►
From YouTube: Node.js Tooling Group Meeting 2021-07-09
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).
B
B
B
Okay,
well,
let's
start
going
into
the
agenda,
then
first
one
is
the
contentious
116,
which
is
the
whether
or
not
we
should
have
some
sort
of
linter
or
other
mechanism
for
notifying
folks
about
new
features
that
we've
added
to
core
like
rimra
for
mcdurk.
B
I
haven't
seen
the
less
contentious
issue.
Well,
yeah.
C
B
That's
jordan,
that's
the
story
of
of
doing
anything,
trying
to
do
anything.
C
No,
he
the
one
that
said:
don't
do
this.
Okay,
anyways.
C
Yeah,
I
haven't
done
anything
else
on
this,
though
I'm
still
still
not
sure
what
you
know.
The
path
forward
is.
B
I
think
go
ahead.
Oh
sorry,
go
ahead
joe,
I
was
gonna
say
a
blog
post
I
think,
is
always
a
good
call.
Just
kind
of
standardizing
on
doing.
B
A
Yeah,
it's
interesting.
I
can't
remember
when
we
talked
about
it
a
few
maybe
a
few
months
ago
and
kai
was
there
too
from
the
slint
team,
but
it
seemed
like
we
had
decided
that
eslint
was
maybe
not
the
best
place
to
do
it,
but
I'm
just
not
sure
how
to
approach
this,
and-
and
you
know
we
don't
need
to
solve
it
today,
but
something
to
think
about.
C
One
of
the
issues
with
the
s
land
for
for
better
or
worse,
is
that
you
would
have
to
install
it
in
your
project
and
install
whatever
like
preset
that
contains
these
rules
right.
Yes,
maybe
that's
not
as
discoverable
because
of
that,
but
on
the
other
hand,
it
lets
people
decide
if
they
want
to
opt
into
it.
A
I'll
just
say
real
quickly,
I
mean
what,
if
we
created
like
some
sort
of
npm
package,
that
you
run
to
check
your
code
to
see.
If
your,
like
you
know,
first
person
code,
not
your
third-party
code
is,
is
calling
things
that
we
know
we've
implemented
or
something
that's.
C
A
Then
it's
like
sort
of
you
know
we're
providing
a
service.
You
opt
into
it.
If
you
want,
if
you
want
to
run
it,
you
can
run
it.
C
It's
like
there
are,
there
are
tools
like
one
that
I
use
is
called
the
depth
check.
I
think
that
will
note
like
hey
here's,
some
dependencies
in
your
package,
json
that
you're,
not
that
you
don't
appear
to
actually
be
using,
so
I
think,
maybe
be
kind
of
similar
to
that
yeah
exactly
you
were
gonna.
That
sounds
like
a
great
great
idea.
I
like
that
idea.
I.
B
I
was
gonna
say
well
I
like
that
idea,
a
lot
and
that
seems
to
really
low-hanging
fruit
and
non-controversial
which
is
really
nice.
Another
recommendation
would
be
if
the
like,
if
library
authors
could
somehow
opt
into
it
like
if
we
were
to
land
argument,
parsing
and
node,
I
would
happily
put
a
warning
in
yard's
parser.
That
says
you
know,
are
you
sure
you
want
york's
parser?
You
could
totally
just
use
the
one
built
into
node.
A
B
C
C
Yeah,
I
know
this
was
a
while
ago.
I
don't
remember
his
exact
reasoning.
B
But
it
wasn't
so
much
about
that.
It
was
it's
more
about,
I
think
him
being
like
overly
involved
in
the
process
of
merging
it
into
node
core,
and
this
was
a
couple
of
years
ago
and
I
think
isaac
has
actually
taken
over
mcdermott
now
from
james.
Maybe
I
I
don't
know
having
a
little
module
seems
low-hanging
fruit,
not
controversial.
B
A
B
C
A
Yeah,
we
can
think
of
it
as
a
two-pronged
approach.
You
know
maybe
work
on
this
package,
but
also
encourage
package.
Authors
were
applicable
to
to
add
some
language
as
well
and
then
maybe
some
guidance
for
node
core
developers.
You
know
when
they
do
add
something
to
just
be
mindful
of
this
process
that
you
know
we
can
help
support
them
to
use
core
core
functionality.
C
Yeah
in
this
in
this
pr
here
say
like
we
discussed
this
in
the
tooling
meeting
today
and
we've
come
up
with
this
idea
of
creating
like
a
standalone
like
little
cli.
That
could
point
out
these
kinds
of
things.
Yup.
B
C
A
It's
it
does
something
yeah,
it's
been,
it's
been
a
couple
of
months
since
I
worked
on
it,
but
yeah
it
was
doing
something.
B
C
You
want
to
work
on
parsers
as
well,
so
yeah.
This
would
probably
be
a
good
chance
for
me
to
to
do
some
work
in
both
of
those
things
so
yeah,
I'm
on
it.
B
B
Then
we
can
write
a
blog
post
about.
I
still
think
we
should
write
blog
posts
about
new
features.
We
add,
but
we
can
write
a
blog
post
about
this
cli
when
it's
ready.
A
C
To
me,
I
can
put
together
like
a
little
prototype
or
something
in
my
own
account
and
then,
if
we
like
it,
if
we
and
we
like
this
idea,
we
can
maybe
move
it
over
to
the
like
pkg
js
org.
Is
that
we're
using
right
now
for
ourselves
yeah?
So
it's
probably
a
good
place
for
it,
and
if
we
do
decide,
we
want
to
go
ahead
with
it.
C
Looks
like
you
are
in
yeah
in
this
this
issue
yeah.
I
was
just
writing
something
right
now.
B
Recursive
cop
recursive
copy,
I
I
sent
an
email
to
jp
richardson
who's.
The
person
who
originally
wrote
fs
extra
along
with
ryan
zim,
who
is
the
person
who's,
been
doing
the
most
maintenance
recently
on
fs
extra
and
they
both
were
supportive
of
us
pulling
in
the
copy
functionality
from
fs
extra
as
a
foundation
for
recursive
copying,
behavior
node.
B
I
was
looking
through
that
code
base
and
it's
really
it's
really
nice
and
self-contained.
So
I
think
I
I
was
starting
to
take
a
stab
at
pulling
it
into
node,
but
didn't
get
super
far,
but
I
was
going
to
take
another
stab
this
weekend,
maybe
cool.
I
was
writing
an
issue
where
it
wraps
a
bunch
of
helpers
like
it
has
it
has
kind
of
a
util
library
that
then
the
copy
stuff
relies
on.
So
I
was,
I
think,
trying
to.
B
I
need
to
figure
out
how
much
of
that
util
library
I
need
to
also
pluck
across
to
it.
It
seems
possible
unless
the
util
library
then
has
its
own
dependencies,
it's
pulling
in
so
as
long
as
it
doesn't
turn
into
an
infinite
recursion,
where
I
pull
in
all
of
fs
extra
to
make
it
work,
which
I'm
hoping
it's
not
looking
like
that'll,
be
the
case
today,
like
there's
a
bunch
of
helpers
in
a
utility
file
and
then
there's
the
copy
method
itself,
which
is
pretty
self-contained.
B
So
I
think
I
can
kind
of
do
surgery
and
pull
out
just
that
functionality,
cool,
but
I'll
need
to
spend
a
few
hours
actually
trying.
B
B
B
A
B
B
So
yeah
I'm
going
to
try
without
the
graceful
fs
for
the
time
being.
The
other
question
that
came
up
is
so
node.js
today.
If
we
look
at
the
node.js
file
api.
B
I
don't
think,
there's
a
move
file
or
a
copy
folder
directory.
So
so
I
was
thinking
of
adding
a
copy,
there's
a
move
file
or
sorry,
there's
a
copy
file
method.
Today
I
was
thinking
of
adding
this
as
copy,
which
would
be
like
analogous
to
copy
dash
r
kind
of
behavior
and
then
the
other
question
I
had
was
whether
we
want
something
like
move.
C
B
C
Yeah,
I
mean
one
I
guess
argument
against
that
is
like
you
could
implement
move
like
much
more
efficiently
right
by
just
actually
like
moving
the
thing
in
place
like
like
by
doing
it
that
way,
you're
like
duplicating
everything
and
then
deleting
one
copy
of
it
right,
whereas,
like
I
think,
a
proper
move
implementation,
the
benefit
there
is
that
it
would
be
like
a
lot
more
efficient.
C
Because
I
know
like
I
mean
I'm
certainly
no
expert
in
this,
but
I
know
in,
like
you
know,
posix
or
like
in
unix
like
shells
right,
if
you,
if
you
move
like
a
giant
directory,
it
just
happens
instantly
right
because
it's
not
actually.
B
Yeah
it
looks
like
fs
extra
smarter
than
just
doing
the
remove.
It
only
doesn't
remove
if
you're
moving
across
device,
and
otherwise
it
does
right.
A
rename
operation.
C
Is
what
does
do
rename
you?
Oh
I'm
just
looking
in
the
so
there
is
a
rename
command
in
node
in
the
fs
api.
B
B
C
B
I'm
tempted
to
start
with
copy,
though,
because
if
the
use
case
I'm
thinking
of
that
this
will
be
super
useful
for
us.
You
have
a
fixtures
folder
for
some
library
that
needs
to
interact
with
a
directory
structure
of
stuff
that
fixtures
folder
is
always
there
in
your
test
directory.
But
as
you
set
up
your
test,
you
do
a
copy
dash
r
of
that
folder
run
your
test
against
it,
and
then
rim
raft
that
folder
afterwards,
because
it's
that's
your
tear
down
behavior.
B
So
I
think
that
will
solve
a
really
useful
use
case.
First,
moves
more,
I
mean.
Actually
my
move
is
useful
too,
like
at
npm
we
would
have.
May
I
don't
know,
maybe
drop
files
down
in
a
temporary
location,
confirm
they're,
correct
and
then
move
them
to
their
final
resting
place
afterwards
or
something
that'd
be
really
nice
to
be
able
to
do
as
well.
B
Of
course,
if
you're
setting
that
up
in
production,
you're
probably
doing
it
on
the
same
device,
maybe
and
copying
it,
and
then
you
can
do
the
rename
right
so
anyways
yeah.
So
so
I
think
that
that
to
me
that
works
coming
along
pretty.
Well,
though,
I
have
too
much
more
to
add
there,
I
I
think
I
could,
depending
on,
if
I
feel
like
coding
this
weekend
or
if
I'm
not
feeling
like
coding
this
weekend,
I
might
have
a
pr
in
the
next
week
or.
B
Hopefully
I
haven't
checked
how
many
open
issues
I
have,
and
I
don't
do
too
much
work
on
nyc
like
I'm
wondering
if
this
is
a
general
enough
problem
to
keep
on
our
plate
like
it's
not
something
like.
I
don't
find
myself
clamoring
for
it
in
my
day-to-day
work
at
google,
where
I'm
writing
lots
of
open
source
libraries,
I'm
not
wishing
I
better
detect
the
exiting
behavior,
I'm
just
not
just
not
finding
it's
a
problem.
I
have
too
often
yeah.
C
I
was
gonna
say
we
could,
maybe
you
could
leave
it
open
and
just
remove
it
from
the
agenda
as
well.
Yeah.
B
B
I
can
imagine
the
node.js
unit
tests
themselves
using
it
quite
a
bit
right
or
is
this
when
I'm
like
yeah,
you
can
bump
into
really
needing
this
function,
but
the
few
times
you
do
that
you
just
put
in
the
trouble
to
actually
capture
all
the
signals,
and
then
you
know
it's
the
sort
of
thing
you
only
have
to
do
once
and
you
only
have
to
do
it
if
you're
a
utility.
So
maybe
it's
I've
convinced
myself,
it's
off
the
agenda
for
now
all
right,
so
we
haven't
really
been
moving
that
part
at
all.
C
Something
kind
of
related
that
I
do
actually
want,
and
that
is
kind
of
like
a
pain
point
for
me,
is
like
an
async
exit
hook.
Type
thing:
I
don't
know.
If
I
don't
know,
if
either
of
you
saw
there
was
a
thread
on
twitter,
I
think
it
was
last
week.
C
I
think
my
even
tyranny
that
started
it.
He
was
asking
like
what
are
the
you
know
if
you
could
have
anything
you
wanted
in
like
node
or
npm
or
whatever?
What
would
it
be?
It's
kind
of
an
interesting
thread.
There
were
some
some
good
suggestions
in
there,
but
that
was
one
that
came
up
and
that's
something.
C
You
know
that
I
oh
yeah
there
we
go
thanks.
Joe
you're
welcome
yeah.
There
are
some
good
good
suggestions
in
here,
and
I
mean
the
other
thing
like
some
of
the
suggestions
I
was
happy
to
see
are
things
that
we're
actually
working
on,
which
is
kind
of
nice,
but
yeah.
That
was
one
kind
of
I
just
bring
it
up
because
you
know
we're
talking
about
like
exit
behavior.
C
B
B
Well,
one
of
the
other
reasons
people
have
pushed
back
on
it,
I
think,
is
you
can
get
in
these
situations.
People
who
know
better
about
the
this
part
of
node
than
myself
are
we're
concerned
about
getting
in
positions
where
the
application
basically
never
exits,
or
it
doesn't
actually
exit
gracefully
because
you're
creating
some
sort
of
you're
not
shutting
down
the
event
loop,
you're
waiting
to
shut
down
the
event
loop,
and
then
you
do
things
that
can
start
to
put
more
work
on
the
event
loop.
D
I
got
summoned
by
listening,
so
the
event
loop
would
be
stopped
running,
but
you
can
still
run
micro
task
queue
stuff.
So,
basically,
no
I
o
jobs
can
be
responded
to.
You
could
queue
up
things
to
flush
to
disk,
that's
fine
and
behind
the
scenes,
as
long
as
after
the
micro
task
queue
is
done.
No
more,
I
o
is
queued
up.
You
should
be
okay,
other
than
that
node
just
isn't
set
up
to
kind
of
have
the
process
exiting
while
the
event
loop
is
running.
D
If
that
makes
sense,
so
you
could
do
like
a
big
chunked
async
right
to
disk,
but
you
can't
really
read
from
disk.
Basically
is
the
problem
without
us
going
and
changing
how
exit
behavior
works
across
core.
B
D
It's
not
just
chunks
of
node
core
itself.
It's
also
people
expect
things
to
kind
of
die,
and
so,
if
you
start
doing
things
asynchronously
like
should
a
server
be
kept
alive
during
the
state,
even
though
the
process
may
be
exiting
because
it's
in
a
bad
state
like
you,
you
want
to
manually,
shut
down
all
the
connections
and
stuff
like
that,
because
node
itself,
as
a
runtime,
doesn't
know
what's
safe
to
keep
responding
to.
I
o
wise.
B
C
C
D
Kind
of
you
can't
use
the
micro
task
queue,
but
we
probably
could
make
it
so.
The
micro
task
queue
promise
awaits,
could
finish
no
timers.
C
D
D
You
might
be
able
to
work
around
it,
but
just
the
way
things
are
architected
right
now,
there's
no
isolation
on
what's
kind
of
the
dispose
going
on
at
the
end
of
the
process
versus
what
you
know
was
running
during
the
main
life
cycle
of
the
process,
and
so
that's
the
bigger
issue.
I
think
so
all
the
I
o
that
you're
doing
for
your
application.
C
All
right:
well,
thanks
for
the
insight
on
that
one
anyways
yeah.
I
wasn't
necessarily
advocating
we
take
this
on
just
brought
it
up,
because
it
is
kind
of
related
and
also
wanted
to
mention
that
that
twitter
thread
there's
some
there's
some
good
stuff
in
there.
If
you
haven't
seen
it
yet,
I
would
recommend
looking
through
it.
We
got
another
request
for
like
a
compile
type
behavior
like
we've
talked
about
in
the
past.
C
Anyways
yeah
we
can.
We
can
move
on
to
the
next
thing,
though,.
B
B
Stands
to
reason:
are
we
okay,
with
moving
on
to
the
next
item.
C
D
So
one
side
thing
with
source
maps:
it's
not
on
the
agenda
yet,
but
with
work
imports
coming
along,
we
need
to
figure
out
the
story
there.
We
can
do
that
in
a
different
meeting.
B
From
an
http
resource,
not
just
oh
sorry,
brad,
I'm
not
trying
to
talk
over
you.
D
Http
https
or
potentially
james
snell
had
blob
urls
somewhere
in
his
pipeline.
Those
also
could
get
confusing
because
anything,
that's
not
off
disk
or
directly
represented
on
on
the
source
text
itself.
B
B
That
makes
sense
so
so
I
feel
like
we
should
be
able
to
use
most
of
the
same
machinery
where
we're
populating
the
source
map
in
the
same
cache
that
gets
populated
when
we
happen
to
find
a
source
map
on
disk.
We
just
would
need
to
have
that
cache,
be
populated
in
this
part
of
the
system
that
now
loads
these
resources
from
something
other
than
fs.
B
The
implementation
is
synchronous
right
now,
but
I
mean
it's
ultimately
just
putting
stuff
in
a
map,
so
there's
there's
no
reason
that
something
running
asynchronous
shouldn't
be
able
to
just
use
the
same
methods
to
put
something
into
that
map.
I
don't
think
it
shouldn't
be
an
issue
cool,
so
that's
good
to
know
future
work.
Maybe
that
means
I
just.
We
just
need
to
be
keeping
an
eye
out
for
when
that
work
starts
landing.
B
Okay,
doesn't
look
like
there's
any
new
outstanding
bugs
for
source
maps,
but
there
continues
to
be
some
improvements
where,
if
anyone
who
happens
to
be
listening
on
the
call
on
youtube,
this
could
be
a
good
place
to
start
contributing
to
node.js,
there's
specifically
source
maps.
Don't
work
well
in
the
rebel.
So
if
anyone
wants
to
play
around
with
the
rupple
code,
health
wanted
on
source
maps
in
rebel.
C
B
D
The
function
you
can't
use
import
meta
either,
there's
a
variety
of
things.
You
can't
do
a
weight
got
unblocked
recently,
but
it
hasn't
landed.
But
if
you
do
use
a
weight,
it
invalidates
other
parts
of
like
javascript
the
language
because
we're
doing
a
source
code
transform.
This
is
what.
C
Yeah
this,
that
was
the
most
recent
issue
that
I
ran
into
was
there's
no
top
level
await
and
I
think,
there's
a
flag
to
enable
it
but
yeah
it
breaks
like
like
you
can
reassign
const
all
of
a
sudden.
If
you
enable
that
flag.
B
Sounds
like
maybe
just
a
good
way
to
put
it
is
that
good
way
to
a
great
great
place
to
contribute
to
node
would
be
having
more
engineering
help
on
the
rebel,
because
it
sounds
like
there's
there's
areas
where
it
can
be
improved
outside
of
source
maps.
So
if
people
are
looking
for
a
way
to
get
involved
in
open
source,
this
could
be
a
good
experts
wanted.
B
Ronda
my
favorite
topic,
which
is
the
argument
parsing
topic.
A
Yeah,
no
news,
no
news,
I
I
should
have
more
time
next
week.
I
just
I've
been
kind
of
juggling
a
bunch
of
things
that
are
settled,
but
no
news
at
the
moment.
C
C
I
guess
on
that
note
too.
I
have
no
developments
on
this,
but
I
just
wanted
to
mention
brought
it
up
last
time.
That
could
be
an
interesting
thing
for
us
to
submit
like
a
talk
to
node
conf
remote
about
that
cfp
closes
at
the
end
of
july.
I
think
so
I
think
so
yeah
so
I
mean
I,
you
know
I'd
be
interested
in
doing
that
or
doing
that
with
someone
else.
I've
not
actually
ever
written
a
cfp
before
or
submitted
one.
So
if
anyone
has,
I
can
probably
use
some
help.
A
I've
done
tons
of
them,
so
I'd
be
happy
to
help
okay
happy
to
help
write
review
speak
whatever,
whatever
you
want.
C
B
I
would
defer
to
you
and
joe:
I
think
I'd
rather
sit
this
one
out
feeling
like
my
open
source
time.
I'd
rather
be
writing
open
source
these
days,
since
so
that's
only
temporary,
but
maybe
once
we're
in
well,
I
don't
will
in-person
conferences
be
a
thing
again.
I
don't
know
like
they
are
an
environmental
cost.
I
I
don't
know
what.
C
Well,
they
didn't.
They
say
that
open
js
world
next
year
is
going
to
be
in
austin,
yep
yep.
It's
already.
A
Already
you
know
planned.
I
know
there
are
some
other
lf
events
that
are
happening
already
this
year
in
october.
There's,
like
you,
know,
open
source,
something
or
other
yeah.
A
C
B
That's
a
non-sequitur
story,
but
I
think
it
would
be
great
for
you
to
put
in
a
talkie
in,
and
I
don't
know.
Obviously
it
sounds
like
joe's
up
for
it
either
way
so
yeah.
C
We'll
I'll
make
sure
that,
because
it's
you
know
only
like
three
weeks
away
or
something
until
the
the
cmp
I'll
make
sure
it's
in
my
calendar,
so
we
don't
miss
it
and
maybe.
C
C
Yeah
I'll
make
sure
yeah
for
sure,
even
if
we
just
get
this
submitted
and
and
someone
else
who's
worked
on,
it
more
wants
to
do
the
actual
talk
or
something
like
I'm,
probably
bad,
although
I
don't
know
if
they
let
you
swap
out
speakers
at
the
last
minute,
but
we're
all
friends
yeah.
Someone
can
just
pretend
they're
me.
B
Great
so
that
brings
us.
We
already
talked
a
little
bit
about
the
thread
that
tyranny
put
out
any
new
business
or
anything
we
didn't
cover.
B
Oh
well,
we
can
end
it
there.
Then
I'm
excited
to
start
doing
a
little
bit
of
work
on
copy
recursive.
I'm
gonna,
I'm
hoping
it's
not
too
hard
to
drop
in
and
at
least
get
a
conversation
started
about
it.
I
guess
I
can
just
call
it
copy,
because
we
don't
have
a
corollary.
We
only
have
copy
file
so
just
be
copy,
cool
and.
C
Right
yeah,
I
think
that
makes
sense
that
kind
of
fits
too
with
what
we
did
with
like
the
rm
command
right,
since
it
offers.
B
B
C
B
We
call
it
we
already
have
copy
file,
so
copy
feels
really
natural,
which
would
obviously
copy
a
file
or
copy
a
directory
structure
cool
if
we
don't
have
to
relitigate
that
oh.