►
From YouTube: Open RFC Meeting - Wednesday, October 20th 2021
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
Good
good,
good
and
we're
live
on
youtube.
Welcome
everyone
to
another
mpm
openrc
call.
Today's
date
is
wednesday
october
20th
2021
we're
going
to
be
following
along
in
the
agenda
that
I
posted
here
in
chat
feel
free
to
add
yourselves
to
the
attendees
in
that
md
doc,
quick
reminder
and
acknowledgement.
These
calls
and
all
rc
calms
are
covered
under
our
code
of
contact,
and
we
just
ask
that
you
please
be
respectful
of
each
other
when
you're
speaking.
A
A
A
D
Yes,
I
I
originally
opened
the
issue
as
the
discussion
again
and
miles
directed
me
to
actually
make
file
an
issue
on
the
rfc's
repo.
So
this
is
that
essentially,
I
kind
of
have
two
questions.
There
are
a
number
of
action
items.
One
of
them
is
unassigned,
so
I
don't
know
if
that
means
like.
I
should
pick
it
up
or
if
there
were
like
waiting
for
more
sort
of
feedback
and
then
the
other
one
is.
I
think
it
was
isaac
who
suggested
this
sort
of
like
hacky
work
around.
D
That
would
get
us
get
me
to
where
I
wanted
to
be
once
the
npm
pack
list
stuff
landed,
I
think
that's
landed
in
8.1,
at
least
it's
3.0
it's
in
the
changelog,
so
I
sort
of
expected
it
to
work
and
it
does
not.
A
So,
let's
start
with
the
first
item,
there
then
funding
a
owner
of
that
second
action
item,
which
was
just
the
rc
around
redding.
A
I
don't
think
we
actually
need
need
that
as
rfc
like
do,
we
like
it's
going
to
be
a
very
small,
like
it's
almost
better
just
to
like
write
the
implementation
and
have
us
focus
on
the
pr
right
like
it's
like.
Do.
We
need
to
actually
document
that
prepare
scripts
are
going
to
run
on
on
workspaces
or
link
steps.
B
B
The
linked
multi-app
mono
repo,
it
looks
like
they're
not
listed
as
bundle
dependencies,
so
I
wouldn't
expect
that
they
that
they
would
get
included
in
the
packed
turbo,
or
maybe
I'm
not
looking
at
the
right
thing.
But
oh
it's
in
in
packages,
that's
the
one
that
has
the
dependency.
B
I
see
and
it
is
listed
as
a
bundle
dependency.
Okay,
so
I
need
to
I
mean
that
could
just
be
a
bug
with
what
pack
list
is
doing,
because.
B
If
it's
installed
and
it's
there,
it
should
be
resolving
that
that
dependency.
B
And
linking
and
including
it
in
the
tarball,
but
do
you
want
to
talk
a
little
bit
about
kind
of
what
was
the
the
more
like
the
higher
level
goal
here
or
what
is.
D
Yeah
I
mean
that
that's
kind
of
back
to
darcy's
question
or,
like
the
second
action
item
from
from
last
week's
meeting,
was
like
in
this.
In
this
example,
I'm
just
trying
to
sort
of
get
the
bundle
dependency
part
working
into
the
packed
hardware
of
app
a,
but
essentially
in
a
kind
of
more
real
world
scenario.
D
App
a
is
something
that
I
want
to
like.
I
guess.
If
I
can
generate
a
tarball
with
all
the
right
sort
of
dependencies,
I
would
untar
it
zip
it
up
and
push
it
to
aws
or
untar
and
copy
it
into
a
docker
file
and
deploy
it
somewhere
like
it
would
just
be
a
nice
way
to
specify
sort
of
local
dependencies
on
things
that
might
have
to
build.
D
So
the
part
about
running
for
peer
scripts
is
just
in
a
world
where,
for
example,
package
c
is
in
typescript,
and
I
want
to
like
compile
it
before.
I
bundle
it
and
ignore
all
the
source
code
in
the
in
the
bundled
deployed
dependency.
B
Version
right
so
that
that
is
my
expectation
of
how
it
should
be
working.
That's
that's
the
design
if
there's
some
reason
why
it's
not,
then
that
might
just
be
a
bug
we
have
to
fix,
but
automatically
running
the
prepare
script
in
those
bundle.
Dependencies
is
something
that
we
should
probably.
A
Okay,
so,
given
that
we
need
an
owner
to
actually
write
it,
I'm
not
sure
matt.
If
you
have
enough
context
to
write
something
small
like
a
small
rfc,
would
you
want
to
yeah.
D
I
I'm
happy
to
start
it
if
people
won't
sort
of
guide
me
through.
If
I
miss
anything
or
don't
understand
what
I'm
talking
about,
I'm
not
like
super
familiar
with
npm
internals,
but.
C
A
Yeah,
you
were
poking
me
the
other
day
about
this,
about
family
cool
and
also
nice,
to
see
you
yeah.
So
if
you
want
to
matt,
take
ownership
of
creating
the
rfc
then
or
at
least
like
a
first
draft
and
then
yeah
cc
mike,
so
that's
mike
mimic
his
ad
and
you
can
also
poke
us
as
well
like
just
at
isaac
myself,
and
we
can
try
to
help
like
round
out
what
that
looks.
Like
should
be
pretty
straightforward,
though
it's
like
like.
A
Essentially
you
want
that
script
to
be
that
event
to
fire
off
when
running
yeah
like
npm,
publish,
essentially
or
pat
ampump
right.
B
A
Okay,
yeah
and
we
can
asian
conditionally
be
working
on
on
on
that,
even
as
like
you're
writing
the
rfc,
like,
I
think
it's
pretty
straightforward
here.
You
go
jordan,
okay,
so
action
there
is
just
we'll
give.
E
Yeah,
so
I
definitely
want
it
written.
I
am
happy
to
be
the
one
to
write
it,
but
I
have
a
personal
struggle
with
writing
docs.
So
if,
if
I
must
write
it,
I
will
write
it.
If
anyone
else
wants
to
help
me
write
it.
That
would
be.
B
I
find
the
this
is
a
little
bit
meta,
but
maybe
worth
mentioning
in
the
context
of
this
call
a
good
way
to
kind
of
get
over
that
initial.
Just
like
breaking
the
glass
for
writing.
An
rfc
is
just
think
of
it
as
a
ever,
so
slightly
more
fancy
issue,
if
you,
if
you
just
write
up
in
the
rfc,
exactly
what
you
would
say
in
an
issue
like
this
is
this
is
the
way
it
works.
Now,
here's
why
it
needs
to
be
different.
Here's
how
I
think
it
should
be
different.
B
E
B
As
writing-
docs,
it's
like
somehow
mentally
harder.
An
issue
feels
like
complaining
and
that's
super
easy
to
do.
A
Cool,
I
think,
we're
at
the
and
then
of
this.
Unless
there
was
anything
else
you
wanted
to
follow
up
with
matt
just
feel.
A
D
Yeah,
I
know
I
think
this
is
good.
I
guess
the
I
guess.
The
other
question
is
I
mean
it's
not
even
a
really
good
question,
I'm
just
I.
I
think
I
should
file
an
issue
on
pack
with
or
npm
cli
about
how
npm
pack
is
working
with.
B
B
No,
no,
I
mean,
if
you
cd
into
the
thing
that
has
the
link
depth
so
that
the
app
a
run
npm
install
and
then
again
pack.
A
A
D
B
D
D
A
B
Ball,
a
lot
of
these
problems
go
away.
D
I
guess
that's
jordan's
workspace.
D
B
That
is,
that
is
jordan's
white
whale
for
sure.
It's
a
it's
a
good
one,
and
I
agree
with
it.
It's
just.
We
kind
of
just
haven't
had
the
time
and
energy
to
to
dig
into
it,
but
we
do
have
an
increasingly
large
list
of
things
that
will
be
solved
by
that,
so
it's
kind
of
like
yeah,
it
seems
like.
Maybe
we
should
just
do
that
and
then
npm
pack
doesn't
even
really
need
to
change
from
what
it
is
now.
I
guess
it
would
need
to
change
to
run
the
prepare
scripts,
but.
B
B
So,
within
the
context
of
that
workspace
node
we're
saying
I
depend
on
a
hyperlink
to
this
particular
folder
and
then
npm
goes
well
good
news
at
the
top
level.
I
have
a
hyperlink
to
that
specific
folder,
so
you're
good.
We
don't
need
to
put
another
hyperlink
in
there
right
then
later
hyperlink,
another
sim
link
in
there
then
later,
when
you
run
npm
pack,
it
says,
is
there
I
have
a
bundle,
dependency,
I'm
going
to
be
looking
for.
Is
there
anything
in
node
modules
and
then
it
says
no,
there
isn't.
B
So
I
guess
you
don't
want
it
in
the
package
and
it
moves
on
what
we
there
are.
There
are
a
couple
of
ways
we
could
address
that.
The
first
is
we.
We
sim
link
dependencies
into
the
workspace
node
modules
that
would
that
would
make
most
of
this
problem
evaporate.
B
The
other
thing
I
was
kind
of
thinking
that
was
like
not
a
great
idea
is,
we
could
say
well,
npm
pack
lists
could
say
hey,
I
have
a
bundle
dependency,
but
I'm
actually
running
in
a
workspace
and
this
bundle
dependency
is
found
up
a
level.
B
B
Well,
so
here's
where
this
gets
challenging,
it's
it's
during
a
simple
tree,
walk
and
it
could.
Yes,
it
could
like
do
require.resolve
and
find
that
it's
actually
up
above
the
working
folder
that
it's
in
right,
but
what
it's
creating
is
a
tarball.
So
now
what
it
has
to
do
is
say
this
file
this
this
particular
file.
That's
like
dot
dot,
slash,
node
module,
whatever
has
to
be
in
the
tarball
at
dot,
slash,
node,
module,
slash
whatever
and
all
of
its
dependencies
need
to
be
brought
up
as
well,
and
it's
it's
I
am.
E
Would
it
be
a
good
interim
solution
to
have
npm
install
for
anything
that
is
a
bundled
dependency
like,
and
maybe
this
is
only
an
issue
with
workspaces,
but
in
but
in
that
scenario
sid
link
those
in
and
then
defer
the
sim
linking
thing.
The
bigger
sim
linking
question
to
the
rfc
which,
because
that
theoretically,
would
be
compatible
with
it.
B
Now,
if
everything
gets
sim-linked
everywhere
into
the
workspace
folders
a
la
the
workspace
layout
plan,
then
that's
no
longer
an
issue
right
because
it
will
be
there
and
we'll
just
be
like
well,
I'm
bundling
a
dependency,
so
I
follow
all
siblings
right
and
that
thing
that
you're
bundling
will
have
its
dependencies
some
linked
rather
than
being
at
the
higher
level
and
so
on,
and
it's
kind
of
a
a
baby
step
towards
better
isolation.
B
Yeah,
so
I
think
that
that's
sort
of
the
the
end
goal
here
has
to
just
be
the
more
comprehensive
workspace
layout,
rfc.
E
B
A
Michael
saw
you
posting
a
couple
times:
do
you
have
anything
you
want
to
share
other
than
your
support
and
help
in
writing?
These
rfcs.
C
I
don't
know
if
it's
particularly
useful,
but
to
like
isaac's
train
of
thought,
just
to
add
words.
I
think
that
there
are
already
scenarios
so
this
this
does
specifically
only
happen
in
workspaces
and
there
are
already
scenarios
when
so
darcy-
and
I
trashed
this-
I
don't
know
whatever
that
was
a
few
weeks
ago
and
there
are
scenarios
where
it's
slightly
different
than
like
packing
stuff
and
build
and
bundle
depths.
C
But
the
idea
of
like
in
workspace
is
a
dependency
having
a
dependency
on
a
dependency
inside
of
the
workspace
and
there's
a
scenario
where,
when
you
run
npm
install
or
mpc,
I
can't
remember
what
we
were
doing,
but
it
it
does
do
the
same,
linking
of
everything
all
together
sort
of
by
accident.
It
seems
like
from
this
conversation
that
wasn't
like
by
design
it
just
sort
of
does
it
so
the
you
know
it
seems
like
there's
already
some
tiny,
tiny
baby
steps
forwards
it
wasn't
implemented.
I
can't
remember
what
it
was.
C
A
C
Yeah
and-
and
so
there
seems
like,
there
are
already
baby
steps
in
place
that
like
sort
of
allow
this,
the
notion
of
this-
the
workspace
layout
that
that
jordan's
talking
about
so
it's
just
words
to
sort
of
add
to
the
context.
I
think,
but
yes,
fully,
supporting
all
these
ideas.
D
No,
I
guess
yeah,
I
guess
sorry,
you
answered
my
question.
No,
there
is
no
bug,
it
needs
to
run
npm.
Well,
maybe
there
is
a
bug.
Maybe
there
isn't?
I
don't
know.
There's
okay.
B
A
Thanks
yeah
just
features
so,
which
hopefully
means
that
yeah
we
can
land
this
without
like
waiting
for
like
a
major
or
something.
B
There
was
a
there
was
a
bug
in
my
in
the
hacky
work
around
that
I
gave
you,
but
that's
not
technically
supported
codes
so
that
that
being
that
you
have
to
actually
run
npm
install
in
the
app
folder
itself,
which
you
really
shouldn't
have
to
do
like
the
point
of
a
workspace.
B
The
point
of
using
workspaces
is
that
you
don't
have
to
go
to
a
bunch
of
different
folders
and
run
npm
install
explicitly
in
all
of
them,
but
this
in
this
case,
you
do
in
order
to
have
that
sim
link,
local.
A
I'm
not
sure
what
exactly
the
workaround
looks
like,
but
I'm
pretty
sure
if
you
just
run
npm
exec
and
then
npm
install
dash
w
foo.
It's
going
to
do
the
same
thing
right
as
probably
what
you
were
provided.
Isaac.
A
Is
that
like
what
you're
suggesting?
Because
I
know
that
we
remove
like
the
workspace
config
from
anything
that's
run
with
exec
or
like
run?
A
A
Yeah,
exactly
what
like
michael
just,
showed,
yeah.
A
B
A
Nom
nom
nom
yeah,
I'm
not
sure
if
that's
the
same
as
as
the
hacky
implementation
that
you
shared
isaac,
that's.
B
That's
a
little
cuter.
I
I
suggest
you
just
cd
into
the
folder
and
running
npm
install
and
I.
A
So
that's
another
workaround
for
you
right
now.
Matt
is
just
like
npm
exec
and
then
npm
install
and
then
dash
w
right.
D
A
Is
well
supported,
execute
something
inside
of
my
workspace.
That's
what
that
is
supposed
to
do.
Okay,
did
anybody
have
any
other
topics
they
wanted
to
bring
up
or
or
use
this
time
at
all.
A
F
C
A
C
Over
everyone
that
didn't
speak
up
the
so
I
just
wanted
to
so.
We
briefly
touched
on
it
the
notion
of
npm
installing
a
dependency
into
a
workspace.
So
if
you
type
in
pmi,
whatever
your
workspace,
is
and
then
the
name
of
the
depth,
the
hope
is
that
it
would
just
do
what
you
would
like,
which
is
mostly
true.
C
But
if
you
try
and
npm
install
a
workspace
app
into
another
workspace,
it
doesn't
do
what
you
think
it
would
do
it
actually
just
blows
up
and
so
darcy
and
I
had
to
come
up
with
a
hack
to
make
this
work.
And
it's
a
really
great
one,
very
cute,
much
like
the
npm
exec
solution.
But
I
just
wanted
to
sort
of
say
it
into
the
ethos.
C
So
it's
in
people's
context
that
it
would
be
lovely
to
support
the
idea
of
installing
one
workspace
package
into
another
workspace
package
in
the
event
that
you
have
something
similar
to
matt's
case.
Where
I
want
one
and
another
and
like
what's
the
nomenclature
for
doing
that,
and
how
do
I
make
sure
that
npm
does
it
appropriately
and
updates
the
lock
file
and
some
links?
C
And
all
these
things
accordingly,
because
I
think
it's
really
common
now
that
we
have
this
work
like
workspace,
mono,
repo
forbidden
term
thing,
it
would
be
nice
if
they
could
like
play
nice
with
each
other
inside
the
context
of
a
workspace
dependencies.
I
mean.
A
Yeah
this
was
interesting
to
navigate
through
with
you
the
other
day
and
I'm
not
sure
if
anybody's
tried
that
yet
in
terms
of
essentially
installing
one
workspace
into
another
and
then
trying
to
see
what
the
layout
like
what
actually
happens
to
the
dependencies
yeah,
I'm
pretty
sure
cross-linking
workspaces
is
going
to
be
a
common
use
case
for
for
folks.
So,
like
the
easier
we
make
that
the
better
I'm
sure
or
the
more
like
deterministic,
it
is
like
the
better
so
yeah,
I'm
not
sure
if
anybody's
run
into
this.
A
A
It's
like
it
seemed
like
we
had
to
do
the
order
of
operations
just
just
right
to
get
it.
The
layout
proper,
which
was,
I
know,
frustrating
for
mike
and
myself,
even
as
I
was
testing
it
so
seemed
like
yeah.
It
was
blowing
up
and
I
forget
which,
which
case
it
was
like.
If
we
were,
I
don't
know
how
we
were
doing
it
if
you're
actually
maybe
cd'ed
into
the
directory.
That's
like
what
was
happening
or
if
you
had
a
package
lock
already
somehow
like
it,
was
tripping
over
itself.
C
C
I
just
I
took
a
screenshot
of
it,
but
the
context
is
missing
in
our
chat,
but
it's
basically
attempting
to
do
a
node.target.dev
equals
node.dev,
but
the
node.target
isn't
a
thing:
it's
null
or
undefined,
or
something
like
this.
So
we
can't
do
the
dot
dev
call
on
top
of
it.
B
Yep
I
I
ran
into
that
in
another
issue.
There
should
be
some
guards
around
that,
because
there
are
cases
where
you
have
a
link
and
it
is
missing
and
in
those
cases.
B
Is
is
null,
node.target
is
still
referencing
the
node
itself,
so
there
are
some
ways
that
we
could
we
could
handle
that
either
we
have
to
add,
checking
and
know
that
link.target
can
be
no
or
we
can
make
link.target
refer
back
to
itself
like
nodes,
do
if
if
its
target
is
missing,
but
that
might
lead
to
some
other
kind
of
weird
behavior
so
as
much
as
I'm
kind
of
loath
to
like
sprinkle,
a
bunch
of
a
bunch
of
guards
throughout
the
code
base
for
something
like
that,
it's
probably
the
lower
risk
approach.
A
Yeah
and
I
I've
seen
lots
of
folks
actually
referencing
that
kind
of
error,
though
in
issues
recently
so
yeah.
I
think
it's
a
fairly
recent
regression
yeah.
C
Yeah
the
work
around
the
wii
is
just
like
with
the
npm
pack
or
pkg
command,
and
it
does
like
the
outcome
of
that
does
exactly
what
like
it.
It
structures
the
workspace
and
the
repository
with
the
log
file
and
it
updates
all
things
the
way
I
would
expect
it
to
so.
It's
sort
of
the
the
line
item
that
I
replaced.
C
What
would
be
whatever
command
we
come
up
with
or,
however,
we
make
that
work
in
the
future.
I
just
sort
of
will
one
for
one
replace
it,
so
I
think
it
should
be
pretty
pretty
simple
and
to
say
that,
like
if
the
hack
is
this
simple
as
it
is,
I
don't
I
imagine
that
it'll
be,
I
don't
know.
What's
it,
I
feel
like
that.
Jinx
is
gonna,
say
easy
to
fix,
but
I
don't
know
we'll
see
I
imagine
it
will
be
yeah.
C
That's
very
clearly
defined
what
the
what
the
thing
is
that
we're
trying
to
accomplish.
A
Remember
you
saying
like
it's
like
if
we
had
npm
ad,
because
we
want
to
essentially
install
or
add
a
new
depth
to
a
workspace
yeah.
The
way
we
got
around
that
was
yeah
using
pkg,
which
is
very,
very
hacky.
I
would
say
not
not
very
accurate,
but
it's
nice
that
we
have
that
command,
that
you
just
manipulate
the
package
json
and
we
added
the
workspace
to
that
cool.
A
If
not,
I
can
give
you
all
25
minutes
back
of
time
and
hopefully
chat
with
you
next
week,
cheers.