►
From YouTube: 2021-02-26-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
A
B
A
So
the
first
one
is
related
to
renaming
to
node
api.
I
think
jim
you've
been
doing
some
work
on
that
front
recently.
Yes,.
C
I
have
a
pr
open
on
abi,
stable
node,
and
I've
received
comments
both
from
legendicass
and
nicola,
that
I
should
shadow
the
badge
images.
So
we
have
a
set
of
badge
images
right.
Last
week
we
had
talked
about
not
renaming
them,
but
the
suggestion
is
that
we
create
a
shadow
set
in
the
same
directory
that
are
not
named
an
api
but
node
api
right
and
then,
as
we
move
forward,
we
would
add
those
yes.
A
A
A
D
D
D
Do
we
need
to
communicate
these
change?
No,
do
we
need
to
to
write
a
a
blog
post,
for
example?
I
don't
know
so
just
to
to
say
from
now
on,
an
api
will
be
called
node
api
right
so
yeah
to
to
keep
everyone
updated
about
this,
to
avoid
confusion,
because
maybe
so
someone
could
think
that
no
dpi
is
a
another
api
right
and
we
just
rename
it
to
follow
our
policy.
A
D
D
Well,
we
will
start
thinking
about
about
this.
I
think,
if
you
agree.
D
E
A
A
A
A
A
A
D
A
A
A
F
A
Okay,
tracking
issues
for
modules
we've
noticed
reported,
I
don't
know
if
there's
anything
new,
there
don't
see
anything
new,
but
we
typically
also
take
a
look
at
10
pm.
A
A
E
E
Come
up
with
a
previously
still
issue
in
another
add-on:
api?
Okay,
I
noticed
that
it's
been
checked
out
in
the
series
list,
but
I've
also
made
a
acceleration
to
the
the
number
is
301.
E
So
I
it's
yeah
right
yeah,
so
currently
we
the
most
most
most
of
the
callbacks
in
our
no
dialogue
api,
it's
not
being
moved,
then
let's
say
this
callbacks
is
being
copied
so
that
we
cannot
use.
We
cannot
capture
nappy
reference
in
those
lambdas,
so
we
just
make
those
callbacks
being
moved
around
that
it
says
those
lambdas
not
just
not
copying
them,
those
lambdas.
E
We
can
capture
reference
in
the
lambdas,
so
the
top
the
example
of
this
issue.
Top
of
the
top
of
this
issue
can
work
with
this
work
with
the
solution
I
just
have
submitted
it.
E
E
It's
just
far
too
much
burdensome
to
implement
such
a
such
a
generic
solution.
To
that,
so
I
just
hoping
that
we
can
make
those
apis
exposed
by
another
item.
Api
can
receive
non-copyable
lambdas,
so
that
may
is
the
solu
is
the
condition
that
the
developers
may
face
right.
Does
this
potentially
break
any
existing
code?
E
Absolutely
not
it's!
It's
we,
the
the
change.
I
believe
it's
fully
compatible
with
the
previous,
since
the
std
move
is
just
a
non-op,
oh
on
runtime.
So
there
is
nothing
changed
on
random,
but
it's
changed
on
the
build
time
to
like
next
day
we
move
the
van
life
callback
and
we
blue.
We
are
not
using
it
in
the
same
scope,
so
there
is
nothing
changed
regarding
to
the
api
service,
so
I
blue.
It's
it's
fully
compatible.
E
Is
that
the
yeah
it
moved
from
the
let's
say
the
object
at
the
finalizer,
the
scope
of
the
object
finalizer
to
the
details,
panelized
data,
so
there
is
no
copy
during
the
construction
of
finalized
data.
So
those
are
the
reference
in
captured
by
the
lambda
is
not
copied
to
in
this
condition.
So
the
compilation
can
the
compiler
is
satisfied
by
this
known,
runtime
np,
but
they
can
compare
and
work
like
we
expected
okay.
So
if
you
try
to
then
use
it
after
the
move,
the
finalized
callback
will
be
a
empty
one.
E
Like
say,
a
reference
captured
lambdas
to
the
node
and
api,
so
this
change
does
broaden
the
api
service.
Okay,
that
previously
we
not
copied
yet
right
so.
E
Since
previously,
we
just
not
allowing
the
non-copyable
language
to
be
busting,
so
it
will
not
be
will
not
compile
previous
without
this
change
right
so
but.
A
E
Yeah
for
for
the
autoscope
that
it
says,
vandalized
callback
is
being
copied
into
the
object
at
finalizer,
so
previously,
so
for
autoscope
of
the
penalized
callback,
it
will
not
be
will
not
be
affected
by
the
change,
since
the
finalized
callback
is
being
copied
into
the
scope
or
object
at
the
finalizer.
E
A
E
The
other
scope
have
to
give
up
the
ownership
of
the
lambdas
to
get
the
reference
captured
lambdas
to
be
acceptable
acceptable
by
the
node
add-on
api
service
right.
So
I
believe
it's
totally
okay,
since
the
developer
has
to
explicitly
explicitly
obtain
with
the
html.
If,
if
the,
if
their
callback
is
not
in
place,
initialized.
E
E
It
were
always
complaints,
but
if
they
explicitly
use
more
semantics,
then
it
will
compare
with
with
the
pr
but
right
so.
A
E
A
A
A
E
Talked
discussed
the
issue
on
the
meeting
that
there
is
a
crash
on
the
now
that
I
used
to
see
the
number.
E
E
E
B
E
E
Actually,
we
can
handle
exceptions
with
all
javascript
values
in
pure
c
apis,
but
the
node
addon
api
will
create
a
reference
on
the
on
the
chord
exceptions.
So
it
will
fail
on
such
actions
and
it
will
consider
it
a
fatal
action,
since
there
is
nothing
to
do
when
we
fail
on
error
handling,
routine.
E
C
E
I
just
noticed
this
behavior.
In
previously,
we
just
previously
just
crashed
on
teardown
of
the
process.
Addon
and.
E
I
I
think
that
the
gabriel
that
will
help
bring
up
a
issue
that
the
code
error
will
be
released
by
intermediate
endoscope
nappy
endoscope,
so
that
we
have
to
create
a
reference
on
them
so
either
or
not.
We
are
creating
a
new
class
like
a
new
class
for
exceptions.
We
have
to
determine
that.
E
A
Talking
primitive,
primitive
references,
you
mean
like
an
integer
or.
E
Yeah
but
the
primitive
values
they
are
not
always
always
encoded
into
the
local
handles
like
say
a
dynamically
concatenated
string.
They
have,
they
will
be
released
as
or
the
hip
will
explode
with
a
outer
memory.
So
if
we
are
not
able
to
reference
those,
let's
say
a
dynamically
constructed
string,
we
are
not
able
to
pass
it
around
with
c
plus
plus
negative
exceptions
right.
E
The
issue
gabriel
just
brought
up
is
about
the
chocolate
core
and
I've
been
using
javascript
previous
days,
and
I
feel
the
strings
have
been
have
been
able
to
be
referenced
since
or
there
will
be
a
hard
development
like
say
we
cannot
reference
the
dynamically
constructed
string
and
we
just
enabled-
or
we
can
create
a
temporary
object,
and
this
object
will
hold
those
primitives.
A
A
E
Need
to
be
an
object
in
the
noppy
create
reference.
There
is
a
check
that
if
the
value
is
not
object
and
or
is
not
a
function,
it
will
throw
a
not
p
object,
expected
statement.
F
I
don't
I
do
know
that
in
v8
only
you
can
only
have
references
to
objects
like
the
okay,
so
that
would
explain
why
I
guess,
like
at
the
sea
level,.
E
If,
if
we
are
using
non-addon
api
without
c
plus
plus
exception
enabled
then
it's
totally
safe
to
passing
the
exceptions
around,
but
the
c
plus
plus
exceptions.
If
it
is
enabled
and
the
value
the
javascript
exception
is
being
rolled,
then
we
are
unable
to
escape
it
from
outer
endoscope.
A
A
E
A
E
Artificial
temporary
objects,
it
we
have
to
unwrap
before
we
actually
throwing
it
back
into
engine
yeah,
but
sometimes
it's
just
a
relaxer.
We
are
just
get
the
error
and
throwing
back
to
engine
like
say
in
this
case
it
will
be
burdensome
to
just
create
a
another
object
for
it.
So
I'm
not
sure
if
we
are
going
to
go
into
this
part
or
if
we
can
figure
out,
if
let's
say
escape
from
the
henoscope,
I'm
having
test
tests.
A
E
Yeah
it
it's
currently
about
the
process,
so
there
will
be
a
possibility.
You
know,
of
service
attack
on
this
kind
of
thing,
if
there,
if,
if
the
code
is
written
and
throwing
with
and
possibly
arbitrary
drug
is
good.
A
Yeah,
it's
definitely
not
a
so
okay,
but
that's
I
yeah.
I
can
only
think
of
like
okay.
If
we
wrap
it
or
like
you
said,
we
do
something
where
it's.
If
there
is
some
way
to
escope
escape
it,
that
would
be
good.
A
Yeah
right,
so
that's
where,
like
you
know,
legends
cast
was
mentioning.
If
there's
a
way,
we
could
actually,
when
you
exit
a
handle
scope
like
on
purpose,
there's
a
way
to
say
and
here's
something
I
want
to
escape.
A
One,
it
sounds
like
you're.
You've
got
a
few
more
things
to
to
sort
of
try
out
before
you
have
a
recommendation.
A
I
think
I
I'd
like
to
call
out
this
one.
The
vs
2013
compiler
support
the
main
thing
I
want
to
call
our
attention
to
is
this
document
which
I
wasn't
aware
of
so
there
was
another
document
that
so
our
building
document
is
one
I
always
think
about,
which
is
this
one?
That's
master
building
and
it's
it's
the
one.
G
A
A
I
always
thought
like,
as
long
as
we
were
supporting
the
compiler
that
was
used
to
build
node
we'd
be
okay,
but
this
kind
of
says
no.
We
should
have
you
know
we
should
still
be
making
it
so
that
you
can
build
add-ons
with
2015
versus
having
to
move
up
to
2019.
A
C
What
was
the
issue
about?
They
were
actually
asking
for
2013.,
oh
okay,
okay,.
A
C
A
Said
well
looking
at
the
supported
versions,
you
know,
you
know
you
need
at
least
2017..
I
see
I
see,
and
you
know
richard
pointed
out-
well
wait
a
sec
there's
this
other
doc.
That
actually
would
say
the
answer
is
2015.
C
A
Now,
either
way
2013
isn't
supported,
but
it
means
that,
like
we
probably
don't,
have
any
testing
that
tests
with
2015.
right
yeah.
You
know
our
are
having
10
in
the
the
matrix
make
sure
we
have
2017
testing,
but
not
2015
testing
right.
D
What
I
know
is
that
from
2015
yeah
you
have
the
binary
compatibility
before
this
is
not
guaranteed
so
yeah.
If
you
could
use
the
library
compiled
with
vs
compil
visa
studio,
compiler
2015
in
your
new
studio
2019
without
problem,
but
yeah.
There
is
a
problem
with
2013,
because
the
microsoft
does
not
assure
the
credibility.
B
This
is
what
I
I
know:
okay,
but
that's
actually
a
slightly.
A
Different
problem,
in
that
it
like
this
is,
we
could
introduce,
we
could
start
using
some
new
compiler
or
language
feature,
that's
only
available
in
19
and
then
you'd
no
longer
be
able
to
compile
or
say.
Today
we
could
introduce
something
that
used
a
feature.
That's
only
available
in
2017
not
available
in
2015,
and
we
probably
you
know
so
that
you
would
no
longer
be
able
to
compile
your
add-ons
with
2015.
E
A
Plus
feature
or
something
like
that,
so
I
I
think
it's
I
just
wanted
to
make
it
so
that
we're
aware-
and
maybe
we
should
think
a
little
bit
more
about
then
maybe
we
need
to
add
some
testing
on
2015
or
at
least
keep
that
in
mind.
E
A
E
Yeah
great
scent
is
that
that's
right,
nichola
right.
D
Yeah
yeah,
maybe
later
I
can
search
for
microsoft,
documentation
where
I
I
I
read
this,
but
but
yeah
essentially
from
the
200
2015
yeah,
you
can
yeah.
You
have
the
binary
compatibility,
so
you
can
use
the
library
compiled
in
the
previous
version
in
the
new
version
without
without
problem,
but
these
is
not.
D
Valid
for
the
pre,
the
version
previews
of
2015.,
so
yeah
2000.
You
cannot
use
the
library
compiled
with
2013
in
a
project
compiled
with
2015,
for
example,
because
yeah
you
don't
have
the
binary
compatibility
sure
by
yeah.
So.
D
Before
maybe
it
seems
to
work,
but
you
will
get
some
strange
error
at
runtime
yeah.
This
is
the
problem.
Okay,.
A
D
Yeah
I
before
I
won't
discuss
about
the
nabi
leaders,
because
I
am
working
on
that
and
I
saw
your
adjuster
and
yeah.
I
just
yeah.
If
you
can
go
on
my
request,.
B
D
Yeah
so
now
yeah
after
your
answer,
I
under
I
agree
with
you.
So
it's
not
the
case
to
include
the
experimental
function,
and
so
what
I
need
to
do
is
to
remove
from
leaders
what
is
protected
against
the
not
experimental.
Is
it
right.
D
Okay,
so
and
then
I
can
export
the
symbol
using
javascript
like
javascript
object,
like
I
reported
in
the
top
of
the
the
document
right
this
one.
B
D
So
yeah
gamblers
say
that
we
have
two
kind
of
api.
One
is
a
javascript
bbi
and
you
know
there
is
a
api
for
node.
So
yeah
we
have
some
symbols
that.
D
For
the
justice,
avi
and
yeah
other
four
node
api,
so
we
can,
we
can
choose
to
to
export
the
the
the
symbols
per
node
api
version
and
yeah
for
j
javascript
and
for
node.
So
if
you
agree
with
this,
I
can
go
forward
and
yeah.
I
can
complete
my
request
for
the
next
week,
so
yeah
that
looks
good.
A
G
A
A
Okay,
for
me,
it's
okay,
so,
okay,
good
any
any
other
key
things
that
we
should
touch
on
before
we
move
into
the
private
section.
A
A
A
So
I
would
say
just
send
me
an
email
if
you're
interested
and
we'll
go
from
there.
Okay,
I
will
say
then:
okay,
thanks
for
everybody
who
is
watching.