►
From YouTube: Node.js Tooling Group Meeting
Description
A
All
right,
so
here
we
are,
what
is
it?
It's
Friday
September
6th,
and
this
is
no
js'
tooling
group
meeting
we
haven't
met
for
a
while
I
needed
to
find
a
new
meeting
time.
So
here
we
are
I've
set
it
to
I.
Think
the
first
thing
I
want
to
go
over
is
and
now
and
even
know
if
it's
on
the
agenda,
but
the
meeting
time.
So,
given
the
fact
that
at
least
some
people
are
here,
it
can't
be
the
worst
time.
A
A
Pardon
me
feels
that
you
know,
maybe
a
more
frequent
meeting
would
be
better
just
to
kind
of
keep
everybody
a
little
more
engaged
because
I
know
like
at
the
end
of
the
the
three
weeks
I
it
just
kind
of
tends
to
fall
off
and
I
forget
about
stuff.
So
what?
What
do
you
think
is?
Is
this
doable
every
two
weeks
yeah.
B
C
A
Let's
see
here
so
the
big
thing
that
it
seemed
to
get
a
lot
of
attention.
The
last
week
was
glob
support,
in
course,
so
the
the
idea
is
that
oppose
questioning.
Do
you
think
core
should
support
globbing
or
parsing
of
essentially,
what
it
would
be
was
would
be
you
hand
it
a
string,
and
that
string
is
some
sort
of
glob
like
string
and
the
blob
parsing
thing.
It
goes
and
looks
at
the
string
and
parses
the
string
and
then
tries
to
convert
that
to
a
list
of
files.
A
There's
the
the
issue
I
believe
it
was.
What
is
the
issue
38?
It
kind
of
goes
over
the
the
main,
the
main
idea,
which
is
you,
can
use
globs
all
you
want
in
in
your
scripts,
but
they
are.
They
are
interpreted
by
your
shell
if
you're
on
Linux
and
if
you're
on
Windows
they're
not,
and
so
you
would
have
to
pass
those
globs
in
as
a
string,
and
so
that
may
mean
single
quotes
or
double
quotes
depending
on
what
what
what
platform
you're
on
and
so
you'd
give
it
a
string.
A
And
so
all
the
command-line
tools
that
want
to
support
globs
have
to
pull
in
a
globbing
module
and
and
parse
the
the
damn
thing
themselves,
and
so
it's
pretty
standard
across
you
most
most
I
will
say
widely
used.
Man
line
tools
that
operate
on
multiple
files,
we'll
we'll
try
to
support
clops
and
to
do
that
they
have
to
pull
in
a
a
module
and
so
give
the
kind
of
the
ubiquity
of
trying
to
parse
these
things.
A
Wouldn't
it
be
nice
if,
if
Corps
had
facilities
to
to
help
people
do
this,
just
because
it's
it's
essentially
it's
necessary
for
a
good
cross-platform
experience
in
a
lot
of
cases
in
CLI
tools,
and
so
there
were
quite
a
few
people
that
try
and
be
in
on
that
post
and
agreed
and
said
yeah.
This
is
you
know.
This
makes
a
lot
of
sense
because
it's
it's
kind
of
a
it
promotes
cross,
cross
platform,
use
of
node,
there's
statistics
somewhere
and
I.
A
A
Popular
place
to
do
that,
but
you're
you're,
doing
your
development
on
Windows
and
so
node
could
do
a
whole
lot
more
with
your
stuff.
Could
do
a
whole
lot
more
to
make
it
a
better
experience,
but
you
know
we
can
do
what
we
can
and
I
think
I
think
adding
gob
support
is.
This
is
probably
a
good
idea
and
so
I'd
like
to
take
the
temperature.
What
what
do
people
think
and
I
know
I
know
many
of
you
have
kind
of
you
know
commented
on
that
if
you,
but
let's
let's
talk.
B
Yeah,
you
know
I
definitely
commented
on
the
issue
and
said
I'm
in
favor
of
it
I
think
it
makes
sense
to
have
something
kind
of
basic
like
that
in
node,
I.
Think
there's
like
sort
of
two
issues
that
you
brought
up,
though
one
is
like
I
wrote,
a
node
script
and
I
want
to
be
able
to
parse
a
glob
like
string
in
there,
but
then
the
scripts
in
like
package
Jason,
is
sort
of
another
issue.
B
I
think
right
like
that,
would
that's
more
of
like
a
like
NPM
or
yarn,
or
something
that's
like
running
those
scripts.
I
agree
would
be
nice
if
globs
work
consistently
in
there
across
Windows
and
Mac
and
Linux,
but
I
think
that
is
maybe
a
trickier
issue,
but
yeah
just
the
idea
of
having
glob
in
court
as
seem
appealing.
A
It's
kind
of
an
avoidable
I
mean
maybe
maybe
it's
not
unavoidable,
but
it
would
take
work
on
the
CLI
side,
which
is
probably
you
know,
the
NPM
CLI,
which
probably
breaking
changes
and
terrible
things,
and
and
probably
trying
to
do
too
much
and
I'm
going
to
doubt
that
Isaac
wouldn't
want
to
do
anything
like
that.
So
it's.
B
A
B
B
Or
something
which
is
kind
of
unfortunate,
because
those
are
some
things
that
would
be
really
nice
to
be
able
to
fix
yeah
I
think,
regardless
of
that,
even
just
having
glob
support
in
node.
So
if
I
was
writing,
say
a
command-line
tool
or
something
and
I
wanted
to
implement
glob
not
having
to
pull
in
a
package
would
be
nice.
A
C
C
C
C
C
C
A
E
D
C
C
C
F
B
F
F
So
I
guess
kind
of
where
I
am
with
this
proposal.
I
kind
of
think.
It
would
be
good
to
look
at
implementing
the
paths
functionality
first
and
I'm
saying
the
path
functionality,
because
I'm
assuming
that
many
match
or
whichever
variant
would
end
up
being
under
the
nodejs
path
module
rather
than
FS
module,
write.
E
F
C
A
B
An
approach
like
that
might
be
nice
too,
because
you
could
potentially
like
pass
your
own,
like
if
you
weren't,
happy
with
the
match,
or
that
we
had
built
in
just
like
with
some
of
the
other
FS
operations,
where
you
can
pass
your
own
like
functions
into
it.
You
could
pass
like
use
this
matter
instead
or
something.
B
B
C
A
So
sorry,
as
far
as
mini
match,
I
mean
as
far
as
I
know,
the
most
popular
one
that
does
this
type
of
thing
is
the
mini
match
module
BAM,
and
you
said
something
about.
There
are
different
implementations
that
aren't
quite
the
same.
If
there
are
there
concerns
about,
you
know
which
implementation
would
want
to
pick
up
or
yeah.
B
C
C
D
C
A
F
Yes,
so,
let's
say,
for
example,
there's
something
the
implementation
of
path
match
might
have
differences
from
many
match,
and
micro
match,
I'd
like
to
try
and
kind
of
coordinate
with
the
authors
of
mini
match
in
micro
match
and
see,
if
there's
a
way
that
we
can
get
them
to
agree
to
basically
do
a
assembler
major
of
their
own
module
so
that
the
latest
version
of
their
module
follows
the
same
functionality
as
nodejs
and
kind
of
try
and
eliminate
the
differences.
And
that
way
there.
A
A
Actually,
like
stuffing
the
polyfill
in
there,
because
he
it
presents
like
another
potential
way
for
things
to
break
for
a
user,
and
so
if
he
doesn't
polyfill,
he
has
complete
control
over
what
the
API
looks
like.
If
node
breaks
something,
then
then
the
users
of
his
module
will
will
feel
the
pain
you
cannot
I
mean
and
so
I
don't
know.
Maybe
maybe
polyfill
I
mean
I.
Think
that's
a
reasonable
concern.
A
C
B
D
B
F
A
A
A
F
Yeah,
but
as
far
as
as
far
as
like,
for
example,
like
what
actually
works
in
a
package.json
for
the
scripts,
I
think
the
reasonable
thing
would
be
to
assume
that
that's
not
enabled
you
know.
For
any
author,
that's
trying
to
put
out
a
something
that
you
have
scripts
that
are
compatible
with
whatever
systems
running.
C
A
I'll
see
if
I
can
just
write
up
a
comparison
of
these
two
and
let's
say
a
comparison
of
of
micro
match
mini
match
in
just
whatever
Bosch
supports
and
I'm
go
with
Bosch,
because
it's
it's
like
the
default
shell
so
and
then
I'll
go
ahead
and
I'll
dump
that
comparison
into
the
issue
and
I
can
also
write
about.
Let's,
let's
try
to
start
with
some
any
match
stuff,
instead
of
just
going
right
to
two
FS
Cobb
or
what
have
you
as
a
first
step
and
see
what
the
with
the
other
commenters
think
of
that
are.
B
A
F
A
A
A
C
C
C
And
plucks
out
the
source
map,
if
it
looks
like
it's
a
file,
it
loads,
the
file
in
for
you.
If
it
looks
like
it's
a
basically
four
encoded
buffer,
which
is
really
the
codes
it
for
you
and
just
places
it
in
a
cache
of
source
maps
that
are
resolved
file
names
to
these
source
files
are
source
map
files.
Then,
for
this
very
simple,
MVP
release,
it
just
dumps
those
two
discs
for
you
at
exit
time,
if
you're
running
with
test
originally.
B
C
C
When
we
compile
the
code,
we
could
pass
those
source
map
objects
into
the
compilation
step
and
then,
when
a
v8
has
an
exception,
it
could
take
those
same
source
Maps,
we're
now
already
loading
out
of
the
source
files
and
give
you
a
proper
stack
trace
instead
of
a
incretin
stack
trace
when
your
program
is
I'm
kind
of
seeing
it.
It's
like.
C
C
C
C
D
C
B
C
C
C
D
C
C
D
D
B
B
D
C
Which
we
do
with
I
see
why
you
to
be
required
on
extensions
to
see
the
code
as
it's.
C
C
A
C
C
I
think
both
would
be
reasonable,
like
I
think,
if
you're,
if
you're
indicator,
that
you
want
to
dump
the
source
information
to
disk,
it
seems
like
a
reasonable
thing
to
also
get
on
disk
is
that
you
can
combine
the
two
and
have
a
proper
coverage
report,
but
if
you
don't
want
to
dump
stuff
to
this,
I
think
would
make
sense.
First,
you
just
expose
it
as
a
runtime
api
for
for
other
tooling
authors,
basically
and
say
like
it's
there,
a
source
map
for
Fuji
s.
A
D
C
F
C
C
A
B
A
A
I
certainly
am
in
no
position
to
try
to
try
to
lead
another
effort
to
get
say
a
copy
tree
or
something
in
there,
but
is
this
still
something
that
that
people-
and
this
is
kind
of
Raphael's
thing
and
I-
haven't
seen
him
in
a
while,
but
is
this
so
still
something
people
are
interested
in?
Do
we
still
think
it's
a
good
idea
to
try
to
add
a
couple
more
of
these.
These
types
of
commands.
C
A
A
B
B
A
Okay
and
then
argument
parsing,
so
I
haven't
I,
haven't
had
a
chance
to
look
at
argument.
Parsing
I
mean
it's
still.
Something
I
would
really
like
to
see
done.
It's
just
kind
of
it's.
It's
further
down
the
list
on
things
to
do.
You
know
I.
Think
Rebecca
who
was
interested
in
this
is
busy
with
a
new
job
or
something-
and
she
may
not,
you
know,
have
have
much
more
interest
in
working
on
it.
A
But
I
mean
if
there's
I
don't
know.
If
anybody
here
is
interested
in
working
on
that
it
sounds
like
we
all
have
our
own
things,
that
we
need
to
get
done,
I
foresee
more
time
to
to
work
on
some
implementation
of
stuff.
You
know,
or
after
after
maybe
next
month,
the
after
you
know
October,
essentially
I'm
kind
of
I'm
pretty
busy.
Until
then,.
C
Argument,
pricing
for
Google,
because
our
samples
it
would
make
our
samples
we've
gathered,
is
very
great
for
dropping
yards
all
over
the
place
because
we
don't
want
people
after
install
of
dependency
to
run
our
samples.
So
if
we
have
just
a
nice
little
built-in
argument
parser,
it
would
just
make
those
way
cleaner,
yeah.
B
I
think
I
said,
like
a
month
ago,
like
I
had
to
I,
was
working
on
a
tool
for
work
and
I
needed
to
add
a
single
command
line,
option
to
hit,
and
so
I
mean
I
ended
up
just
like
manually
parsing,
like
processed
on
RB,
but
yeah.
It
would
have
been
nice
if
there
was.
Is
that
one
argument
then
grows
into
two,
and
then
you
have
our.
You
have
a
second
one
that
takes
like
a
value
and
yeah
I,
really.
C
A
B
C
C
C
A
A
A
A
Unescape,
maybe
do
not
use
unescape,
okay,
all
right,
so
this
is
already
been
time.
Do
not
use
unescape
to
decode
your
eyes,
use
decode
uri
instead,
all
right!
So
that's
what
I
was
gonna
say:
I
was
like
well,
something
like
that
might
might
make
sense
as
like
a
a
module.
If
it's
like
a
thing
that
is,
you
know,
omnipresent,
but
not
not
actually
standardized,
but
then
again
it's
already
been
taken
care
of.
So
what
do
I
know?
I,
never
even
knew
about
the
code
URI
anyway.
A
C
Great
creating
unit
tests,
if
you
need
like
test
data
generated,
you
can
use
it,
certainly
for
like
FedEx
uses
it
for
packages
like
I've,
got
a
great
post
on
this.
Actually
we
were
doing
a
literature
review
as
part
of
the
pistil
reading
a
literature
review
as
part
of
the
proposal
we've
been
playing
together
and
one
of
the
people
who
wrote
here
is
like
happen
to
notice.
It
was
under
FedEx
package.
While
they
were
doing
the
literature
review
kind
of
funny.
C
C
B
A
A
B
Sorry,
just
before
we
go
I
was
gonna
say
we
kind
of
mentor
brought
up
a
few
people
like
Raphael
and
Rebecca
Roy,
who
would
have
been
at
the
meetings
in
the
past
and
we
kind
of
haven't
had
a
meeting
for
a
little
while.
So
maybe
we
should
make
an
effort
to
kind
of
let
them
know
we
have
a
new
meeting
time
and
sort
of
invite
them
back.
Yeah.