►
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
On
youtube:
okay:
here
we
are
august
21st,
nodejs,
tooling,
group,
meeting
hi
everybody.
So
does
anybody
have
any
sort
of
like
announcement
type
thing
that
they
just
wanted
to
make
people
aware
of.
A
Okay,
so
if
not,
I
think
we
can
start
looking
at
the
agenda.
So
I
remember
last
meeting
we
started
talking
about
the
the
configuration
file
location
business
and
there
was
a
lot
of
chat.
We
spent
probably
half
the
meeting
just
kind
of
hammering
on
that
one,
and
I'm
thinking
that
you
know
maybe
we
should
just
do
a
deep
dive
on
that
one
day.
Just
because
there's
a
there's
a
lot
of
different
ideas,
there's
a
lot
of
different
ways
to
approach
approach,
the
problem
and
yeah.
A
A
Any
objections
there,
all
right
so
yeah,
let's,
let's
just
schedule
a
deep
dive
for
that-
and
I
think
before
that-
and
so
we'll
probably
want
to
come
up
with
some
sort
of
agenda
for
it
and
maybe
list
out
the
various
like
ideas
and
proposals
and
in
some
sort
of
way
that
that
makes
sense
and
try
to
try
to
just
have
a
little
bit
of
structure
for
it.
But
on
the
topic
of
deep
dive,
so
I
did
schedule
a
deep
dive.
A
A
If
you
know
someone
who
cares
about
it,
maybe
bother
them
to
show
up
or
something
because
that's
what
we
will
be
talking
about
next
meeting,
which
is,
I
think,
september
4th,
and
so
what
I
think,
maybe
that
makes
sense,
is
the
next
meeting
after
the
fourth
is
to
have
a
normal
meeting
and
then
the
meaning
after
that,
which
would
be,
I
don't
know,
math,
so
something
just
late
late
september-
would
be
this
deep
dive
for
the
config
file
stuff.
A
If
anybody
wants
to
try
to
do
something
sooner,
you
know
we
could
try
to
work
out
scheduling
for
that,
but
it's
just
kind
of
easy
to
just
take
the
meeting
time
we
already
have
so
yeah.
Does
that
sound
cool
super
okay?
So
that,
like
knocks
off
three
things
on
the
agenda,
unless
anybody
else
has
anything
they
want
to
talk
about
on,
unlike
deep
dive
stuff
or
their
cursive
file
system,
deep
dive
meeting,
I
think
what
we
should
probably
do
again
is
is
before
next
meeting.
A
Somebody
should
work
on
an
agenda
for
that
one.
So
it's
not
just
you
know
free-for-all.
A
So-
and
I
don't
know
if
anybody
wants
to
volunteer
to
propose
an
agenda,
if,
if
not,
I
can
probably
do
it
just
kind
of
come
up
with
something.
A
Okay,
so
moving
on
then
so
the
next
one
we
have
is
the
ffi
stuff.
Last
meeting
it
was.
A
Let's
see
brian
said,
gus
is
working
on
some
stuff.
I
don't
know,
if
guess
kaplan,
I
don't
know
if
any.
If
you
have,
if
you
know
of
any
updates
there,
brian.
B
Yeah
pretty
much
nothing
changed
since
last
week
I've
been
working
on
other
things,
god's
been
working
on
other
things,
so
yeah,
that's
no
progress
there.
This
week.
A
Okay,
does
anybody
know
about
the
ch,
mod
r
stuff.
C
I
feel,
like
I
mean
we
did
a
spike
darcy
and
I
did
a
spike
on
it
and
it
kind
of,
I
think,
fell
off
a
little
bit.
C
Maybe
we
could
make
that
one
of
the
agenda
items
on
the
deep
dive
on
the
file
stuff.
I'm
actually
way
more
excited
about
copy
recursive
honestly.
So
if
we
drop,
if
there
are
like
major
concerns
and
security
around
permission
changes,
maybe
we
just
choose
not
to
do
that.
Like
we
actively
say
this
is
a
function,
we're
not
going
to
implement
or
something
that
might
be
a
way
to
approach
it.
A
Okay,
then
the
next
one
is
the
esm:
module
reloading
module
graph
sort
of
thing,
and
so
I
think
this
was
yesterday
was
this
yesterday
brian
was
like.
Why
can't
I
do
this
and
I'm
like
brian,
have
you
even
been
paying
attention?
Oh.
A
Sorry,
so
it
it
no,
it
was
funny,
but
it
it
brought
up
some
stuff.
So
I
I
linked
the
thread
in
the
in
the
minutes,
but
in
that
thread
I
posted
a
link,
which
I
should
probably
post
here
too,
to
a
blog
post
on
dev
by
gill,
who
showed
how
you
can
use
a
use
test,
double
which
is
a
mocking
library
to
to
do
this.
A
What
it
comes
down
to
is,
you
have
to
use
a
loader,
so,
okay,
so
loaders
are
experimental.
As
far
as
I
understand,
you
can
only
use
one
loader,
which
is
a
drawback,
and
I
believe
that
there's
room
for
improvement
there.
A
It
sounded
like
from
that
thread
like
people
still
wanted
to
to
make
some
changes
to
it
and
improve
it
or
gather
feedback,
but
in
order
to
do
this
to
make
it
happen,
you
have
to
use
this
like
a
special
loader
and
essentially
what
their
implement
implementation
does,
and
I
didn't
look
too
hard
at
it,
but
it
looks
like
it
essentially
like
injects
a
header
into
every
module
and
or
something
like
that,
and
it
stuffs
all
of
this.
A
This,
like
crap
into
the
global
into
a
global
somewhere,
and
then
it
uses
that
that
global
to
to
like
intercept
these.
These
calls-
and
I
don't
I
mean
I
haven't-
played
with
it-
I
don't
know
how
well
it
works
or
if
there
are
limitations
to
their
approach,
but
it
is
something
that
you
can
do.
A
That
said,
my
like
the
main
thing
I
was
trying
to
get
at
was:
yes,
okay,
you
can
do
this.
You
have
to
use
this
loader
thing.
If
you
want
to
use
another
loader
right
now,
you
cannot
you
you
can
choose
to
mock
it
mock
at
the
module
level
or
not,
because
if
you
need
to
use
another
loader
you're
out
of
luck,
so
I
was,
I
was
basically
saying
like
look.
We
could
do
what
we
did
with
with
module
level
mocks
and
and
and
reloading
modules
in
common
js.
It's
just
that.
A
The
experience
now
with
esm
is
is
like
it's
downgraded.
It's
not
as
ergonomic
it's
it's
like
painful,
it's
more
painful
than
than
it
was
with
using
common
js,
and
so
the
there
were
a
couple
proposals
in
the
twitter
thread.
I
mean
it's
they're,
not
like
proposal
proposals,
but
just
like
ideas
like
the
one
is:
let's
allow
multiple
loaders,
the
other
one,
which
would
be
really
nice
for
a
test
framework.
A
Is
the
ability
to
basically
use
a
loader
but
call
it
programmatically
so,
instead
of
needing
to
pass
a
flag
on
the
command
line,
your
test
framework
or
just
some
library,
essentially
that
you
import
would
be
able
to
just
like
register
its
loader
by
just
you
know,
importing
the
module
and
calling
a
function
or
something
like
that,
and
I
I
don't
know
the
like
how
feasible
that
is
it
there's
most.
A
Certainly
some
issues
with
you
know
loading
part
of
the
program
and
then
loading
this
loader
and
then
loading
everything
else
with
the
loader
there's,
probably
some
like
problems
there.
I
don't
know
about
the
implementation,
but
I
just
intuition
suggests
that
it's
probably
not
that
easy
to
to
and
and
it's
going
to
like
introduce
weirdness
and
maybe
some
unexpected
behavior,
and
I
expect
there
will
be
some
concerns
there.
But
you
know
if
we,
if
we
were
able
to
get
this
done
with
a
loader
and
if
a
test
framework
or
a
library.
A
So
if
you
want
to
do
this,
mocking
you're
going
to
want
a
library
right
and
if
a
library
can
can
just
like
register
its
loader
and
do
its
thing,
then
I
think
we've
gotten.
You
know
obviously
without
having
tried
this
this
this
this
test,
double
thing.
You
know
it
seems
like
it
would
get
us
pretty
much
to
where
we
want
to
go.
It
seems
like
it
would
give
a
a
very
similar,
a
similar
experience
for
developers
who
want
to
do
this
sort
of
thing
to
to
do
with
loaders.
A
B
B
It
just
seems
like
that.
Isn't
going
to
happen,
I
I
don't
know
like
to
me.
It
seems
the
biggest
problem
is
still
reloading
right.
A
I
mean
it
sounded
like
the
this.
Your
your
test
double
would
allow
you
to
do
stuff
like
that,
I'm
going
to
have
to.
I
want
to
take
a
look
at
it.
I
just
I
haven't
had
time
between
then
and
now,
but
I
want
to
see
like
what
exactly
can
you
do
with
this
thing?
B
I
think
I
think
there's
like
if
adding
the
flag
is
in
the
issue,
and
so
something
is
provided.
Programmatically
then
there's
also
the
question
of
whether
or
not
it's
going
to
work
with,
with
the
loaders
being
on
a
separate
thread.
If
that's
the
thing
that
ever
happens,
the
other
thing
is
that
test
double
and
it's
underlying
library
which
they
might
forget,
all
of
a
sudden,
relies
on
the
communication
between
the
loader
and
the
code
running
in
the
rest
of
the
process.
B
A
Is
so
loader
isn't
a
worker
is
this?
Is
this
a
thing
I've
never
heard
of
this.
B
A
A
Yeah,
I
think
I
think
that's
I
mean
that
seems
fair
yeah,
it's
hard
to
say
much
more
without
without
actually
having
used
it
and
see
and
and
understood
what
it
can
and
can't
do,
and
but
it
would
be
nice
to
not
merge
anything.
That's
going
to
break
it
for
sure
it
is
an
experimental
api.
Yes,
but
it
you
know
it's
not
the
concern
isn't
breaking
that
particular
library.
The
concern
is
breaking
the
entire
use
case.
B
A
A
Okay,
so
ben,
are
you
do
you?
Do
you
still
wanna?
Do
you
still
wanna
go
like
try
to
crash
that
meeting?
Otherwise,
if,
if
you're
not
able
to
do
it
next
week
on
their
meaning
is
like
on
wednesday,
I
can
just
go
and
I
will
go
ahead
and
take
a
closer
look
at
test,
double
and
and
explain.
This
is
what
they're
doing
with
it,
and
you
know
this
is
what
it
can
do.
A
C
C
A
C
E
D
B
Post
the
link
to
that
in
the
chat
there
yeah,
so
it's
not
clear
whether
that'll
be
at
odds
with
this
other
pr
for
worker
threads
or
what
or
something
like
that.
B
But
it's
it's
worth
noting
that
that's
in
there
too,
and
that's
especially
relevant
in
the
case
where,
for
example,
somebody
who's
doing
a
lot
of
shimming
say
an
apm
tool
may
need
a
loader
and
even
testing
that
loader
is
gonna.
Require
mocking
so
like
that's
gonna
require
training
right.
So
that's
the
thing.
That's
gonna
need
to
happen.
So
that's
that's
kind
of
related.
I
think.
A
With
it's
not
just
to
be
clear,
it's
not
just
multiple
loaders,
it's
that
they
have
that
things
have
to
happen
in
an
order
right
and
that's
why
you
say
chaining,
because
things
are
sequential.
You
have
to
do
this
and
then
this
and
then
this
you
can't
just
like
try
to
run
them
all
at
once
or
in
some
like
non-deterministic
order,
and
then
you
get.
Who
knows
what
happens
right
because
they
they
they
like
take
code
right,
so
they
take
source
code
and
they
return
other
source
code.
Is
that
essentially,
like.
B
Right
right,
exactly
the
the
order,
absolutely
matters,
it's
I
kind
of
liken
it
to
like
connect
middlewares.
C
Rule
the
with
require
stacking
of
requires
has
been
surprisingly
stable
in
the
community,
because
there
are
these
interdependencies
and
like
a
lot
of
the
time
it
they
seem
to
work.
Okay,
I
guess
it
kind
of
the
burden
with
the
require
extensions
is
often
like
on
the
library,
tooling
authors,
because
they'll
get
a
bug
report.
That's
like
coverage
didn't
work
with,
I
don't
know
pretty
stack
traces
or
something,
and
then
this
pretty
stack
trace
author
will
make
it
so
there's
happens
after
coverage
or
something
right.
So
this
would
be
better
to
define
that
more
clearly.
A
A
C
No
well,
I
can't
the
name
of
the
the
library
is,
but
basically
sometimes
you
will
have
a
chain
of
people
all
trying
to
hook
the
dot
js
require
this.
This
already
exists
in
the
community
and
and
then
like.
What's
basically
happened
is
there's
some
there's
some
libraries
written
in
the
community
that
can't
handle
creating
a
stack
of
people
who
run
on
top
of
that
extension.
Instead
of
so,
it
sounds
really
similar
to
the
problem
of
like
how
you
resolve
the
order
of
loaders.
A
A
How
does
that
work?
And
I
don't
think
no
nobody
knew
in
that
thread,
but
I'm
sure
somebody
does
have
you
have
you
played
with
loaders
at
all
ben
with
your
your
source
map,
stuff.
C
I
haven't
really,
I
can
speak
kind
of
like
to
and
like
the
past
world
that
we
lived
in
and-
and
you
would
end
up
in
these-
you
actually
have
to
resolve
two
steps
of
source
maps
before
you
get
back
to
the
original
code,
so
you'd
be
in
a
world
where,
like
ts,
node,
adds
a
source
map
for
the
compilation
from
typescript
to
javascript,
but
then
we're
instrumenting,
the
javascript,
so
nyc
adds
a
source
map
for
the
instrument
is
for
the
compiled
javascript
code
to
the
instrumented
javascript
code
and
now
to
get
from
point
c
back
to
the
original
code.
C
You
actually
have
to
apply
two
source
map
transformations,
so
yeah,
it's
a
little
bit
of
a
nightmare,
but
it
works
today
like,
but
you
do
need
to.
Basically,
if
you
have
three
transform,
if
you
had
three
transformers
that
were
all
doing
some
sort
of
code
change,
you
need
to
emit
a
source
map
out
of
each
layer.
A
C
I
I
mean
presumably
like
a
loader,
would
be
using
some
existing
thing
like
roll
uppers
or
like
in
istanbul
instrument
code
or
like
some
probably
it's
plug-in
some
library
that
takes
code
a
and
puts
out
code
b,
and
usually
it's
that
library
that
that
most
libraries
that
do
any
type
of
transformation
like
that
are
going
to
have
an
option
to
emit
a
source
map.
C
A
I
mean
it
sounds
like
yes,
so
test
double,
for
instance,
it
it
transforms
source
and
I
did,
but
I
didn't
see
anything
about
source
maps,
and
I
don't
know
if
they
they
considered
that,
but
that's
gonna,
that's
gonna
gum
it
up
if
they
don't
have
any
way
to
like
provide
a
source
map.
Somehow
and-
and
I
don't
know
like
how
you
would
do
that
from
a
loader.
C
Yeah,
no,
I
do
and
like
you've
brought
up
an
actual
valid
you've
brought
up
a
valid
point
like
because
you
don't
you
need
some.
You
need
some
sort
of
cache
that
knows
that
that
this
code
actually
went
through
two
transformation
steps
so
like
yeah,
maybe
node.js's
source
map
implementation
would
need
to
be
setting
aside,
like
two
source
maps
with
two
different
identifiers,
one
for
step
a
of
the
transformation
one
yeah.
I
haven't
thought
about
this
too
deeply.
It's
a
hard
problem.
I
feel.
F
Yeah,
maybe
I
don't
have
a
good
answer,
so
I
just
joined
I'm
listening.
A
C
Yeah
yeah,
I
don't
know
it's
a
hard
problem
you
need
to
like
it
could
be
like
it
could
be.
The
loader
themselves
is
using
a
library
that
munches
the
source
maps
together
on
each
step,
so
you
do
end
up
with
one
single
source
map
that
can
go
back
to
truth,
but
but
you
need
some
way
to
like
having
three
source
maps
is
a
nightmare
but
building
up
a
source
map
that
is
able
to
trace
from
your
end
transformation.
C
So
a
hard
problem
it
feels
like
it
would
have
to
maybe
be.
I
think
you
would
maybe
fix
it
at
the
loader
level
by
trying
to
get
to
one
source
map
by
the
end,
and
probably
that
feels
like
it
would
be
done
in
the
community
and
not
in
node.js
like
it
feels
like
you'd
have
some
something
like
pirates
or
someone
writing
a
loader
has
some
some
common
approach,
that's
being
used
to
emit
one
source
map
instead
of
three.
C
F
So
I'm
unfamiliar,
I,
I
wrote
a
vs
code
extension
which
did
do
nested
source
maps.
Does
chrome
devtools,
currently
not
unwrap
inline
source
maps
multiple
times?
Does
it
only
do
one
level?
I
don't
know.
That's.
C
F
C
Map
gotcha,
but
but
here's
the
interesting
thing,
though
right
like
if
you're
going
through
a
loader
extension
and
say
there's
two
three
loaders
say:
there's
two
loader
extensions:
loader
extension
a
performs
some
modification
to
the
code,
so
it's
adding
it's
transforming
from
typescript
into
javascript,
but
then
there's
a
second
loader
before
it
ends
up
in
the
module
cache
that
adds
coverage.
C
F
Unwrapping
to
so,
there
are
a
variety
of
ways
this
has
been
done
in
the
past.
The
sources
field
of
source
maps
is
how
I've
always
dealt
with
it.
Originally
out
of
band
files,
don't
really
work
so
great
in
my
experience
and
you
really
do
not
want
to
put
a
whole
sources
field
in
your
generated
source.
F
If
you're
lucky,
because
it
has
to
be
encoded
with
that
weird
stuff.
C
C
C
F
C
D
F
So
really,
the
problem
is
where
to
store
the
source
maps
and
associated
sources.
To
my
knowledge,
currently,
the
v8
api
does
have
the
ability
for
us
to
alter
the
perceived
source
map
of
any
given
thing
we
put
into
the
compiler.
C
F
Even
if
we
do
that,
I
don't
think
we'll
solve
the
multi-stage
transformations
that
I
don't
think
we
can
automate
that
reliably.
I
agree
that
something
like
pirates
or
other
things
that
do
nesting
properly.
It's
likely
going
to
be
the
future.
A
C
Then
well,
I
think
what
you'd
end
up
with
like,
I
think
what
pirates
is
essentially
saying
is
like
anyone
who
wants
to
override
require
extensions
has
to
go
through
a
third-party
dependency
to
do
so
so
you're,
basically
just
creating
a
bottleneck
where,
where
the
composition
happens
in
one
place,
so
I
mean
what
could
happen
is
there's
just
a
loader.
That's
a
loader!
That's
designed
for
composing
other
loaders
right,
yeah,.
F
Somebody
wrote
that
was
it
rich,
he
wrote
one
that
did
you
just
give
it
an
array,
basically
of
loaders
and
it
would
compose.
F
F
So
this
is
actually
fairly
important.
The
idea
of
composing
loaders
being
done
in
user
land
versus
in
core.
If
we
do
land
the
ability
to
specify
multiple
loaders
in
core,
you
may
not
be
able
to
enforce
that
bottleneck.
C
F
There's
a
slight
contention
there.
Overall,
we
are
leaning
towards
allowing
you
to
declare
multiple
loaders.
In
such
a
case,
though,
you
would
only
be
able
to
bottleneck
at
one
of
those
segments.
You
could
not
bottleneck
all
the
loaders
gotcha.
C
A
It
is
cool
that
you
can
get
you
can
solve.
At
least
some
of
this
stuff
with
loaders
is,
is
I
mean?
Is
it
the
right?
Is
the
right
thing
like?
Is
it
the
right
tool?
I
don't
know
it
may
be
the
only
tool
available
to
us
so.
F
So
one
thing
I
think
that
could
be
actionable
is
instead
of
purely
returning
source
and
format
from
loaders.
We
could
add
fields
that
you
want.
It
won't
solve
all
the
issues
but
proposing
something
like
a
source
map
field
that
we
can.
Let
you
directly
pipe
it
in
rather
than
inline
it
as
a
comment
at
the
end,
might
at
least
simplify
things.
A
Okay,
so
I
think
I
wanted
to
wrap
this
one
up.
Does
anybody
have
anything
else
they
want
to?
They
want
to
bring
up
on
the
esm,
module
reloading
topic.
A
All
right,
so
next
one
on
the
list
is
the
spawn
sink,
spawn
hooking
patching.
A
So
last
we
heard
that
we
knew
stephen
was
working
on
something
related
to
this,
like
this
ultimate
hook,
stuff
and
ben
mentioned
that
we
need
to
make
sure
steven
understands
like,
or
we
want
to
make
it
known
to
stephen
what
our
use
cases
are,
and
I
don't
know
if
anybody
has
talked
to
him
since
I
have
not.
C
An
action
item:
someone
should
actually
take
that
action.
I
know
yeah.
I
suppose
I
can
take
it.
If
does
anyone
else
have
no
I'll
mention
it
to
stephen?
I
think
someone
said
on
the
call
last
week
that
stephen
already
had
some
of
these
use
cases
on
mine,
so
I
can
just
reach
out
to
him
on
twitter
or
something
confirm.
A
Yeah,
I
know
he's
been
around
and
I
think
he
knows,
but
I'm
not
sure-
and
I
don't
have
any
current
visibility
into
what
he's
doing.
I
suppose
I
could
go
hunt
for
it,
but.
F
A
Okay,
well,
thank
you.
Ben
looks
like
this
better
way
to
detect
a
process
is
exiting.
Is
there
was
nothing
on
this
last
week
as
far
as
I
understood
and
corey's
not
here
this
week,
but
as
I
understood
like
he
had
found,
I
don't
know,
I
don't
actually
know
where
this
left
off,
but
there
was
no
notes
for
it
from
from
sorry
last
meeting.
C
I
can
give
a
little
context
that
so,
when
we
added
coverage
support
to
node,
I
basically
hacked
in
this
approach
to
catching
all
types
of
exits
with
be
they
signal
or
like
just
the
normal
process.
Dot
exit,
zero
or
one
or
whatever,
sorry,
just
choking
on
something.
C
Joyee
then
implemented
this
kind
of
in
a
better
way,
and-
and
I
think
it
ended
up
being
like
pretty
specific
to
how
we
approached
coverage
to
make
sure
we
wrote
the
content
to
disk,
regardless
of
how
we
were
exiting,
but
it
was
like
ended
up
being
like
I
think
someone
was
inside
the
c
layer.
It
wasn't
in
the
wasn't
like
something
easily
hookable
inside
userland,
so
it
would
be
nice
to
have
something
that
was
what
was
hookable
inside
of
userland
and
we
talked
about.
C
Maybe
it
being
another
thing,
you
could
just
have
an
online
process
that
was
like
on
like
any
exit
or
something
I
don't.
I
don't
know
what
it's
called,
but
there's
kind
of
code
kicking
around.
That
does
this
for
some
edge
cases
like
making
sure
coverage
reports
are
written
in
disk,
but
I
don't
think
it's
too
exposed
to
user
land.
A
Do
you
have
like
a
like
an
issue
or
something
or
just
like
a
change
set
or
something
you
could
point
to
with
like
what
what
she
did.
A
A
So
the
last
time
last
meeting
there
was
a
note
here
about
removing
the
source
map,
support
from
the
agenda
and
then
ian
said
something,
and
I
don't
know
if
something
about
interleaved
stack
traces.
And
so
I
don't
know
if
this
wants
to
stay
open
or.
C
I
feel
like
maybe
keep
it
open
because
there's
another
thing
that's
come
up,
which
is
like
trying
to
use
v8s
source
map
field
instead
of
kind
of
reinventing
it
ourselves
like
if
it's
practical,
because
they
do
have
the
option
to
set
a
source
map,
as
you
create
the
isolator,
whatever
it's
called.
The
other
thing
I
think
to
ian's
point
like
it
might
be
worth
syncing
with
well
quest
actually
and
wes
is
pure
ghee,
and
I
guess
jordan.
C
Jordan
was
working
on
this
too,
that
there
is
there.
It
was
like
a
little
bit
of
a
tc39
effort
to
standardize,
I
think
stack.
Traces
specifically
might
be
worth
talking
to
them
and
being
like
hey.
Is
there
like
a
slightly
more?
Is
there
a
standard
being
thought
of
around
how
we
should
print
the
stacked
trace,
because
then
we
could
follow
that,
and
at
least
speak
to
that.
So
I
don't
know,
I
don't
think
I'm
quite
ready
to
root
this
off
the
agenda.
E
C
A
A
C
C
C
C
A
I'll
reach
out
to
them,
or
maybe
even
send
a
pull,
request
myself
and
see
what
it
what
what
it
looks
like.
A
Okay,
anything
else
on
the
source
maps.
C
I
don't
think
so.
I
haven't
had
much
time
to
work
on
them.
I've
been
wanting
to
look
at
what
what
it
would
mean
to
set
that
plague
in
v8.
That's
it
less
bullish
about
removing
the
experimental
flag
for
some
of
the
reasons
we
just
brought
up
all
right.
A
Okay,
this
last
one
here
is
arc
parsing
and
I
believe
like
where
that
left
off
was
I'm
supposed
to
send
a
pr,
and
I
have
not
done
that
and
that's
all
I
know
I'm
busy.
Okay,
I
have
other
things.
C
A
Totally
hear
you
I've
other
things
that
I'm
supposed
to
be
doing
so
I
will
get
to
this
when
I
get
to
it
and
hopefully.
C
I
I
would
like
to
just
put
a
strong,
I
think
would
be
awesome,
because
I
think
it
will
let
us
experiment
with
some
of
those
other
neat
ideas.
We
have
like
being
able
to
run
a
little
http
server
with
arguments
as
a
built-in
node
thing,
that'd
be
cool
yeah.
C
A
All
right,
so
that
is
the
agenda
and
we
have
eight
minutes.
So
does
anybody
have
anything
that
they
want
to
bring
up
or
show
and
tell
or
whatever.
C
F
So
the
policies
is
gonna
start
moving
towards
unflagging
in
the
next
year.
I
don't
know
if
people
even
use
that,
but
it's
had
a
few
minor
changes
to
get
it
up
to
speed
with
recent
stuff,
so
it
should
be
more
usable,
especially
after
one
pr
lands
on
monday.
So
yeah.
A
It
is
you
know,
I
have
even
not
even
looked
at
policies
is
that
in
in
the
docs
already.
F
Yeah
it's
in
the
docs,
but
it's
kind
of
hard
to
generate
stuff
for
without
scoping
mechanisms,
which
is
what
lands
on
monday.
A
A
All
right,
and
so
that
will
make
policies
much
more
useful.
It
sounds
like
them.
F
So
it
doesn't
actually
add
much,
but
it
just
means
that
a
human
might
be
able
to
write
a
policy
file.
F
A
All
right:
well,
that's
cool!
We'll
check
that
out
yeah,
I
don't.
I
don't
have
anything
to
announce
really
I'm
just
kind
of
heads
down
on
just
just
some
some
mocha
stuff,
but
yeah.
So
yeah.
D
A
So
I
didn't
look
into
that.
I
had
cc'd
you
on
twitter
about
the
feedback
about
workspaces.
I
didn't
really
understand
what
the
problem
was.
D
A
Behaves
well,
if
you
don't
get
it,
I
can
reach
out
to
miroslav.
Is
my
colleague
at
ibm
and
I'll
I'll,
just
ask
him
what
the
hell
he
was
talking
about.