►
From YouTube: 2022-04-08-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
yeah.
Welcome
to
the
node
node,
add
on
api
team
meeting
for
april
8th
2022
we'll
follow
our
standard
approach
of
taking
a
look
at
the
things.
We've
had
tagged
for
milestone,
11.
A
Stop
that
and
so
before
we
get
started,
anybody
have
any
announcements
or
anything
that
they'd
like
to
talk
about.
A
Okay,
the
first
thing
I
think,
just
before
we
lunch
into
the
regular
thing
is
I
did
want
to
before
I
forget
mention
you
know.
The
collaborator
summit
is
coming
up
in
june.
I
don't
know
if
how
many
people
are
going
to
be
able
to
make
it
there,
but
I
just
wondered
if,
if
we
should
think
about
having
a
session
in
the
collaborator
summit
or
something
like
that,.
A
It's
well
it's
so
there's
a
collaborator
summit
after
opengs
world
which
and
there's
two
days
and
so
the
node.
The
there's
space
set
aside
for
the
node
project,
so
there'll
be
topics
on
a
number
of
different
things,
but
we
we
could
propose
a
like
a
one
hour
session
or
something
like
that
related
to
node,
add
on
api,
or
it's
basically
up
to
you,
know
cloudberry
collaborators
to
propose
what
would
make
sense
to
have
a
session
and
talk
about
on,
but
the
different.
The
main
difference
between
what
we
do
week
to
week
would
be.
A
A
There's
the
link
to
the
conference
itself
and
sorry
there's
the
two
days.
The
first
two
days
are
the
actual
conference
and
then
the
following
two
days
are
like
a
collaborator
summit,
not
just
for
node
but
for
the
other
opengs
projects,
but
I
do
know
that
we
have
like
two
rooms
booked
for
those
two
days
for
node.
So
we
can
have.
Quite
you
know
a
two-day
node
collaborator
session
set
of
sessions
that
we
can
fill
in
and
propose.
A
D
C
I
briefly
looked
at
this
issue
about
nam.
Okay,
sorry,
I
didn't
go
in
details,
but
what
I
understood
from
nan,
what
they
do
they
mentioned
his
macro,
but
macro
is
nothing
more,
but
just
they
kind
of
hide
in
this
signature
of
of
the
getter
and
the
signature
of
the
getter
has
additional
property
so
think
about
our
getter.
Signature
has
only
kind
of
a
callback
info
right.
It
pretty
much
has
all
this
information.
C
What
type
of
arguments
were
receiving
and
stuff,
but
in
none
they
have
additional
parameter,
which
is
of
type
of
kind
of
a
v8
string.
I
mean
whatever,
like
wrapped
up
v8
string,
which
effectively
a
property
name
and
what
this
developer
they
asking
about.
Can
they
access
this
property
name
inside
of
getter
function?
So
if
you,
if
you
look
on
the
example,
they
show
like
a
they,
have
its
property
property
parameter
in
in
the
middle
line.
Right.
It's
you
see.
C
So
what
happened
like
this
non-getter
macra
inside
has
this
property
parameter.
This
property
parameter
name
is
a
type
of
string,
as
you
see
here,
like
the
converting
detail,
like
the
string
type,
the
converting
to
none
string
using
it.
Somehow.
A
C
Think,
yes,
yeah,
that's
interesting
point,
so
we
effectively
today
are
required
to
have
five
different
static
methods
for
all
our
properties,
but
in
their
case
they
can
have
one
static,
static
method,
sorry
static
method
which
they
can
share
across
all
properties.
A
C
So
what
happened
like
we-
I
don't-
I
don't
see
like
in
our
both
apis.
Like
I
see
a
lps
and
c
plus
plus
low
api,
we
don't
have.
We
don't
define
property
itself
instead
of
we
defined
like
property,
descriptors
and
descriptor
technically
receives
property
name
right
yep,
but
our
whatever
signature
files
function.
What
we're
passing
it
does
not
receive
this
property
name.
C
Yeah
we
can,
we
can
probably
do
go
to
c
api
like
in
a
core
just
a
kind
of
capability
of.
C
C
No,
no,
we
need
to
look
at
define,
create
property,
descriptor
property
descriptor,
something
about
problems.
C
So,
whenever
creating
proper
descriptor,
we
do
have
name
and
we
have
and
stuff.
But
the
question
is
like
in
their
case,
like
we
see
an
api
callback
in
some
sense,
receiving
this
property
name
right
and
I'm
not
sure.
If
we
do
it.
A
C
No
callback
was
still
there
in
in
types.
Sorry.
C
A
C
So
I
think,
like
I'll
call
back
in
for
not
supposed
to
have
it.
A
A
C
See
yes,
but
among
these
arguments
we
don't
have
property
name,
so
I
believe
what
nun
does
like
they
can
afford
they.
Never
they
allow
you
to
build
a
pro
object,
they're,
probably
doing
something.
On
top
of
it
like
additional
infrastructure.
A
We
just
know
that
it's
yeah
you
get
right,
so
you
know
which
property
you're
being
asked
for
strictly
because
you've
gotten
the
the
method.
You
know
the
method
that
you
bounded
to.
C
C
C
C
A
So
what
you're
saying
like
on
this
on
the
the
node
add-on
api
side,
we
might
be
able
to
do
something
where
we.
C
So
we
pretty
much
if
we,
if
we
register
appropriate
descriptor
yeah,
we
can
probably
some
like
imagine
if
we
don't
use
this
property
descriptor
exactly
the
same
as
on
c
side,
but
we
maybe
have
a
little
bit
more
extended
one
which
we
persist
along
with
object
itself,
and
whenever
this
method
would
have
a
given
register,
we
somehow
can
write
those
property
names.
So
this
case
we
can
potentially
we
will
have
probably
different
signature
for
the
property
getter
which
to
which
we
can
give
you
stuff.
But
thank
you
answer
short
answer
to
this
question.
C
No,
we
can't
do
it
today,
but
we
we
may
think
about
like
if
we
want
to
edit
right.
C
B
C
C
A
C
Okay,
so
they
have
a
method
which
called
if
you
look
in
nand.h
called
set.
Accessor
set
accessory.
A
E
Foley,
I
think
that
there
is
god
kevin
that
are
yeah
are
waiting.
C
Okay,
okay,
so
right
this
method
they
used
to
to
register
this
getter
option,
etc.
Yeah
and
I
guess
they're
doing
some
kind
of
some
template
magic
here.
A
C
I
don't
know
if
we
want
to
do
it
like
it's.
I
kind
of
a
little
bit
it's
what
I
kind
of
like
about
current
node
apis
that
we
have
pretty
much
as
close
as
possible
to
javascript
object.
Javascript
model
does
not
provide
this
information.
C
If
people
want
to
do
it,
they
can
potentially
do
it
on
their
own,
but
I
don't
know
a
little
schedule,
it's
it's,
but
I
will
defer
to
you
guys.
E
Is
this
possible
with
other
just
engine
other
than
v8.
A
E
C
Yeah,
there's
no
easy
way
to
do
it
ourselves,
like
yeah
right,
I
mean
like
somebody
who's
using
our
apis.
It's
not
easy
for
them
to
do
it
actually
right,
because
there
is
no
way
to
do
it
unless,
unless
the
kind
of
the
only
way
I
can
think
of
is
actually
then
the
level
of
people
who
have
this
they're
providing
this
function
pointer,
they
can
use
some
template
magic
make
right.
So,
instead
of
like
having
specific
functions,
they
may
actually
have
this.
C
Template
and
kind
of
this
study
function
is
a
template,
and
this
function
effectively
have
a
template.
Parameter
would
be
whatever
is
given
during
registration
and
then
something
possible,
along
with
spain,.
C
So,
let's
think
about
it,
like
I
guess
my
my
question
like
do
we
want
to
do
it
or
not
if
we,
if
we
have
some
ideas
that
we
would
like
to
have
with
this
facility,
let's
let's
think
about
it,
but
I
guess
first
question:
do
we
want
to
do
that.
A
A
From
doing
it
yeah,
so
I
guess
like
here
like
we
might,
I
think,
we'd
be
oh
we'd,
be
interested
in
somebody
providing
a
solution,
but
it
might
be
that
they
would
go
into
documentation
versus
anything
else
like
if
there
was
some
way
to
do
this
through
template
magic.
If
we
had
a
here's,
a
few
lines
that
you
can
use
to
achieve
this,
I
think
that
would
be
really
interesting
and
I'd
want
to
put
it
in
a
dock
somewhere.
C
So,
thankfully
speaking,
maybe
we
just
need
to
provide
this
macro.
I
guess
like
in
the
end,
what
probably
goes
like
the
way,
I
would
think
it
could
potentially
be
done.
Like
so
measure.
You
have
a
macro.
You
provide
into
this
marker,
whatever
like
property
name,
you
want,
and
this
prop
your
name
can
be
magically
used
for
to
to
specif
to
kind
of
specialize
some
some
some
template
parameter
for
some.
A
C
A
A
Right,
okay!
Well,
I
think
I
mean,
because
I
mean
correct
me
if
I'm
wrong,
but
I
look
at
that
as
a.
It
is
kind
of
an
interesting
thing
where,
like
if
you're
gonna
have
a
whole
bunch
of
getters
and
if
you
can
basically
use
the
same
method,
given
that
you
know
the
name,
I
can
understand
how
that
might
be
a
common
use
case
right.
C
C
A
C
A
D
A
A
A
There's
anything
radically
different
here,
just
still
sort
of
that
slow.
A
F
A
Sounds
great
debug
testing
I
haven't
done
and
then
okay,
so
we
can
look
at
the
issues
I
do
have
one
which
it
would
be
great
if
I
don't
know
if
somebody
has
had
a
chance
to
take
a
look
at
it,
but
it
would
be
great
if
somebody
could
take
a
look
at
this
pr,
because
I
think
we
definitely
want
to
land
it
before
we
do
the
next
release.
G
Yes,
I
have
been
trying
to
work
on
that
one,
the
formatting,
all
the
code
which
has
is
which
had
the
issue
with
the
types
say,
type
typed
thread,
save
function,
yeah,
yeah,
and
so
what
I
did
was
I
installed
an
ubuntu
vm
and
tried
to
run
the
tests
and
I
am
not
able
to
replicate
it
after
a
couple
hundred
attempts
of
running
the
tests
right.
G
One
thing
that
I
noticed
is
that
I'm
not
entirely
sure
if
it's
using
the
node
version
that
it
says
that
it's
running
on,
if
you
can
go
back
once
you
finish.
G
G
G
Right
and
I
looked
at
the
the
github
workflows
and
it
does
correctly
install
the
right
node
version,
but
I
don't
know
why
it's
using
node
16
here
and
then,
if
you
look
at
my
last
comment,
this
there's
a
deprecation
warning
that
I
was
talking
about.
If
you
scroll
down
a
little
bit
this
setting
process,
config
is
deprecated
that
only
shows
in
node
16..
G
G
So
here
we
can
see
that
you
know
it's
linking
with
node
16,
which
I
suppose
is
okay,
because
it's
just
the
headers
or
whatever.
Yes,.
G
Right
but
then,
when
it
actually
runs
the
test,
so
there's
this
set
up
dependencies
task
job.
I
guess
it's
their
step
well,
this
is.
G
G
Yes,
so
on
line
23,
I
think
it
had
or
yeah
so
there
it
installs
the
dependencies
and
then
30
for
use
node.js.
It
uses
nvs
to
add
and
use
the
correct
version,
and
we
see
that,
like
it
console
logs
in
the
build,
also
the
correct
version
of
node,
but
I
don't
know
why
yeah
and
that
and
use
node.js14.
G
B
A
G
A
G
Right
because,
in
our
test
javascript,
we
probably
we
output
the
the
the
napi
process
version,
but
not
the
node
version,
but
yeah.
We
see
there
it's
12.,
okay,
so
I
have
not
been
able
to
replicate
and
I
don't
really
know
like
I
looked
at
the
test
and
everything
looks
fine,
so
I'm
not
sure
exactly
what's
going
on.
G
D
C
Can
be
even
like
you
can
install
them
in
your
machine,
but
that's
because
you
have
a
little
bit
different
hardware.
Exactly
I
didn't
see
it.
I
had
to
deal
with
some
such
issues
like.
G
In
the
github
ci
right
yeah,
so
this
sort
of
you
know
inconsistent
failures
and
whatnot
as
I'm
unable
to
replicate
it.
I
guess
I
can
try
getting
a
debian
vm
to
see
if
it
behaves
like
this
ci
does
then.
But,
like
I
mentioned,
I
looked
at
the
test
and
I
can't
see
anything
wrong.
G
The
only
thing
I
can
think
of
next
would
be
just
like
maybe
making
a
separate
branch
that
adds
a
bunch
of
like
std,
see
outs
in
the
in
the
native
side,
just
for
like
additional
debugging,
or
something
like
that
just
to
like
print
some
stuff
to
see
to
try
to
see
where
it's
failing
or
where
it's
hanging
yeah
and
then
just
keep
on
pushing
stuff
to
the
ci
to
see
where
it
fails.
C
C
Intermediate,
I
guess
my
question:
do
I
don't
know
linux
that
much
and
on
windows
we
typically
have
something
called
watson
dump
like
so
thankfully
the
name
we
have
some
kind
of
crash,
we
can
collect
some
dumb
file
and
then
we
can
actually
have
a
snapshot
of
memory
and
stuff
load
inside
of
the
debugger
and
see
what
exactly
happened
like
you
can
at
least
look
at
this
stuff.
I
wonder
if
linux
has
something
like
this
yeah.
A
G
Yeah
we
would
need
to
kill
it
and
because
it's
not
actually
crashing
what's
happening
is
that
it's
just
not,
for
example,
resolving
some
promise
somewhere
or
either
on
the
javascript
side
or
resolving
a
future
in
the
native
side.
So
since
none
of
that's
happening,
the
process
is
just
stopped
waiting
for
something
to
occur.
That
never
does
so.
There
is
no
actual
core
dump
that
occurs.
C
Issues
like
we
waiting
for
result
and
waiting
forever
yeah,
it's
not
okay,
it's
not
a
crash.
It's
something
but
not
being
completed.
G
G
So
I
guess
now
I'll
try
getting
a
debian
vm
to
see
because
it
seems
like
debian
is
usually
failing
the
most.
If
I
can
reproduce
it,
then
I
could
do
these
like
adding
the
console
logs
and
stdc
outs
without
having
to
continuously
push
and
wait.
You
know
flood
up
the
ci's
and
whatnot.
D
A
C
This
one,
where
it
failed
kevin
you
mentioned
you're
using
futures
and
c
plus
plus
I
believe
that
is.
That
is
the
case.
Yes,
so
I
wonder
like
if
different
versions
of
linux
may
potentially
using
different
implementation
of
a
standard
library
and
maybe
there's
some
kind
of
issues
with
implementation
standard
library.
This
future,
because
I
know
that
the
people
generally
hate
standards
future
and
probably
not
so
much
love
and
and
maybe.
G
Some
issues
there
right.
I
guess
this
specific
test
uses
mutexes
and
condition
variables.
There's
not
this
one
doesn't
have
futures
then,
but
you're
right
like
maybe,
there
are
different
implementations
for
standard
library
for
those
those
things
and
we.
G
E
For
example,
in
I
don't
know
we,
we
changed
the
minimum
version
of
gcc
in
a
node.
G
A
A
C
Whenever
you
want
to
wake
up
this
variable,
like
some
cases,
people
saying
you
need
to
do
it
under
lock
and
some
other
cases,
people
saying
without
lock,
I've
seen
like
some
slight
timing,
bugs,
depending
how
it's
actually
done,
and
maybe,
depending
on
internal
implementation,
can
be
some
slight
timing
issues
just
trying
to
kind
of
brainstorm
some
ideas.
C
What
can
go
wrong
with
using
actual
variables
because
we're
waiting
we
pretty
much
we
under
log,
but
while
we're
waiting,
we're
kind
of
releasing
these
clocks
and
we
need
to
kind
of
send
the
signal
to
wake
it
up
right
like,
and
there
are
two
school
thoughts
like.
Should
it
happen
under
local?
Without
look,
let's
see,
okay,
so
it's
maybe
and
maybe,
depending
on
different
versions
of
library
and
stuff.
People
have
a
little
bit
different
implementation,
different
requirements,
a
little
bit
which
may
affect
some
timing
just
slightly
and
and
can
be
different
of.
G
Right,
okay,
the
thing
is
so
like
the
because
we
have
a
the
the
original,
the
non-typed
thread
say
function
sum
class:
it's
basically
it's
pretty
much
the
same
test,
so
it
uses
both
the
the
mutexes
and
the
conditional
variables
in
the
same
way
that
they're
used
in
the
same
way
in
both
tests.
G
G
C
One
case
you
have,
even
maybe
the
whole
structure
is
same,
but
maybe
like
just
few
extra
instructions
being
executed
and
now
like.
Maybe
the
different
versions
of
linux
have
a
little
bit
different
kind
of
scheduler
for
different
operations
and
we're
just
hitting
this
unfortunate
situation.
More
often
right.
G
Right,
what
was
the
pr
that
did
the
enhanced
test
framework.
D
G
That
was
the
the
unit
test
filter
right,
okay,
yeah.
A
Yeah,
I
I
I'm
just
more
like
you
know,
it's
timing,
these
things
to
be
timing,
but
maybe
maybe
what
makes
sense
if
he
can
try
debian,
because
that
seems
to
be
where
it
happens.
If
it
doesn't
happen,
then
what
I
think
the
next
step
is,
we
can
ask
to
get
you
access
if
you,
if
you're
willing
to
take
a
look
on
the
machines
themselves,.
A
A
I
mean
I,
I
don't
think
it
has
to
do
with
that
particular
pr,
but
it
is,
it
is
happening
enough
that
we
probably
really
need
to
fix
it
right.
G
Some
way
yeah
and
then,
regarding
this
oddity
of
not
using
the
right
node
version,
I
was
just
going
to
simply
modify
the
test
to
also
output
the
node
version,
because
we
do
yeah.
I.
G
A
A
C
A
C
A
D
C
We
have
this
function.
First,
first
function
factory
measure
the
size
of
string.
You
need
second,
like
people,
then
can
use
two
string
to
say
reserve
how
many
bits
you
need
and
then
do
this
stuff.
I
can
provide
answers
like
it
should
be.
Okay,.
B
D
A
C
C
A
Yeah,
sorry,
I
I
don't
know
why,
but
it
was
in
a
different
window.
So
this
one.
C
C
So
if
you
look
on
the
types
first
like
so
basically
for
them,
I
propose
a
couple
of
things
like
first
of
all,
I'm
introducing
this
new.
You
know
we're
pretty
much
saying
and
again,
I'm
not
good
with
names.
Better
names,
suggestions.
Welcome,
so
exactly
we
have
two
types,
maybe
not
even
even
we
should.
Maybe
we
should
have
flags.
I
don't
know
it's.
It's
also
kind
of
I'm
open
to
ideas,
but
then
I
looked
at
that
we
had
was
all
this
stuff.
C
Every
time
we
were
sending
all
this
native
data,
finalizers
and
stuff,
like
so
many
different
fields,
it's
like,
let's
do
maybe
one
struct,
which
encapsulates
all
them
in
one
single
place
and
what
I
found
even
inside
of
our
code.
We
just
simply
can
pass
a
pointer
to
the
struct
rather
than
coping
with
values,
again
again
every
single
time,
so
we
can
hopefully
even
do
some
internal
optimizations
by
using
this
strap.
A
C
No,
but
because
this
tracks
will
be
typically
being
on
the
call
stack,
so
it's
so,
if
you
think
from
the
perspective
of
then
we
call
a
c
function,
we
we
need
to
put
all
these
parameters
on
on
a
stack
anyway.
So
here,
instead
we
said:
okay,
there
is
a
structure
on
the
stack,
we're
just
sending
pointer
to
this
track,
so
think,
we'll
probably
extra
pointer
and
again
just
for
abs
safety.
C
Right
and
there
are
a
bunch
of
methods
which
kind
of
whatever
we
used
to
have
finalizers
before
so
now,
all
these
methods
are
receiving
this
new
struct
and,
like
last
time,
we
we
this
we
discussed
all
this
new
method
has
pretty
much
the
same
name
but
different
prefix.
Instead
of
an
apis
that
have
a
node
api
right,
okay
and
internally,
depending
on
this
flag,
we
can
actually
call
either
synchronously
or
asynchronously,
and
thankfully
this
call
finalizer
became
non-virtual
method.
Now,
it's
all
inside
of
a
js
native
api.
C
Now
I
had
to
do
some
magic
about
set
immediate
because
I
think
respect
have
to
do
set
immediate.
So
I
put
this
if
you
look
inside
of
just
native
api
internal,
so
I
have
some
kind
of
prototype
there.
So
if
somebody
wants
to
implement
just
native
api
v8
outside
of
node,
they
just
need
to
provide
this
set
immediate
template
implementation,
but
inside
of
our
code
we
pretty
much
just
doing.
We
just
call
it
set
immediate
from
from
this
environment,
know
the
environment.
C
C
A
C
E
C
Synchronous
call
you
see,
I'm
introducing
something,
what
they
call
finalizer
context
and
this
guy,
if
you
scroll
up
a
little
bit,
does
a
few
things
so
most
important
that
it
prohibits
using.
So
oh,
it
actually
preserves
current
air
state
because
imagine
like
we,
we
executed
some
javascript.
Maybe
there
is
some
kind
of
pending
javascript
exception,
maybe
something
else.
So
thank
you.
C
What
we
do
we're
preserving
the
current
air
state
somewhere
here
locally
and
it's
just
cleaning
top,
then
we
okay
handle
scope
and
stuff,
and
then
we're
saying
we
would
like
to
prohibit
any
javascript
call.
Apparently
all
our
methods
inside
of
our
preamble,
we
check
if
we
can
call
a
javascript
or
not
okay.
So
in
query.
C
Okay,
so
what
what
happened
like
before
in
this
base
class
in
just
native
api,
we
always
return
true.
So
now
I
have
introduced
additional
field
prohibit
called
js,
and
now
we
instead
of
turn
through,
we
call
this
prohibit
called
js
so
and
then
in
derived
class.
If
you
look
at
node
api
cc,.
C
Okay,
sorry,
I
introduced
this
alias
super,
and
so
previously
it
was
just
using
node
environment
can
call
into
js
right
now,
like
I'm,
I'm
saying
like
in
our
base
class.
We
can
also
can
prohibit
for
this
while
we're
running
for
analyzers.
Essentially,
so
it's
a
kind.
C
C
H,
oh
sorry,
it's
internal
internals,
yes,
node,
api
internals
so
because
node
api
is
probably
class
internal,
so
it's
actually
class
declaration.
Yes,
it's
what
I
usually
have
whenever
writing
classes,
it's
more
convenient
to
always
refer
to
super
instead
of
kind
of
every
time.
Remember
what
exactly
public
class
kind
of
part
of
the
class
name.
B
A
Okay,
but
yeah,
okay,
so,
but
I
think
it'll
take
some
walk
to
go
and
take
any
any
other
things
you
want
to
kind
of
highlight:
we'll
have
to
go
and
take
a
more
closer
look.
Obviously,
but.
C
My
biggest
question
is
about
all
these
external
buffers
because
our
then
locating
external
buffer.
You
know
an
api
and
native
api
cc.
We
can,
according
to
this
node
api
cc,
and
I
can't
try
to
do
stuff,
but
I
think,
let's
it
has
kind
of
very
unique
pattern
for
destruction.
C
So
what
parts
or
which
parts
are
you
talking
about?
Oh
if
this
node
api
cc,
if
you
scroll
down
a
little
bit,
I
think
it's
there's
some
something
about.
Oh
no,
it's
yeah!
C
C
But
at
this
point
I
think
it's
a
draft.
I
really
want
just
us
to
discuss
it.
If
it's
approach,
we
want
to
proceed
because
currently
I
don't
have
additional
tests,
I
don't
have
documentation,
change
and
stuff.
I
know
that
I
still
have
big
to-do
lists.
It's
more
like.
Could
you
preview
it
and
see
like
doesn't
make
sense
to
move
forward
with
all
these
right
work.
A
D
Okay,
any
others
that
we
should
hi
that
people
want
to
highlight.
A
C
A
Think
yeah,
I
asked
you
a
question
which
was
like
there
was
like
a
non-blocking
comment
and
he
said
that's
a
good
point,
but
so
I
wasn't
sure
whether
you
know
I
should
go
ahead
and
land
it
or
you
want
to
maybe
address
that
point
or
I'm
sort
of
good
either
way.
But
I
wanted
to
know
what
you
were
thinking
like
land
this
and
do
and
do
a
follow-on
pr
or
or
not,
or
you
want
to
do
something
before
this
lens.
C
I'm
not
sure
how
to
implement
it
without
innumerable,
because
it
actually
imagine
it
gives
us
all
different
methods
like
this
set
of
methods
it's
different
and
because
I
originally
did
it
and
it
didn't
work
so
I
had
to
edit
was
simplest
way
to
work
around
it.
So
I
can
probably
do
it
for
current
hopes.
Only
okay.
A
C
Let
me
add
this
this
test,
like
I,
I
then
I
responded
to
it.
I
can
do
this
like
instead
of
going
to
research
for
the
chain,
I
can,
I
can
add,
a
couple
new
tests:
okay,
two
tests:
okay,.
A
A
C
C
It
was
more,
I
remember
like
with
permanent
permit,
but
persistent
rf
count.
It's
also
yeah.
There
is
pr
for
it
as
well.
C
So
I
have
new
iteration
for
this
stuff,
based
on
our
last
discussion.
So
previously
I
remember,
I
have
created
four
different
methods
since
now,
like
I
am
changing
one
of
the
methods
which
now
can
accept
reference
type
and
also
add
another
method
to
to
return
this
type.