►
Description
A
B
C
D
Believe
that
that
version
that
release
also
included
an
updated
version
of
NPM
that
fixes
an
issue
that
people
are
having
with
NPM
CI,
not
listening
to
environment
variables,
which
was
breaking
chip
in
some
weird
ways.
So
just
so,
people
are
aware
of
that.
When
that
new
update
comes
out,
if
you've
been
having
issues
upgrading
to
anything
higher
than
I,
think
it
was
like
10.4.
The
problems
been
consistent
for
a
couple
months
now.
A
B
E
Things
that
are
things
are
going
pretty.
Well,
you
guys
can
hear
me
hope.
Yes,
okay,
perfect
things
are
going
pretty
well,
we
landed
a
run
on
Travis
that
runs
and
it's
it's
in
big,
bright
letters.
Python
3
is
experimental
in
capital
letters
but
at
least
we're
starting
to
run
Python
3
and
it's
building
its
it's
compiling
all
the
C
files,
etc,
etc.
It
does
spell
at
the
end,
because
async
is
a
key
word
in
Python,
as
well
as
in
as
in
JavaScript,
and
so
that
that
screws
this
up.
E
But
it's
nice
progress
and
we
landed
a
few
other
PRS
that
were
pretty
essential.
To
make
that
happen,
then
there's
three
more
that
are
open
and
interviewed
and
tested.
So
I
expect
those
to
land
in
the
next
24
hours
or
so
I
think
I'm
struggling
a
little
bit
with
with
the
configure
file
itself,
which
which
does
strange
things
to
Travis
CI
but
I've
gotten
around
it.
E
But
just
you
know,
running
directly
configured
PI
and
that
lets
me
do
a
bunch
of
things
so
I've
been
trying
to
make
test
make
install,
and
then
we
run
a
few
of
the
core
Python
files.
Like
tools,
tests,
etc,
and
so
so
I
think
we're
we're
making.
You
know
the
progress
that
we
hope
to
make,
and
it's
nice
that
we're
at
least
giving
other
contributors
the
ability
to
see
see
our
progress
in
a
way
that
doesn't
fail.
Travis
CI,
but
at
least
runs
Python
3
on
Travis,
CI,
I.
E
A
A
F
F
B
A
E
A
G
Yes,
so
not
much
has
changed
over
the
last
week,
I
think
there
was
no
new
comments
and
I'm
not
exactly
sure
about
how
often
this
move
for
bill
calls
each
introduce
some
potentially
a
controversial
changes
and
I
would
like
some
more
review
on
that.
There
were
some
comments
about
the
process
that
it
makes
in
two
aspects,
like
first
of
all,
ditches
fence
to.
G
Be
accessible
only
in
the
first
week
and
the
second
one
is
that
is
it's
not
so
I
flag,
but
an
API
I,
just
a
regular
method.
I
think
that
I
answered
both
of
those
on
github
and
I
would
like
to
keep
it
that
way,
and
after
that
there
was
no
new
input
there.
So
how
should
I
move
forward
with
that
because
I'm
not
yet
sure
from
the
technical
point
of
view,
these
still
needs
tests
and
benchmarks,
I
think
I.
A
G
For
now
we
can
need
to
move
some
distinctions
later.
I
think
I
mentioned
it
in
the
videos
me
challenge
in
the
github,
like
the
restriction
that
block
this
from
being
who's
gene
code,
after
that
hostak
could
be
removed
in
a
semi
life
change,
but
it
can
be
introduced
and
self-
change.
So
this
is
why
I
would
prefer
to
keep
it
that
way
for
the
time
being
and
then
see
if
that
consignee
issues
to
make
a
system
like
if
someone
is
not
happy
with
that,
and
they
do
lose
cases.
G
G
A
G
Yes,
I
think
there's
a
good.
This
is
not
going
the
way
it
is
designed.
It
is
not
going
to
be
used
in.
You
know
the
models,
so,
even
if
we
remove
that
it
won't
got
much
like
it
will
be,
a
similar
need
to
move
the
implication.
Self
I
do
not
have
to
wait
for
gold
be
removed
from
the
it's
a
cinch.
You
can
ensure
a
label
that
experimental
for
now
it'll
give
us
more
freedom
to
remove
it.
C
H
G
The
thing
is
that,
even
if
someone
ends
up
using
that,
they
could
easily
move
that
on
no
generation
change,
because
it
is
not
going
to
use
them
deeply.
Just
the
top
level
application.
So
I
don't
think
that
an
additional
warden
intended
that
I
want
to
see
people
starting
to
experiment
with
touching
the
applications
in
to
see
how
it
works
and
to
cache
buffer
problems
in
their
dependencies.
Using
is.
A
Just
the
challenge
with
a
warning
to
is,
it
would
show
up
for
users
who
can't
do
anything
about
it
and
if
the
app
developer
has
decided
they
want
to
turn
that
on
its
I
guess,
the
only
challenge
could
be
that
it.
The
code,
might
not
then
run
with
the
later
version
of
note
if
it
was
removed,
right,
yeah,.
B
H
H
H
A
H
G
B
G
B
A
I
C
G
I,
don't
think
the
change
is
a
good
wage
or
I.
Don't
think
that
you
even
have
a
good
way
to
determine
chain
development
and
sharp
action
mode
and
because,
like
people,
could
use
common
line
tools
in
their
development
attacks,
and
why
should
they
see
your
warrants
ideality
through
their
common
tools?
If
Daniels
and
I
don't
think
they
should
even
tell.
B
G
B
B
B
C
B
B
I
B
H
Week
be
fantastic
to
get
something
to
get
a
final
decision
on
this
I.
Don't
think
the
arguments
against
doing
this
have
changed
at
all.
It
increases
the
size
of
the
binary,
increases,
downloads
and
and
all
kinds
of
things.
I
know
that
rod
is
suggestions
for
potentially
dealing
with
the
additional
bandwidth
usage
of
the
local
download.
H
The
arguments
in
favor
also
have
not
changed
right.
You
know
having
his
additional
locales
available
all
when
we're
doing
localization
work,
so
I'm,
not
sure
I'm,
not
sure
if
there's
been
anything
that
has
moved
a
needle
and
that
either
way
on
this
discussion,
but
it
is
one
that
we
need
a
decision
on
to
do
it
or
not.
D
Research
that
I
can
add
some
some
stuff
that
I
found
out.
We
could
take
it
or
leave
it.
I
was
so
over
and
Google
that
this
would
have
potentially
of
some
sort
of
effect
on
our
server.
This
runtimes,
just
adding
to
the
binary
size
by
itself
is,
is
not
an
issue,
but
I
see
you
under
the
hood
is
actually
using
a
map
to
dynamically
load.
D
It's
not
necessarily
a
reason
by
itself
to
block
it,
but
just
kind
of
like
an
unexpected
edge
case
of
doing
this,
something
that
floated
onto
my
radar
actually
earlier
today
that
we
may
or
may
not
want
to
explore
I'm,
not
sure
if
it
would
be
helpful.
But
someone
sent
me
a
link
to
a
project
where
they're
courting
I
see
you
two
webassembly
I
just
popped
it
in
the
chat.
D
I'm
not
saying
that
this
is
necessarily
a
direction
that
we
should
go,
but
I
think
that
that
may
be
a
direction
that
can
have
things
like
loaded
into
the
binary
and
not
rely
on
like
nmap
operations
to
dynamically
load.
Stuff
I've
also
heard
some
like
James,
as
you
mentioned,
like
the
feedback
from
those
like
who
are
doing
a
lot
of
things
with
containers
not
being
super
thrilled
about
the
exploding
size
of
our
binary,
which
has
their
hand
up
now.
B
You
just
you
just
you
just
answered
what
I
was
gonna
ask,
which
is
you
know,
it
seems
like.
Opposition
to
this
has
been
whittled
down
over
the
years,
like
you
know,
been
used
to
be
opposed
now,
he's
now
he's
completely
the
opposite
of
opposed
and
so
on,
and
so
forth
and
I
was
gonna.
Ask
who's.
Could've
left
that,
like
doesn't
want
this
to
happen,
and
it
seems,
like
you
just
pointed
out,
and
you
know
I
mean
even
Rob
doesn't
seem
to
it.
So
much
as
just
hey
here
are
all
the
concerns.
B
H
D
And
it's
even
if
you
just
used
assault
a
small
segment
of
it,
it
loads
the
entire
everything
into
memory.
So
as
it
is
right
now
with
small
ICU
we're
getting
kind
of
we're
already
in
the
place
where
we
get
hit
by
that,
but
we'd
increased
the
hit,
I
guess
by
double
or
three
times,
you're
a
thief
drop.
You
know,
jump
alright!
Okay,
that's.
D
G
D
Use
it,
the
memory
gets
used
by
right,
okay,
whoops,
the
entire
ICU
into
memory,
it
eats
the
RAM
and
you
can't
use
it
right.
In
some
cases
you
take,
the
full
hit
is
what
you're,
saying
and
and
where
this
becomes.
A
big
issue
is
like,
especially
as
more
serverless,
runtimes
or
trying
to
run
node
with
as
small
of
a
profile
or
in
kubernetes,
trying
to
run
it
with
a
small
of
a
profile
as
possible.
B
D
H
A
H
Now,
longer-term,
you
know
I
think
just
moving
things
over
to
webOS
emember
I've
seen
those
conversations
too
much
and
I.
Don't
think
that
actually
solves
the
problem,
because
the
way
that
I
see
currently
works,
you
would
still
have
to
load
all
that
in
the
memory.
The
the
more
important
thing
I
think
that
needs
to
be
revisited
guys.
You
level-
and
this
is
something
that
has
come
up
there
multiple
times,
but
never
progress,
because
because
nobody
put
the
resources
into
it,
it
is
changing
the
way
the
resources
are
loaded
to
use
an
incremental,
on-demand
loader.
H
A
This
problem
ears-
and
maybe
we
could
take
this
back
to
the
thread
because
we
should
get
a
bunch
of
this
Meishan
in
there.
But
my
last
question
is:
is:
is
it
possible
to
bundle
Foley's
ICU
in
but
have
an
option
so
that
it
only
gets
em
mapped
in?
If
you
turn
it
on.
A
H
Could
do
we
could
effectively
bundle
the
full
ICU
module
kids
there's
a
way
of
basically
just
getting
the
the
full
data
file.
Essentially,
it's
a
separate
DLL,
but
it
basically
what
the
X
policies
module
was
friend
and
dynamically
linked
to
it
rather
than
static
right.
If
we
did
that,
if
we
distributed
that
the
full
ICU
with
with
it-
and
this
have
a
command
line
flag,
just
use
it
instead
of
using
the
statically
linked
small
ICU
there,
it's
an
option
it
would.
D
B
I
think
I
think
discussing
the
particulars
here
is
really
like
we're
just
going
to
repeat
it
in
the
issue
tracker.
What
we
really
need
to
discuss
here
is,
is
the
discussion
progressing
and
should
we
just
let
it
go
or
do
we
need
to
make
a
decision
or
do
we
need
to
do
something
else
to
push
the
conversation
forward,
but
actually
trying
to
have
the
conversation
here
is
probably
not
not
a
good
good
approach.
Rube.
H
G
H
A
I
I
think
I.
You
know
what
I've
heard
is
as
rich
was
saying
earlier,
opposition
seems
to
have
been
going
down
the
concerns
which
miles
has
raised,
but
you
know
we
need
to
put
that
discussion
into
the
github
issue
for
sure
cuz.
It
has
I,
don't
think
it's
there
yet
is
still
related
to
the
in
memory
footprint
and
and
is
any
so
if,
if
there
was
no
in
memory
footprint,
so
basically
that
the
binary
size
went
up,
but
the
in
memory
was
only
based
on
a
command
line
flag.
A
B
A
I
mean,
from
my
perspective,
I
I'd
be
okay
with
that
it's
you
know
it,
and
it
I
have
the
feeling
that
you
know
I,
since
people
have
been
less
worried
about
the
download
size
that
maybe
that's
a
comp,
you
know
compromise,
which
is
our
next
step,
I,
just
wondering
if
anybody
else
here
like
options,
nope
that
doesn't
work
for
me
cuz,
because
if
there
isn't
maybe
that's
what
we
go
back
to
the
tractor
with
and
James
was
saying,
you
know,
as
James
was
saying.
That
would
be
enough
of
our
consensus
to
move
forward.
Well,.
H
B
B
A
B
B
A
A
B
There's
a
lot
of
context
around
the
east
and
I
think,
and
these
got
added
you
know
in
the
last
24
36
hours.
I,
don't
think
anybody
other
than
maybe
Matteo,
maybe
James,
maybe
but
like
small
number
of
people
are
gonna,
be,
are
gonna,
be
prepared
to
like
look
at
those
in
weigh
in
at
this
time.
I
think
we
should
just
note
that
there
on
the
agenda
leave
the
TSC
agenda
label
on
them.
They'll
come
back
next
week,
but
but
ask
people
to
look
at
those
issues,
get
educated
form
an
opinion.
B
Even
if
your
opinion
is
you
don't
care,
because
that
let's
make
it
easier
to
make
a
decision
and
have
a
productive
discussion
either
in
the
tracker
or
if
nothing
happens
in
the
tracker,
then
we
can
talk
about
it
at
the
meeting
next
week
and
figure
out
how
to
move
forward
with
these
either
or
land
them
close
them
change
them.
Whatever
it
takes,
I
agree.
H
100%
there
are
a
very
large
number
of
PRS
right
now,
open
against
both
streams
in
HP,
HP
HP,
one
that
all
kind
of
need
to
be
taken
together
to
see
where
they're
all
going
before.
We
can
really
make
too
much
decision
on
these
so
and
because
there
are
such
a
large
number
of
them,
I
know
myself.
Matteo
we've
just
had
a
backlog.
We
have
not
been
able
to
get
through
all
of
them.
A
A
B
A
A
Okay,
so
moving
on
to
the
next
issue,
which
is
under
no
GS
that
admin,
the
open,
GS
/,
opened
JSF
proposal
to
move
governance
of
the
travel
fund
to
the
cross
project
council.
This
was
on
the
issue
last
week
and
you
know
it's
really.
I
haven't
seen
any
concerns
or
objections
from
the
TSC
side,
so
this
is
just
here
still
for
awareness.
A
You
know
I
think
within
the
next
few
days.
If,
unless
there's
any
objections,
it'll
be
configured
that
the
the
TC
and
I
don't
think
I've
heard
any
from
the
calm,
calm
side
you
know
or
in
agreement
with
this
and
it'll
move
forward.
So
if
you
do
have
any
concerns
feel
free
to
reach
out
to
either
myself
or
Matteo
or
a
comment
in
the
issue
itself.
Okay,.
B
D
D
C
A
Just
interactive
the
other
thing
is:
there
was
a
memory
leak
reported
on
function,
creation,
which
is
this
actually
turns
out
to
be
something
we've
known
and
is,
is
that
v8
doesn't
collect
function?
So,
there's
a
bit
of
a
discussion
going
on
in
that
front
to
see
if
that's
just
the
way
it
is,
or
or
at
least
we
need
to
engage
the
B
side
of
people
to
figure
out
what
the
history
is
on
their.
A
A
H
On
that
one
okay
I
know
that
there's
there's
a
number
of
efforts
that
are
that
are
ongoing,
I'm
working
with
a
new
contributor
that
I've
been
mentoring
here.
Just
a
bit
on
I
know
he's
going
through
the
Z
lied,
module
doing.
Some
promise
promise
application
of
that
and
I'm
hoping
to
get
another
another
imported,
but
we
have
to
go
through
I
know.
Efforts
are
ongoing.
This
just
okay
just
take.
H
Jeremiah
is
looking
for
input.
I
owe
him
some
input
on
the
new
stream
stuff,
but
basically
he
needs
other
people
taking
the
time
to
look
through
it
and
quite
a
feedback.
He's
kind
of
reached
a
point
where
stuff
forget
that
people,
okay
and
I'm
just
gonna,
say
I'm.
Sorry
I
have
not
been
able
to
sleep
anyway,
if
you're
out
there.
If
you're
listening.
C
A
Okay,
V,
so
movie
aid
open
web
standards,
I,
don't
know
miles
fee,
anything
to
say
on
that.
D
So
I
guess
a
high-level
question
to
everyone
like
since
we're
spinning
up
a
standards
group
up
and
like
at
the
foundation
level
underneath
the
cross
project
Council.
Would
it
make
sense
for
us
to
maybe
spin
down
the
initiative
as
a
nodejs
specific
initiative,
or
is
it
good
to
maybe
keep
it
as
general
updates
to
the
committee
I'm,
not
sure
what
the
best
course
of
action
would
be
right
now,
because
at
least
for
myself,
you
know
any
of
the
work
that
I'd
be
doing
would
be
at
that
CDC
standard.
D
D
D
H
B
H
Ongoing
making
really
good
progress,
I'm,
probably
two
weeks
away
from
opening
APR
against
no
just
note,
with
quick
behind
a
configure
flag
so
we'll
have
to
you
know
it
won't
be
built
by
default.
There's
lots
of
bugs
in
it.
You
know
in
it
and
lots
of
things
to
work
out,
so
I
don't
want
to
have
it
enabled
by
default,
but
once
we
have
that
PR
and
look
like
behind
the
config
flag
would
be
able
to
move
main
development
over
to
I
mean
repo.
So
happy
happy.
A
C
H
On
building
just
as
a
side
note,
this
is
something
we
can
talk
about
later
near
forum.
We
had
a
infrastructure,
we
were
automatically
going
out
and
building
distribution
images
within
30
minutes
add
like
the
docker
image.
Is
that
kind
of
thing
within
about
30
minutes
of
when
I
release
happened,
we're
actually
spitting
that
down
and
wanted
it
to
see.
If
there's
interest
in
having
those
scripts,
you
know
if
we
can
donate
those
scripts
or
whatever
bring
them
to
the
foundation.
H
A
B
H
H
H
A
I
think
it's
worth
opening
like
we
said
an
issue
in
the
repose,
the
main
the
main
delay
I
know
now
in
terms
of
the
official
docker
images
is
just
we've
got
to
get
the
the
docker
builds
done
in
docker
itself
and
I.
Think
that's
the
only
way
that
they're
official
so
yeah,
let's
open
it.
If
you
can
have
the
full
discussion
that
I
think
is
the
end
of
the
discussion.