs
with
node
v8
with
the
canary
branch.
So
I
tried
like
well,
let's
first
try
running
with
the
node
one
and
then
running
the
canary
one
and
then
back
and
forth,
and
nothing
seemed
to
recreate
so.
A
A
Hopefully
the
clearer
has
has
done
it
then
we'll
we'll
know
that
within
a
well,
hopefully
today
sometime.
If
so,
I
will
close
that
one
out
any
pi
7
this
one.
I
think
we
decided
to
leave
open
just
as
you
I
know,
you're
working
to
flow
it
back
through
the
different
versions.
E
Yeah,
I
think
I
think
12
has
it
now,
so
it's
landed
on
12..
We
just.
We
just
need
two
pr's
on
10
and
I
think
we
should
be
golden
after
that.
A
A
B
I
added
one
there,
michael
oh
good.
What
was
that
so
mateo
has
this
histogram
library
on
a
rickety
old
man.
It
was
like
almost
prehistoric,
and
so
it
is.
It
is
now
completely
modern.
It
is
completely
up
to
using
pre-buildify
and
github
actions.
So
I
I'm
really
pleased
with
the
work
that
I
was
able
to
do
on
this.
One
excellent
and
the
code
was
generated
from
my
automated
tool,
so.
B
A
B
If
you're
interested
to
see
what
that
code
looks
like
take
a
look
at
that.
A
F
B
Or
not,
I'm
I'm
still
keeping
it
I'm
still
keeping
it
private
nicola
until
I
can
get
it
farther
along.
So
it
is
not
open
source
at
this
time.
A
Awesome
job
yeah
noted
on
api
watch
still
like
we're
over
2
million,
which
is
quite
good.
I
guess
that's
one
thing.
We
should
talk
a
little
bit
about
I
forgot
last
week
because
we
should
talk
about
the
blog
post
that
we
were
going
to
put
together
right,
but
I
think
we
have
that
in
our
list
here.
I
just
missed
it
the
last
time,
so
we
got
that
node
conf
workshop
proposal
sounds
like
jim.
You
have
the
answer.
A
We'll
we'll
keep
it
in
mind,
for
you
know
other
places
that
we
might
deliver
it.
So
I'm
going.
A
A
Else
has
right
so
maybe
before
we
dive
into
the
I'm
going
to
move
this
up,
because
this
is
what
happened
last
time,
kind
of
forgot
it
before
we
move
on
to
the
burn
down
the
list
of
issues
here,
nick
opened
this
blog
post.
Maybe
we
can
spend
a
few
minutes
kind
of
brainstorming.
What
we
think
the
flow
could.
A
Sorry
yeah
nicole,
can
you
give
us
access
to.
G
Yeah
yeah,
just
just
watching.
A
G
A
G
But
it's
it's
a
blog
post
about
know
the
dom
api
or
an
api
in
in
in
general,.
E
Nicola
may
I
also
have
edit
access
to
my
personal
account,
so
I
think
I
sent
you
a
request
for
it.
G
G
G
E
Yeah
so
here
we
can
put
context,
sensitivity
and
like
this.
F
A
A
E
G
G
A
E
No
okay,
okay,
okay
context,
sensitivity,
okay,
pools
tools
to
support
making
native
add-ons
oops
context
sensitive.
How
about
that.
A
E
H
Yeah
I
brought
up
this
the
documentation
and
examples
because
didn't
we
do
a
lot
of
work
to
migrate
those
over
and
get
them
consolidated,
but.
G
E
I
think
I
think,
like
things
like
pre-build
defy
and
and
all
that
support
that
we
should
mention
that
that
makes
it
way
easier
to
to
maintain.
A
D
A
G
A
E
B
B
B
A
G
G
G
I
can
you
cannot
be
gripped
that
now
is.
It
is
an
api.
G
I
I'm
finding
the
project
that
the
model
that
have
a
good
number
of
downloads
per
week
so.
G
Yeah
because
500
000
download
per
week
and
sharp
a
80
1900.
A
Was
that
that
was
100
like
100
per
week.
A
B
Okay,
yeah
yeah
yeah
esql
light
3
has
283k.
G
Okay,
you
missed
the
key
sorry
sql
283
key,
because.
H
A
So
in
terms
of
writing,
the
blog
post,
like
this
seems
like
a
good
start
and
an
outline
we
can
always
add
more
stuff.
Does
anybody
want
to
volunteer
for
different
sections.
D
H
A
H
A
Just
looking
how
we
so
that
leaves
like
I'm
happy
to
look
at
the
intro.
A
Okay,
so
that
sounds
good.
Why
don't
we
like
each
try
and
work
a
little
bit
on
those,
and
then
we
can
come
back
and
review
it
next
week
to
see
what
we've
got
and
sort
of
put
it
pull
it
together
as
an
overall
thing.
So
we.
A
G
A
A
Okay,
so
I
think
that's
a
good
start
on
that
one.
Now
we
can
just
flip
over
to
looking
at
issues
that
may
be
in
the
repo.
We
need
to
talk
about
no.
G
H
G
H
Saw
gabe
that
you
had
made
some
reviews
come,
and
so
I
guess
we're
doing
back
and
forth
on
that.
So
I'm
aware
yeah.
E
H
Okay,
is
it
still
worthwhile
to
create
that
test
that
you
mentioned
the
ref
unref
tests,
though.
E
Yeah
yeah,
probably
yeah,
if,
if
we
don't
have
it
for
for
thread,
say
function
ex
then.
Definitely,
though,
though,
that's
one
of
those,
I
think
where
you
have
to
spawn
a
process
because,
like
you,
know,
you're
you're
influencing
the
event
loop
there.
I
don't
know
if
you
can
do
that
in
the
same
process
as
the
test,
but.
H
E
H
Not
at
all,
okay,
so
that's
that's
that
and
then
cleaning
up
the
the
text
and
whatnot
and
the
documentation
yep.
I
got
that.
A
Yeah
that
sounds
good
because
yeah
it
would
be
good
to
get
that
out.
I'm
just
yeah.
I
think
we
look
through
there's
nothing
else
here.
That
really
should
be.
You
know
we
should
do
that
before
the
release.
So
if
we
get
that
in
and
then
doing
the
release
release
right
after
makes
sense.
A
A
F
A
Okay,
that's
great
because
that's
one,
I
think
if
you
do
that,
the
rest
of
the
team
can
can
prioritize
looking
at
it,
because
it's
related
to
async
resource
moving
forward.
I
think
it's
kind
of
sort
of
called
it
so
napi,
something
like
yield.
That's
not
new!
So
I,
unless
there's
something
there.