►
From YouTube: Open RFC Meeting - Wednesday, May 18th 2022
Description
In our ongoing efforts to better listen to and collaborate with the community, we run an Open RFC call that helps to move conversations and initiatives forward. The focus should be on existing issues/PRs in this repository but can also touch on community/ecosystem-wide subjects.
A
A
I've
also
copied
and
pasted
the
meaning
note
stock,
which
is
also
here
in
chat,
feel
free
to
add
yourselves
as
attendees
and
we'll
be
trying
to
take
notes,
still
gently
as
possible
apologize
for
the
spam
and
just
quickly
want
to
note.
B
A
A
Also,
once
a
quickly
did
the
floor
for
announcements,
we
have
a
v9
road
map
which
we're
updating
regularly
here
and
have
been
doing
some
planning
trying
to
get
out
the
last
few
features
and
bugs
in
v8
before
we
cut
over
to
sort
of
cutting
b9
releases
and
want
to
get
this
on
folks
radar
feel
free
to
give
us
feedback.
A
Let
us
know
about
any
breaking
changes
that
might
affect
you,
I'm
going
to
try
to
do
a
better
job
here,
as
we
move
from
v8
to
b9
to
make
sure
that
folks
know
about.
What's
changing
so
feel,
free
to
comment
in
that
issue
and
yeah
we'll
we
can
try
to
address
any
concerns
you
have
in
the
comments
there.
Also.
A
I
wanted
to
note
that
our
team
is
headed
to
austin
in
early
june
to
meet
up
with
community
and
also
to
be
a
part
of
the
collab
summit
and
that's
for
openjs
world.
I
believe
it's
june
6th
to
the
10th
is
the
week
that
folks
will
be
there.
I
believe
there's
some
discount
codes
etc
going
around
if
you
haven't,
got
a
ticket
yet,
and
we
can
try
to
find
the
right
folks
to
get
you
hooked
up.
A
If
you
haven't
already
decided
to
head
out
there,
there's
also
virtual
options
for
folks
that
don't
want
to
meet
up
in
person.
I
think
this
is
going
to
be
a
great
conference,
so
hopefully
folks
will
attend
and
we're
hoping
to
to
meet
y'all
in
person
after
a
few
years
away,
so
want
to
give
a
floor
for
anybody
else.
If
anybody
will
set
any
announcements.
A
If
not,
we
can
just
dive
into
the
agenda.
We
had
issue
number
581,
rrfc,
add
messages
or
opt
out
support
for
non-registered
fireball
urls.
Oh,
and
I
think
this
is
one
of
the
issues
you
made.
C
Yeah
the
feedback
from
last
week
was
that
it
was
good
to
proceed
to
an
rfc
I'll,
be
able
to
do
that
for
next
by
the
time
of
next
week's
meeting.
C
A
That's
right,
yeah!
That's
fine!
Awesome!
Thank
you
very
much.
Moving
on
to
our
fc
566
command
specific
configuration.
This
is
my
rfc.
I
haven't
seen
any
updates
from
anybody
else.
I
know
that
this
is
basically
blocked
and
I'll
actually
add
the
comment
here.
Action
items
needs
commands.
A
Yeah,
just
I,
I
think
that
we've
referenced
this
in
a
couple
calls,
but
I
think
a
blocker
to
this
is
clearly
documenting
the
commands
that
call
other
commands
and
potentially
would
share
config.
So
I
think
that
that
audit
needs
to
be
done.
I
haven't
had
time
to
look
into
it,
but
I'm
just
going
to
document
that
here
in
the
pr
and
then
maybe
remove
the
agenda
label,
so
it
doesn't
pop
up
next
week.
If
there's
no
updates,
it's
command
on
it.
A
Moving
on
to
the
next
item
that
we
had
so
this
is
rfc
564.
the
dependency
selector
circuit
text
and
npm
query
command.
I
think
the
update
here
is
that
myself
and
roy
went
over
the
last
two
weeks.
I
think
had
two
sessions,
maybe
two
or
three
sessions,
cleaning
up
the
rfc
itself.
A
I
think
we
addressed
a
number
of
the
concerns,
including
the
biggest
one
which
what
we
were
concerned
about
was
attributes
query
dynamically
created,
I
think
pseudo
selectors,
which
essentially
would
have
eaten
up
the
name
space
like
yeah
super
selector,
name
space.
A
That
was
originally
a
strategy
that
I
thought
could
be
used
to
try
to
dynamically
iterate
over
objects
and
arrays,
which
is
not
something
that
like
html
or
css
sorry
has
that
capability
to
do
the
way
that
we
were
sort
of
able
to
navigate
that,
and
if
you
look
at
the
rc
as
it
is
written
today,
and
I
can
potentially
even
share
my
screen
here
so
long
as
I
get
rid
of
everything
else
that
might
be
sensitive.
A
Okay
and
apologize
for
the
resolution
and
the
size
of
it,
so
the
changes
here
that
were
made
that
are
probably
the
most
significant
are
the
introduction
of
the
attributes.
Where
is
it
sorry
the
attribute
selector?
I
think
this
is
the
old
rendered
version
actually
close.
This
actually
update
this
rendered
version.
A
So
yes,
the
biggest
change
here
is
the
attribute
selector,
our
query
selector.
So
it's
essentially
colon
adder,
which
would
be
utilized
in
the
same
manner
that
had
originally
been
proposed
to
iteratively.
A
Walk
objects
and
arrays
using
sort
of
that
original
idea
of
the
key
mapping
to
sort
of
these
nested
query
selectors.
A
A
D
D
A
I
mean
I'm
okay
with
us,
changing
that
I'm
totally
easy
with
what
gets
used
there
yeah.
It
is
a
special
case,
like
essentially
iterating
over
arrays.
How
can
we
reuse
the
attribute
selector
syntax
without
introducing
something
net
new
and
so,
like?
I
thought
using
yeah
like
I
guess,
a
special
key
or
something
like
that
would
would
make
sense.
You
could
also
have.
D
D
D
Is
a
yeah,
that's
and
you
could
have
a
key
of
you
know
a
string
that
was
empty
square
brackets,
and
so
that's
a
fair
point.
So
maybe
like
I
don't
I
don't
know
what
the
solution
is,
but
it
would
be
really
ideal
if
somebody
was
able
to
come
up
with
something
that
couldn't
possibly
be
an
object
key.
But
I
think
the
other
question
is
what
happens
in
an
object
where
the
key
has
a
space
in
it?
Are
you
forced
to
put
it
in
quotes
yeah?
You
can
have
literally.
D
D
Like
like,
you
could
do
square
brackets
with
a
space
between
them
automatically,
it
has
to
be
in
quotes
if
it's
an
object
key
and
not.
If
not-
and
I
don't
know
this
is
a
bike
shed,
but
it's
it
seems
like
a
useful
property
to
result
to
end
up
with.
If
someone
can
come
up
with
a
good
idea,
yeah.
A
Okay,
I'll
leave
that,
as
kind
of
like
an
unknown
question
or
a
question
to
consider
other
than
that,
we've,
like,
I
think,
really
landed
on
in
a
good
place.
We've
kept
one
of
the
things
that
I
know
that
you
also
have
some
pushback
on.
Jordan
was
like
the
classes.
A
We've
kept
these,
but
we've
changed
like
the
description
to
try
to
make
it
more
reasonable.
Why
we
are
making
this
distinction
for
this
set.
D
To
be
clear,
I
think
the
functionality
is
excellent
and
necessary
and
I
think,
as
you've
described
it
and
like
those
specific
names
like
after
the
dots
in
the
bullet
list.
Those
are
great,
I'm
literally
just
allergic
to
the
word
class
here,
and
I
think
because
it
means
so
many
different
things.
It
was
a
mistake
for
javascript
to
pick
that
name
for
its
its
inheritance
syntax.
It
was
a
mistake
for
the
web
to
pick
it
for
css
like
it's
just
such
a
massively
overloaded
word
and
nobody's
who's.
D
Familiar
with
css
is
going
to
be
confused.
I
think
if
they
see
a
selector
starting
with
a
dot
and
then
there's
no,
the
word
class
does
not
appear
anywhere
in
npm
stocks
to
describe
it.
Everyone's
just
going
to
understand-
and
I
think
like
like
in
the
worst
case-
and
I
think
that
in
the
best
case,
we'll
have
spared
people
a
lot
of
confusion.
So,
like
I,
I
feel
a
lot
more
strongly
about
the
naming
here
than
I
do
about
the
item
that,
like
the
word
class,
should
be
like
verboten.
D
You
are
reasonably
cribbing,
css
syntax
here
right,
but
it
does
not
map,
no
matter
what
it
will
never
map
completely
across,
and
so
I
don't
think
that
that's
an
a
high
priority
and
that
should
be
a
high
priority
to
have
that
mapping,
because
it
just
has
to
be
something
that
feels
reasonably
intuitive
if
you're
familiar
with
one
and
you
look
at
the
other-
and
I
don't
think
the
word
class
helps
that
case.
If
anything,
I
think
it
hurts
it.
A
A
D
B
D
B
D
But
like
something
can't
be
both
pro
well,
I
guess
it
could
be
in
multiples,
yeah,
okay,
so
fine,
but
at
any
rate
it's
like
what
kind
of
dependency
is
it
right?
That's
what
that,
and
it
could
be
one
or
more
of
these
right.
You
can
have
both
prod
and
death
right.
So
on
so
they're
different
classifications,
I
mean
yes.
D
Yeah
yeah,
I
mean
I'm
not
saying
I
have
the
answer
off
top
my
head.
I'm
just
saying
that
I
think
almost
anything
is
better
than
the
word
class
and
like
even
type,
would
be
better
than
class,
although
type
is
a
little
generic
too,
but
yeah.
I
just
really
think
it's
important
to
pick
a
non-confusing
term
here.
B
D
Mean
think
of
how
often
on
this
very
call,
we've
all
stumbled
over
the
terminology
for
workspaces
that
you
know
I
claimed
was
was
mistakenly
chosen
like
this
stuff,
really
matters
it
matters,
it
affects
how
we
talk
about
it
forever.
It's
not
just
in
the
docs,
and
the
word
class
I
think,
is
just
going
to
cause
us
a
lot
of
problems,
and
this
is
our
this
is
our.
This
is
like
before
landing
for
implementing
this,
like.
That
is
our
only
opportunity
to
fix
this
forever.
A
I
don't,
I
would
totally
be
okay
with
forever
now
understanding
that
a
class
of
dependency
is
one
of
these
six
types
and
if
they
expand,
we
have
a
future
new
class
that
gets
introduced
to
this.
This
set.
Our
understanding
of
what
that
means
to
us
is
something
that
would
be
defined
in
your
package.json
as
a
new
like
set
of
dependencies.
D
D
A
I'm
concerned
about
types
and
the
reason
why
I
would
push
back
on
types
is
that
we
get
into
types
that
has
the
same
problem
in
that
it
then
we've
also
used
that
to
say
git
depths
are
a
type
of
dependency
or
link.
Depths
are
a
type
of
dependency,
and
so
like
that,
expands
out.
The
types
like
like
beyond
this
set
for
sure
like
it's
like
those
are
a
different
set
or
different
classification.
D
D
D
Like
I
said,
I,
I
don't
think
it's
important
what
it
is.
I
think
it's
important
what
it's
not
so.
If
types
isn't
ideal,
then
great,
not
types,
but
like
groups
works.
You
know,
there's
lots
of
other
choices
we
can
toss
around,
because
it's
because
these
are
like
a
subset
of
kind
of
a
larger.
D
A
I'll
add
a
note
about
that's
explicitly
as
well,
and
then
I
think
other
than
that.
I
believe
that
roy
has
already
begun
to
do
some
initial
work
on
this.
So,
like
that's
one
thing
to
note
is
yeah:
it's
we're
excited
to
potentially
have
something
to
share
and
show
next.
You
know
when
we
see
folks,
potentially
in
person
at
the
conference
and
yeah,
I'm
rory,
I'm
not
sure
if
you
want
to
share
anything
on
that,
but
yeah
just
other
than
development
has
kicked
off
this
week.
C
A
Cool
and
obviously
like
some
of
these
concerns
that
are
like
left
over
can
pretty
easily
be
quickly
fixed
and
changed,
like
I
don't
think
you're
at
the
point
where
you're
implementing
the
sort
of
those
iterative
operations
or
even
the
classes
that
are
the
groups
whatever
you
want
to
call
them
cool,
moving
on
then
to
rfc
548s.
A
I
know
we
had
discussion
last
week,
potentially
I'm
not
sure
if
enough
was
on
that
call,
but
yeah.
I
think
this
comes
down
to
us
just
implementing
this,
like
somehow
I
forget
what
was
it
like?
What
were
the
notes
from
last
week.
B
Yeah,
I
think,
last
week,
jordan
highlighted
the
the
fact
that
npm
has
the
dependency
graph
information
and
could
do
a
better
job
of
just
properly
using
that
order.
Instead
of
just
not
ordering
anything
or.
A
Right
so
then,
what's
what's
the
action
on
our
end,
just
like
adding
an
rc
for
this
or
we
get
backlogged,
I
mean.
D
A
D
B
B
D
Is
detect
determined
then
you
can
do
both
of
those
features
easily,
and
I
really
want
the
topological
order.
One.
Some
other
people
really
want
the
include
transitive
dependencies,
one
those
can
be
discussed
independently
and
shipped
in
either
order.
I
I
just
am
advocating
that
we
build
the
core
part
of
the
core
dependency
detection
part
so
that
both
it
can
lead
to
both
features,
because
I
think
both
have
valid
use
cases.
B
D
D
Right
right
and
then
with
with
both
of
these
features,
only
child
workspaces
are
candidates
right.
It's
never
digging
into
node
modules.
It's.
E
D
Circling
the
cycle:
yeah
one
option
is
that
it
explodes
and
says
this
is
a
circular
dependency.
You
use
this
flag
to.
You
know
run
in
the
other
in
the
like
current.
Whatever
the
status
quo
is
called
right,
like
first
the
opposite
of
topological
order,
or
we
just
run
them
together
in
parallel
or
I
don't
know
I
mean
I
never
have
circular
dependencies,
so
lots
of
people
do
like
and
the
issues
we
get
about.
D
The
open
question
is
how
many
of
the
people
that
have
circular
dependencies
should
have
them
and
want
them
versus
have
them
by
accident.
That's
a
that's
a
more
editorial
question
which
maybe
npm
doesn't
want
to
get
involved
in,
but
my
personal
intuition
is
that
virtually
nobody
actually
should
have
them
and
so
exploding
in
this
with
it.
You
know
exploding
in
this
case
and
with
an
escape
hatch
of
like
just
run
them
all
in
parallel
seems
fine
to
me,
but
we
should
get
input
from
those
who
think
that
they
want
circular
dependencies.
A
Okay,
I
think
I've
captured
I'm
gonna,
create
two
separate
issues
for
this
and
backlog
them.
One
is
for,
it
seems,
like
include
transitive,
is
fine
for
a
flag
between
bike
shed,
the
actual
name
and
the
other
is
the
topological
ordering
which
right
now
is
ordered
by
like
for,
like,
I
don't
know
how
we're
actually
ordering
it,
but
I
think
it's
directly
provided
it's
in
the
order.
You've
passed
those
the
workspaces
but
providing
this
flag
would
essentially
then
reorder
the
workspaces.
A
So
I
see
the
I
see
the
benefit
in
both
these
things.
I
think
so
yeah
we
can
backlog
them
cool
and
I
think
actually,
I
can
probably
just
get
rid
of
the
agenda.
We
talked
about
this
quite
a
bit
and
once
I
backlog
these
two
items,
then
it's
pretty
much.
A
We've
accepted
the
fact
that
we
want
this
in
in
core,
so
yeah
cool,
so
I've
just
removed
the
agenda
label
moving
on
to
rfc
546..
This
is
cleaning
up
file
ownership
roy.
I
know
you
brought
this
up
just
as
a
general
thing
to
keep
in
mind,
there's
been
some
discussion
and
even
some
diagrams
made
by
nelf
he's
even
messaged
here.
E
Just
so
it's
on
record
here
I
did
find
the
comment
I
was
thinking
of
about
the
last
time.
We
were
talking
about
running
things
in
topological
order.
Isaac
explain
why
that
may
not
be
able
to
work.
It
may
apply
here.
It
may
not
because
we're
talking
about
limiting
it
to
workspaces,
but
I
added
the
comment
to
the
rfc
too.
A
I'm
not
sure
if
there
was
any
update
here.
Oh
no,
as
of
8
6,
we
validate
the
shape
of
the
advancing
tree,
make
sure
it's
valid.
I
know
he
was
doing
work
in
that
area
and
we'll
throw
we
see
entirely.
A
Something
was
removed
or
entirely
wrong
version.
This
should
lay
the
groundwork
for
further
verifications
like
resolve
and
integrity
values.
So
I
think
we
can
take
the
label
agenda
label
off
this.
We're
essentially
doing
the
work
to
fix
this
up.
We
could
potentially
even
close
this
if
fitzy
is
okay
with
it,
based
on
the
fact
that
we're
continuing
to
add
more
checks,
alright,
so
you're
unused.
Did
you
have
a
comment.
E
A
A
Okay,
moving
on
to
item
number
eight,
this
is
rc
519.
This
is
mine
for
package
distributions.
Unfortunately,
this
has
sort
of
stalled.
I
have
not
had
any
time
to
really
look
at
this
spend
time
trying
to
clean
up
query,
which
is
the
last
lost
rc.
I've
really
put
any
effort
into
I'm
not
sure
if
I
can
get
somebody
else
interested
in
this
I
know.
A
Actually
there
was
an
update
from
roy
where
he
had
reached
that
to
mile
from
yarn,
and
there
was
some
interest
by
mile
to
actually
collaborate
on
this
rfc
to
you
know,
align
with
yarn
to
ensure
that
we
built
something
that
potentially
both
package
managers
are
aware
of
and
implement
sort
of
similar
functionality
around.
So
I
know
that
they
have
a
precursor
rfc
for
variance
package,
variance
which
is
very
similar
but
the
shape
of
our
config
and
is
a
little
bit
different.
A
B
Yeah,
no,
that's
pretty
much
it.
I
think
we,
if
you
can
expand
on
the
rfc
for
more
optional
dependency
types,
like
that's,
probably
a
good
thing
to
do
meanwhile,
but
it
might
be
better
to
actually
wait
see
if
mile
can
come
back
to
us
right,
yeah,
the
the
yarn
version
of
it.
The
iron
variants
has
been
open
for
quite
a
while
too,
so
maybe
it's
for
the
best
that
we
all
wait
a
little
bit
more
and
we
can
collaborate
and
come
up
with
a
single,
unified
solution.
A
So
no
real
update
there
moving
on
to
breaking
change
for
bin.
This
is
479..
This
is
command,
should
not
return
non-existent
pass.
I
believe
we
have
this
q
dot.
Right
is
the
reason
this
got
added
to
the
agenda.
Still,
I
think
we
were
wondering
whether
or
not
I
don't
know
what
the
outcome
was,
but
we
were
wondering
what
the
value
was
of
removing
this.
E
E
D
Thing
I
can
think
of
is
that,
if
someone's
using
it
in
a
subfolder
to
figure
out
where
the
route
is
because
the
presence
of
a
package
json
isn't
sufficient
for
that,
and
I
don't.
I
don't
actually
know,
though,
how
well
npm
bin
works
for
that
like
if
I
have,
if
I'm
in
a
nested
folder
with
a
package
json
that
just
has
a
type
common,
js
or
type
module
and
nothing
else,
and
I
type
npm
bin
in
that
folder.
Do
I
get
just
dot?
D
Well,
if
I
type
npm
install,
I
know
it
goes
up
and
finds
the
dependencies
and
installs
them
so
without
a
prefix
or
writing
kind.
I'm
just
trying
to
grasp
it
straws
to
see
if
there's
any
use
cases
for
npm
bin.
That
aren't
do
this
dangerous
thing
and
put
it
in
your
path
that
there's
been
articles
warning
you
not
to
do
for
a
decade
like
if,
because
if
there
are,
if
there
are
no
other
use
cases,
then
we
should
burn
it
with
fire
right,
kill
it
with
fire,
but
like
the.
E
A
Maybe
I
changed
this
then,
where
is
this
and
let's.
B
Last
time
we
talked
about
it,
we
were
talking
about
just
like
people,
usually
using
npx
instead-
and
I
know
like
neil
mentioned
with-
maybe
we
should
just
add
support
to
the
location
flag
for
to
npx,
so
that
folks
can
point
to
different
locations
to
be
reading
the
packets
from.
A
So
what
you're
saying
is
sorry:
npm
config
supports
a
location
flight
today
for
folks
that
don't
know
so,
we
can
specify
where
you
look
and
there's
a
set
of.
I
think
four
locations
that
are
known
to
npm
projects,
user
global
and
I
think
or
something
else.
A
And
then
so,
what
you're
saying
is
yeah
now
npm
exec
would
get
the
would
now
support
that
location,
flag
and
understanding
as
well
to
be
able
to
execute
within
those
locations.
Is
that
yeah,
so
that
would
override
any
other
path
of
prefix
that
I
like
was
defined
like
location,
would
essentially
trump.
That
is
that
what
you're
saying.
A
A
Okay,
let
me
add
that
quickly
as
well.
A
A
Okay
done
and
done,
updated
and
closed
moving
on
to
our
last
item
with
10
minutes
left
of
the
call
this
is
issue
575,
future
requests
run
scripts
should
short
circuit
on
scriptera.
I
think
this
is
similar
to
what
we're
just
talking
about
before
right
or
no.
This
is
a
little
bit
different.
This
would
be
providing.
A
It
is,
I
think,
orthogonal
enough
right
that
if
I
set,
if
I
ever
had
false
topological
order
right
ordering
and
and
only
I
had
and
I
had
specified-
let's
say
a
few
different
workspaces
that
I
wanted
to
execute
some
script
in,
and
I
knew
that
like
as
soon
as
it
fails
the
first
time
I
don't
want
to
continue
execution,
I
don't
want
to
voice
cycles
or
whatever.
Then
this
might
be
the
flag
that
I
would
also
define
to
say.
D
Yeah
this
seems
like
a
so
there's.
Whatever
the
I
don't
know
what
the
current
mode
is
called
the
the
one
the
topological
is
not,
I
don't
know
parallel
so
like
in
parallel
mode.
D
You
should
certainly
might
want
to
opt
in
to
short
circuit,
and
you
could
use
a
flag
to
do
that
in
topological
mode
you'd
always
have
short
circuiting
there's.
No,
I
I
can't
conceive
of
a
reason.
You'd
want
to
opt
out
of
short
circuiting
there
and
so
like.
D
It
seems
like
a
flag
like
something
that,
if
that
will
be
much
easier
to
add
when
there's
command
specific
syntax,
but
like
only
allowing
like
all
the
places
that
the
topological
flag
is
allowed.
So
maybe
actually
maybe
that's
another
way
to
say
it.
It's
a
third
mode,
there's
parallel,
there's
parallel
with
short
circuiting
and
then
there's
topological,
and
that
way
it's
the
same
flag.
It
affects
the
same
set
of
commands
which
isn't
just
run
script.
D
It's
also
start
and
test
and
publish
and
pack
and
all
these
others
and
and
it's
sort
of
mutually
exclusive,
with.
A
I
don't
have
any
immediate
thoughts,
but
I
do
think
that
there's
value
in
providing
enough
life
where
we
change
the
behavior
to
fail
early,
and
I
knew
that
when
we
were
implementing
this
and
we
decided
not,
we
decided
not
to
fail
early
or
fail
that
you
know
some
people
would
eventually
want
that.
Behavior.
A
Yeah
I
mean
that's
like
like
more
clearly
defined
like
if
you
want
to
define
some
other
type
of
ordering
right
in
some
other
kind
of
like
manner,
like
the
reason
why
I
reference
wire
it
in
my
comment-
and
we
talked
about
the
time
bit
last
week-
I
think,
was
that
they
are
tacking
on
to
essentially
run
scripts
like
that
that
understanding
of
like
which,
which
things
should
run
not
just
it's
like
build
pipelines
right
at
that
point.
A
It's
not
just
run
this
command
in
the
dependency
chain
or
in
the
chain
that
makes
the
most
sense
like
based
on
my
yeah
topological
order,
but
it's
actually
like.
Oh,
I
want
to
run
this
thing.
I'm
just
saying
this
thing
and
for
whatever
reason
one
fails
and
you
don't
have
to
have
them-
be
direct
dependencies
like
if
you're
spinning
up
your
server
before
you
want
to
run
your
app
applications
tests.
It
may
not
be
that
your
application
says
that
you
have
a
dependency
on
your
server
right
right.
A
C
Is
there
is
there
a
case
where,
like
short
circuit,
is
off
by
default,
and
then
you
add
it
when
you
want
it
and
then
topological
just
by
default,
turns
it
on
so
it's
still
like
a
boolean
but
to
jordan's
point
where
I
think
it
makes
sense
like
you're
kind
of
part
of
probably
why
you're
leveraging
topological
order.
Is
you
know
for
that
kind
of
short
circuit
feedback
loop
in
general?
So
is
it
just
a
matter
of
npm
coming
with
the
default
and
then
topological
just
turns
it
on
so
it's
off
by
default.
C
If
you're
doing
what
you
were
just
describing
darcy,
you
could
turn
it
on
and
put
it
into
kind
of
your.
You
know
chained
workflow.
C
If
you're
running
in
parallel
mode-
and
you
want
it,
you
could
also
add
it
and
then
topological
still
uses
it.
It's
just
implied
that
it's
going
to
turn
it
on
for
you
and
if
you,
for
whatever
reason,
don't
you
want
topological
without
short
circuit,
you
just
say
short
circuit
off
or
something
like
that.
So
you
can
opt.
B
A
Yeah,
we
might
want
to
look
into
what
npm
run
all
does
and
also
concurrently,
I'm
not
sure
what
the
configuration
that
both
of
those
tools
have,
but
I
know
that
they
are
tools
dedicated
to
advancing
and
expanding.
Like
your
run,
script
capabilities
right
and
they've
been
around
for
a
long
time.
The
ecosystem
is
just
them
and
like
we're
getting
into
the
territory
of
like
okay,
we
I
want
to
run
scripts
parallelly
like
concurrently
of
each
other.
I
want
to
run
them
in
sync.
A
C
Yeah
yeah
I
mean
it
could
just
be
like
fail
early
or
something
like
that.
You
have
failed
whatever
it's
just
the
behavior
I
want.
Is
this
thing
to
just
stop
as
soon
as
something
goes
wrong?
You
know
just
because
either
I
want
the
feedback
loop
or
you
know
something
really
trivial
breaks
and
my
test
then
run
for
30
minutes
in
my
ci
just
to
fail.
Because
of
that
you
know,
I
think,
that's
the
value
of
the
the
short
kind
of
circuit
behavior.
C
C
A
Yeah,
I
just
want
to
be
mindful
because
there's
so
many
there's,
probably
like
five
or
six
or
seven
different,
you
know
issues
and
conversations
that
we've
had
over
the
last
couple
years
about
roughly
around
the
same
thing,
which
is
like.
Do
we
run
things
in
parallel?
Do
we
run
them
synchronously
or
what
is
the
order
we
run
so
like
those
are
all
kind
of
like
similar
conversations
and
then
in
terms
of
like
how
we
handle
errors
that
also
plays
into
all
those
things.
A
So,
like
jordan,
said
like
maybe
just
in
this
call,
even
maybe
those
are
all
just
the
same
flag
and
they're
just
different
modes
and
like
so
you
could
even
say
you
know
like
yeah
topological
ordering
fail
early
or
something
like
it's
one,
big
long,
whatever
like
very
explicit
name
but
yeah.
A
I
think
that
what
needs
to
be
done
here
is
a
little
bit
more
research.
I
would
love
to
look
at
like
the
other
tools
and
what
the
other
tools
are
doing
and
like
a
lot
of
people
come
with
us
sort
of
just
trying
to
solve
the
one
problem.
But
like
we're
going
to
want
to
solve
this,
I
think
holistically
for
one.
D
A
D
That's
good
because
my
experience
with
concurrently
in
parallel
shell
is
that
they,
when
everything
passes
they're
great.
When
something
fails,
I
have
seen
the
overarching
command
pass,
which
made
me
rip
it
out
of
all
my
projects,
because
I
don't
want
ci
to
ever
pass
when
something
fails
so
like
how
to
juggle
that
and
manage.
That
correctly
seems
critical
if
npm
ever
runs
two
things
at
the
same
time,
but
like
I.
A
Just
something
we
do,
I
think
we
do
have
we've
been
asked
for
quite
like
we've
been
asked
for
concurrently
and
and
run
I'm
jim
reynolds.
Like
parallel
implementation.
We've
been
asked
for
that
quite
a
bit
and,
like
the
the
you
know,
those
the
popularity
of
those
tools
also
tells
us
that
the
ecosystem
obviously
wants
that
from
right,
like
from
a
tool.
So.
A
We're
doing
there
to
the
rest
of
the
ecosystem
right,
like
so
like
any
strategy
that
we
find
that
works
for
us,
including
like
how
to
you
know,
obviously,
speed
up
the
execution
and
synchronicity
or
asynchronously
like
do
operate
operations.
It's
like.
Let's,
provide
the
same
kind
of
tools
to
the
ecosystem,
so
anyways
I
apologize.
We
are
essentially
at
time
we
did
get
through
the
entire
agenda,
though,
which
is
great.
I
think
this
is
one
we'll
keep
on
the
agenda
and
we'll
like
continue
to
talk
through.
I
think
there's
probably
some
work.
A
We
want
to
consider
how
we
again
address
this
more
holistically,
but
I
think
the
idea
of
having
some
sort
of
flag
to
like
obviously
short
circuit
would
make
sense.
It's
probably
the
easiest
one
to
actually
implement
compared
to
the
other
topological
ordering
sort
of
config
but
yeah.
Thank
you.
Everybody
for
jumping
on
today
feel
free
to
continue
to
add
rc's
and
also
add
to
the
conversation
and
the
existing
ones
and
hopefully
see
you
next
week
and
want
to
quickly
note
before
we
do
go.
A
We
will
not
be
running
a
call
the
week
of
the
open.js
world,
so
we'll
be
doing
it
in
person
and
hopefully,
maybe
even
virtual,
as
well
call
for
our
cloud
summit
session,
which
may
be
the
thursday
or
the
friday
of
the
week
of
june
6th.
So
that's
two
weeks
from
now
two
or
three
weeks
from
now,
but
we
will
be
here
next
week
so
same
time
same
place,
hopefully
see
you
then,
and
yeah.
Thanks
for
jumping
on.