►
From YouTube: Node.js Tooling Group 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
Okay,
I
think
we're
live
hey
everyone.
This
is
the
node.js
tooling
meeting
it's
friday
february.
18Th!
Oh
now,
we're
live.
I
don't
know
anyways
yeah,
let's,
let's
get
to
it.
The
first
first
on
the
agenda,
so
I
mentioned.
Actually
when
we
started,
I
was
waiting
for
people.
I
was
just
cleaning
some
stuff
up,
so
there
was
this
issue
about
test
failures
in
fs
render
recursive
and
I
know
ben
you
put
up
a
pr
for
that.
A
B
There's
yeah,
the
problem
still
seems
to
be
happening:
oh
okay,
it
doesn't
and
it
doesn't
make
that
much
sense.
Well,
I'm
not
quite
sure.
What's
doing
it,
I
had
said
I
have
an
idea
of
at
least
to
try
to
make
the
prom
easier
to
debug,
which
is
serializing.
B
I
think
it'd
be
interesting
to
serialize
the
tests
right
now,
they're
running
in
parallel,
and
I'm
one
I'm
wondering
if
there's
just
windows,
there's
a
bunch
of
logic
inside
of
rimder
for
windows
that
retries.
If
there's
contention
on
the
folder,
which
we
pulled
over
from
the
rimraff
library
and
this
failure
seems
to
happen
periodically,
I'm
guessing
it
has
to
do
with
more
than
one
thing
trying
to
delete
the
folder
at
once
and
it
hitting
its
maximum
retries
on
that
check
for
contention.
B
So
action
item:
I
should
actually
do
what
I
said
I
was
going
to
do
or
if
anyone
wants
to
jump
in
and
do
that,
but
I
think
it's
like
it's
different
than
how
you
usually
write
tests
in
node.
I
think
they
tend
to
be
parallelized
for
callback
based
tests
like
just
by
the
with
the
way
the
test
suite
works,
but
I
think
basically
we'd
want
to
kind
of
wait
for
one
to
finish
before
we
do.
The
next
might
be
an
interesting
thing
to
try
even
just
to
get
more
information.
A
B
B
A
Okay,
anything
else
about
that.
One.
C
I
think
that's
been
slowly
chugging
along
there's
a
draft
pr
up
for
it.
Now
it's
41.439
in
node
core,
I
think
ian
has
been
ian.
Ethan
has
been
slowly
tripping
away
at
making
it
something
that
he
feels
like
he
can
get
it
get
done
fully
and
then
merge
it
in
and
then
iterate
on
it
from
there
and
subsequent
pr's.
I
think
I
I'm
going
through
the
comments.
Now
it's
been
a
few
days.
C
Apparently,
since
I've
looked
and
there's
been
progress,
it's
been
like
a
month
since
I've
looked
and
there's
been
progress.
One
thing
I
think
I
do
agree
with
is,
I
think
it
should
probably
land
with
the
promise
api,
which
is
feedback
I
think
ben
gave.
C
I
I
I
think
we
should
be
landing.
We
should
not
be
landing
new
apis
without
promise
apis,
so
that
that's
the
only
one
only
one
I
have
only
additional
feedback
I
have
but
yeah
it
seems
to
be
going
pretty
well,
so
yeah
excited
excited
for
this
for
sure.
B
B
I
believe
that
the
benefit
of
opendir
is
that
it's,
it
doesn't
page
everything
into
memory.
So
if
so,
if
you
do
read
dur
on
a
like
a
huge
directory,
it
will.
It
will
just
pull
that
whole
thing
into
into
memory,
and
can
you
know
just
has
issues
yeah
so
so
the
recommendation
people
gave
was
just
use
opener
everything
else
exactly
the
same,
but
you
but
you'd
have,
but
it
would
be
a
like
an
async
iterator.
B
C
I
think
he
said
he
pushed
it
or
maybe
maybe
I
missed
that.
Maybe
we
were
just.
B
Going
to
update
the
title
or
something
yeah
yeah
yeah
anyways,
I
don't
think
there's
anyone
just
it
seems
like
it
has
buy-in
from
a
bunch
of
people,
ethan's
cool
with
that
slight
change.
So
I
think
it's
just
a
matter
of
when
he
has
time
to
work
on
it.
Probably
yep.
C
I
I
did
am
I
misreading
and
that
he
did.
The
thing
comment
from
18
days
ago
is
saying
that
he
did
that
or
is
that
still
pending.
B
C
Actually,
right
above
that,
there's
complete
recursive,
open
door,
sync
and
dirt
changes
as
well,
so
yeah,
that's
I
I.
I
was
confused
by
that
yeah.
B
C
C
B
Someone
went
back
and
switched
copy
dirty
using
opener,
and
it
just
worked
as
a
drop-in
replacement.
B
C
Yeah
same,
I
didn't
know
that
existed
cool
yeah,
I
I
do
agree
and
I'll
comment
this
on.
I
I
think
we
need
we
should
land
with
the
open
door
or
sorry,
not
there.
The
promise
api.
I
think
that
we
should
I
I
might
block
on
that,
but
it
depends
on
if
even's,
able
or
willing
to.
I
guess,
because
you
know
I
don't
want
to
block
the
feature
landing
at
all,
but
I
I
think
we
should
have
that.
D
C
A
Nope,
okay,
so
next
on
the
agenda
is
recursive
copy.
B
I
seem
to
have
buy-in
if
I
want
to
make
it
if
I
want
to
call
it
stable,
so
it
might
just
be
worth
going
back
and
calling
it
stable.
Oh,
the
reason
I
was
saying
I
might
not
call
it
stable
was
because
someone
had
had
suggested
on
twitter.
We
switched
it
to
opener
and
then
I
was
like.
Oh,
I
should
go,
switch
it
to
opener,
and
then
someone
had
already
done
it
like
a
few
weeks
ago.
So
I
don't
think
there's
any
real
blocker
on
calling
it
stable.
A
I
can't
remember
exactly
what
happened
there.
It's
like
it
was
like
a
while
ago,
like
a
few
months
ago,
and
we
weren't
really
able
to
get
a
good
reproduction
from
them.
C
B
A
B
Yeah
james
snell
said
that
he'd
probably
be
okay
with
it
going
to
stable
without
supporting
buffers,
because
it's
kind
of
a
fringe
feature
and
he
thought
it
would
be
like
wait
until
someone
says
they
need
it.
And
then,
if
someone
has
a
good
cause
for
it,
then
because
it's
not
going
to
be
trivial
to
do
so,
it
might
be
better
to
do
it
when
we're
asked
to
do
it
and.
B
B
The
file
system
join
method,
that's
used
to
do
recursive
stuff.
So
if
you're
doing
recursive
rm
you're
like
joining
a
bunch
of
paths
together
to
expand
out
to
the
recursive
depth
of
the
folder
that
doesn't
work
properly,
I
would
say,
probably
in
rimroaf,
if
you're,
using
a
buffer
for
the
path.
Instead
of
the
a
string
for
the
path
and
and
most
of
our
file
systems
happen
to
support
calls
happen
to
support
a
buffer.
B
I
think
it
has
to
do
with
non-uh
character
sets,
but
we
we
support
utf-8
strings.
So
I'm
not
quite
sure
what
non-latin
file
systems
can't
do.
Unicode
paths-
I'm
I'm
not!
Actually
I
don't
know
how
to
make
this
a
requirement
for
us
or
how
to
test
for
it.
Basically,
like.
I
think
you
need
to
have
a
certain.
B
You
need
to
formatted
your
file
system
in
a
certain
way,
and
it's
probably-
and
I
think
it
had
to
do
with
like
some
like
some,
like
maybe
japanese
character
sets
had
issues,
so
I
think
we
could
probably
go
to
stable,
and
but
we
should
keep
our
eye
open.
If
anyone
runs
into
this
issue-
and
we
know
what
it
is-
and
we
can
probably
fix
it
pretty
fast-
which
is
going
to
the
trouble
of
adding
buffer
support.
A
It
would
be
good
if
we
do
want
to
make
that
not
experimental.
We
could
do
that
before.
I
guess
lts
for
note.
16
isn't
for
a
while.
B
Did
you
want
to
get
it
in
for
note
18,
because
that's
soon
right,
that's
april.
C
Happens,
12
or
no,
the
the
new
lts
branch
is
cub,
but
it's
not
lts
yet
so
the
new
lts
release
so
20
becomes
current.
C
B
So
that's
when
you
like,
if
it's
best
to
get
in
at
the
very
first
release.
Otherwise
you
have
bugs
for
the
next
five
years
of
people
who
have
somehow
installed
a
pre-lts
version
of
the
platform
and
complaint,
like
I'm
running
version,
18.0.0
beta0-
and
it
doesn't
have
this
feature.
Yet.
Why
not
right.
B
A
Okay,
anything
else
on
that,
or
should
we
move
on
move
on
yeah,
okay,
so
last
and
but
not
least,
on
the
agenda
as
per
usual
is
argument.
Parsing.
B
A
Yeah,
a
big
increase
in
activity
on
that
repo,
which
is
good
but
yeah
lots
of
discussion
points
being
raised.
B
B
I'm
just
thinking
of
darcy's
replied,
I
think
darcy's
been
quite
busy.
Aaron's
made
a
case
for
us
going
away
switching
to
using
an
an
options
bag
for
configuring.
The
argument
parser
yeah.
B
I
think
it
solves
some
problems
to
be
perfectly
honest
because,
like
it
would
let
us
do
strict
mode
which
jordan,
lj
harp
has
been
making
a
really
strong
case
for
strict
mode
and
he's
kind
of
convinced
to
me
that
strict
mode
might
be
the
right
way
to
go
and
I'll.
Tell
you
why?
Because,
because
we're
requiring
that
you
specify
with
with
value
like
we
kind
of
require
that
you
we
kind
of,
require
that
you
say
what
you
want,
because
we
require
that
you
say
these
options
accept
a
value.
B
These
options
accept
a
flag,
we're
kind
of.
Are
we
requiring
that
of
a
person,
so
strict
mode
kind
of
makes
sense
at
that
point,
because,
unless
you've
defined,
what
you
accept
your
program
is
going
to
not
behave,
maybe
as
intended
or
it
would
be
rare
for
someone
not
to
define
the
options
they
need.
I
think
I
don't
know,
maybe
there's
an
argument.
I
guess
there's
an
argument
to
be
had
around
strict
mode
and
there's
an
art
and
there's
an
argument.
B
We
had
around
whether
you
should
configure
with
an
options
bag
or
a
that's
like
the
way
we're
doing
it
now,
where
you
have
the
key.
That
is
an
array,
basically
like
the
width
value
and
it's
an
array,
the
there's
a
few
other
things
that
take
an
array.
B
Oh,
I
remember
the
argument
if
we
do
support
strict
mode
which
lj
harv
is
making
a
strong
case,
for
you
need
a
way
to
define
flag
arguments
as
well
as
arguments
that
take
values,
because
we
don't
really
have
a
way
today.
If,
if
we
do
strict
mode,
we
now
immediately
need
to
be
able
to
define
both
all
our
arguments
that
take
values
and
all
our
arguments
that
take
take
booleans,
or
else
you
can't
have
booleans
in
strict
mode
right.
B
So
at
that
point
you
need
like
a
width
value
array
and
like
a
with
flag
array
which
starts.
It
starts
to
maybe
be
just
nice
to
define
all
the
arguments
and
say
what
they
are
like.
Is
it
a
boolean
or
is
it
a?
Is
it
a
string
like
maybe
have
a
type
argument
instead,
anyways
aaron
makes
a
case
in
that
issue.
B
There
makes
the
case
in
a
few
places
the
link
to
issue
number
45,
his
pr
number
63.
B
And
I
don't
know
he's
making
a
good
case.
I
think
people
should
join
the
discussion
on
github.
B
C
No,
it
just
looks
like
it's
an
object
that
has
some
properties
that
are
arrays
rather
than
dividing
the
the
commands
themselves.
B
Exactly
which
I
kind
of
agree,
it
might
be
more
ergonomic
like.
I
think
that
was
a
interesting
design
decision
and
I
think
it
kind
of
grew
out
of
like
how
simple
can
we
make
this,
and
then
it
turns
out
that
as
soon
as
a
few
people
trying
to
use
it
make
a
few
suggestions,
it
starts
to
get
kind
of
out
of
control,
because
you
don't
have
to
have
more
and
more
fields
that
have
something
in
an
array
right.
B
C
Finding
them
in
three
different
places,
yeah
yeah,
I
kind
of
agree
with
this.
It's
also
just
like.
I
don't
know
thinking
about
this
from
like
how
I
would
try
to
have
this
explained
to
someone
who's.
Just
starting
programming
and,
like
you
know,
clies
are
like
a
first
step
in
node.
I
feel
like
having
it
explained.
Once
is
like
the
more
logical
thing:
it's
something
that'll
make
be
easier
to
like
rock
conceptually
for
them,
and
then
I
think
in
terms
of
long-term
maintenance.
It
also
kind
of.
C
It
centralizes
things
a
bit
more
closely,
which
is
good.
B
Yeah,
I
tend
to
agree
too.
I
I
think
what
it
does
mean,
though,
is
we
need
like
you
to
chime
in
joe,
and
we
need
to
get
try
to
get
darcy
to
chime
in
and
roy,
maybe
because,
because
everyone's,
like
everyone,
kind
of
decided
on
how
they
was
going
to
take
options
and
now
we're
kind
of
if
this
is
a
pretty
major
change
and
how
it
would
take
options
right
so
yeah.
B
So
it
makes
me
think
that
trying
to
rush
this
over
the
finish
line
in
a
month
might
not
be
likely.
So
maybe
node
19
is
more
likely
yeah,
which
yeah.
C
So
I
think
to
answer
the
question
earlier:
no
19
ships
in
october
18
becomes
lts,
then
and
then
20
ships
in
next
april
and
then
yeah.
C
I
mean
it'd
be
nice
if
we
could
get
it
done
before
april,
but
I
I
don't
think
we
will
but
it'd
be
nice.
I
think
we
can.
We
can
try
to
aim
for
that
and
if
it's
not
going
to
happen,
that's
fine,
like.
B
We're
february
20th
today
I
mean
we
have
an
entire
month.
I
think
what
we
need
is
to
to
prod
darcy
or
roy
and
get
them
to
get
in
in
the
mix
there.
I
think
yeah
and
lj
harp's
been
pitching
in
a
ton
and
like
I
think
I
would
happily
compromise
on
the
strict
thing
personally,
like
he's
kind
of
convinced
me
about
this,
maybe
strict
being
the
default
behavior,
there's
another
person
making
a
strong
case
for
strict
as
a
default
behavior.
C
C
B
C
B
So
that's
terrifying.
It
means
your
arc.
It
means
your
argument.
Parser
is
literally
just
like
one
line
with
no
configuration
which
is
kind
of
a
neat.
Look
if
you're
writing
the
simplest
of
program
where
you're
like
just
you
just
want
someone
to
pass
in
the
flag,
h
or
something
yeah,
you
don't
even
have
to
define
that
you
take
the
flag
h,
you
just
parse
it,
and
if
someone
runs
your
program
passing
in
an
h,
it
will
do
what
you
expected
it
to
do.
B
The
inverse
argument
that
jordan's
making
a
case
for
and
as
well
someone
from
tc39
is
actually
making
up
giving
us
a
lot
of
feedback
right
now
to
someone
I
know
from
pc39:
what's
their
name
kevin,
gibbons,
no
nice!
His
is
making
a
strong
case
for
having
to
define
stuff
as
well.
B
B
That
makes
sense
yeah.
That
makes
sense,
and
I
guess
so
I
guess
the
real
question
is:
is
it
strict
by
default
or
loose
by
default,
two
people
making
a
two
people
of
making
a
strong
argument
for
strict?
I
would
go
either
way
because
I
want
to
get
this
done.
Yeah.
B
C
C
Yeah,
I
I
I
am,
I'm
fine
with
strict.
I
think
I
would.
I
agree
that
I
would
like
to
hear
from
roy
and
darcy.
I
also
think
that,
for
I
mean
I
know
that
they're,
you
know
they're
explicitly
interested
in
using
this
and
I
think
they're
fine
to
set
whatever,
as
whatever
they
want,
as
they're
fine
to
write
that
code.
So
they
don't
probably
don't
care
yeah,
at
least.
B
Yeah,
well,
we
just
like
if
anyone
wants
to
make
a
really
strong
case
for
loose
by
default,
and
yes,
then
someone
can
make
that
case,
but
I
would
compromise
on
it
and
then
the
one
place
I
won't
compromise
is
that
I'm
gonna
push
on
is
like.
I
don't
think
we
should
have
a
requirement
that,
like
mpm's
already
adopted
it
before,
we
pull
it
into
node.
D
D
Mean
I
think
that
was
just
partially
because
darcy
was
saying
he
wanted
to
adopt
at
asap
the
the
sentiment
was.
It
would
be
good
to
have
some
people
using
it
for
when
we,
you
know,
propose
it,
but
I
don't
think
it
needs
to
be
a
strict
requirement
per
se.
B
I
think
I
was
also
thinking
I'd
kind
of
like
to
move
c8,
which
is
a
command
liner
run
for
I
make
for
argument
for
code
coverage.
I
was
thinking
I
kind
of
want
to
try
to
move
it
to
parse
args
that'd
be
cool
because
because
it
would
get
rid
of
most
of
the
dependencies
in
c8,
because
a
ton
of
the
dependencies
are
for
yards.
B
Oh,
this
is
a
complete
non-sequitur,
but
I
actually
had
it.
I
had
another
really
cool
idea
too,
which
was
someone
was
suggesting
that
yarg's
just
be
able
to
generate
the
help
documentation
like
like,
so
you
could
just
generate
what
the
command
line
options
available
would
look
like,
but
I
was
thinking
I
could
do
that
to
pre-generate
what
the
hell
message
would
look
like
with
something
like
yargs
then,
just
put
it
put
it
inside
a
program
like
c8.
B
That
has
no
dependencies
and
just
put
it
in
template
strings,
but
you
can
generate
the
template
string.
That
will
be
your
command
line.
Our
application
for
with
some
secondary
program
like
yargs
so
like
your
actual
cli,
has
no
dependencies
but
you're,
except
for
parse
arcs,
but.
B
C
B
D
B
C
Don't
think
we
should
have
the
to
be
different,
okay,
yeah,
I
don't
think
we
should
have
that
to
be
different.
I
think
I
mean
what
I
do
think
we
should
do
is
we
should
try
to
make
it
make
sure
that
there,
the
hooks
necessary
for
someone
to
be
able
to
build
that
are
available
so
like
we're,
not
basically
allow
make
sure
we're
exposing
what
we
need
to
if
we
need
to
expose
anything
to
allow
someone
to
hook
in
and
say
like
cool.
I
want
to
you
know,
here's
all
the
arguments.
C
Let
me
process
this
and
make
a
pretty
string.
Yeah,
that's
a
good
point!
Yeah.
I
think
we
should
try
to
make
sure
we
do
that.
I
think.
A
B
C
B
C
Yeah
that
makes
sense
and
like
being
able
to
just
make
that
help
output
thing
as
a
dev
depth
and
then
like
build
that
as
output
is,
I
think,
a
a
good
goal
for
us.
I,
I
don't
think
that
needs
to
that
doesn't
need
to
ship,
yet
I'm
sure
we
can
figure
that
out,
but
making
that
great
making
sure
we
have.
That
in
mind.
I
think,
is
good.
B
B
D
Yeah
well,
let's,
let's,
let's,
let's
aim
for
april,
I'm
gonna
have
a
little
bit
more
free
time.
My
my
calendar
is
freed
up
a
little
bit
more,
so
I
will
definitely
try
to
look
at
what's
going
on
here
and
move
things
forward
and
maybe,
as
I
get
a
better
handle,
I
can
ping
darcy
with
very
specific
things
to
look
at
and
see.
If
he's
got,
you
know
two
minutes
or
ten
minutes
to
to
pop
in
and
check.
You
know
comment
on
bits.
B
C
D
B
D
To
be
like
working
through
it,
but
some
of
the
things
that
we're
thinking
are,
you
know
the
directions
we'd
like
to
take
and
the
decisions
based
on
some
of
these
comments
and
pr's.
We
could
document
that
too
it'd
be
helpful
for
me
when
I'm
trying
to
dig
in
further
as
well
cool.
B
If
you
look
at
some
of
the
closed
prs
too,
I
think
there's
some
good
information
because,
like
like,
I
added
support
for
negative
numbers
so
that
I
could
implement
that
feature
in
yards
and
a
couple.
People
make
a
strong
case
not
to
have
negative
numbers,
and
I'm
like
okay,
I'll,
just
I'll,
try
to
implement
that
in
yards.
Like
I
don't
think
I
need
it.
B
I'm
going
to
try
to
figure
out
how
to
do
it,
a
step
up
from
parser.
This
is
what
I'm
getting
at,
like
the
idea
that
you
can
have
some
of
these
things
built
up
a
layer
and
but
but
that
issue,
where
we
kind
of
just
argue
about
whether
or
not
we
should
support
negative
numbers
is
interesting,
because
that
argument
will
probably
come
up
again
with
people
right
so
yeah,
who
is
it
shadows?
Is
it
a
shadow
spawn?
Is
that
the
that's?
Yes,
not
his
actual
name
for
sure?
B
D
B
Yeah
so
he's
been
really
helpful
and
I
like
that.
He
I
like
that.
I
like
that
he's
been
making
the
case
to
when
in
doubt
and
try
to
do
stuff
similar
to
possex,
because
unix
has
been
a
thing
for
45
years
or
something
so.
A
With
some
of
our
other
stuff
too,
like
cp
and
things
like
that,
so
yeah.
B
D
B
B
B
C
Just
so
we
can
have
like
this
ambiguation
there
don't
have
to
like,
if
something's
already
flagged
into
that,
we
can
know
that
this
is
like
cool.
We
don't
need
to
worry
about
this.
Quite
we
can
worry
about
it,
but
we
don't
need
to
worry
about
quite
yet.
C
That
sounds
good
cool
awesome.
I
think
I'm
gonna.
C
B
B
D
A
B
Anyways,
let's
change
my
sleep
schedule
a
bit
yeah
all
right.
Well,
this
is
great.
I
think
this
is
a
really
productive
conversation
about
the
argument
parsing.
So
it
sounds
like
we're,
keep
saying
we're
on
the
home
stretch,
but
I
think
we're
actually
on
a
home
stretch
because,
like
playing
with
it,
has
been
working
well.
So
let's
just
get
folks
to
agree
on
a
few
more
things.