►
From YouTube: Node.js Build WG Aug 9 2016
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).
B
So
welcome
to
the
node.js
foundation
bill
working
group
meeting.
This
is
the
ninth
of
August
2016
and
present.
Today
we
have
myself
Johann
Bergstrom.
We
have
Michael
Dawson
miles
Barnes
and
we
have
Hans
Christian,
flatten
and
I
thought
we
had
rich,
but
no,
it
doesn't
seem
to
be
around
and
I
guess
a
normal
fashion.
We
can
start
with
a
short
Stan
up.
B
You
guys
want
me
to
begin
or
yeah
sure
not
as
well,
okay,
so
on
the
exciting
side,
for
once,
I've
done
some
refactoring
on
our
ansible
repo
and
I
just
pushed
it
to
my
my
private
fork
of
or
not
private,
my
public
Fork
of
the
build
repo
which
gives
us
like
templating
for
ssh.
So
if
you
like
me
stitch
into
a
lot
of
the
machines,
it's
pretty
annoying
to
keep
everything
in
sync.
B
So
now,
I've
got
a
playbook
with
a
an
ass
will
plug
in
that
actually
parses
the
inventory
and
then
just
generates
all
the
hosts
in
the
ssh
repo,
and
it
also
works
with
ansible
sense,
will
actually
use
a
statement
or
a
puppy
I've.
Also
added
a
thing:
that's
very
helpful
for
us
moving
forward,
which
is
that
we
parse
our
host
names,
which
makes
us
able
to
write
logic
without
having
to
get
danceable
facts,
for
instance,
operating
system
or
or
architecture.
So
that
makes
like
our
branching
for
bootstrapping
much
much
quicker.
B
B
So
that's
pretty
cool
I'm
also
landed
Ukrainian
language
supports
for
node.js
org,
which
is
pretty
much
just
a
small
change
to
the
engine,
X
complete
and
then
I've
been
working
with
the
even
Lucas
and
phillip
Johnson,
which
isn't
here
now,
but
that
he
joined
us
recently
and
where
the
a
machine
that
tests
our
commit
messages.
So
hopefully
we
can
soon
have
commit
message:
validation
as
part
of
landing
landing.
B
Something
onto
master
I'm
not
quite
sure
how
to
implement
that,
since
if
we
would
have
it
as
part
of
the
test,
PR
job,
it
would
pretty
much
fail
until
we
do
the
final
squashes
and
the
final
notes,
because
it
expects
reviewed
by
your
URL
and
stuff
like
that,
so
I
was
kind
of
thinking
that
we
should
hide
it
behind
another
ship
books.
That
says,
you
know
with
final
review
or
something
like
that
and
I've
also
done
work
with
Phillip
on
deploying
the
github
boat.
B
We've
changed
it
from
and
we
moved
it
into
the
node.js
organization
and
we've
updated
it.
So
it
actually
does
a
bit
of
logic
on
its
own.
Instead
of
having
to
rely
on
Jenkins,
it's
Jamie,
we
were
having
problems
with
Jenkins,
basically
sending
us
different
communities
shalom's,
probably
because
we
rebase
as
part
of
fetching
the
the
commits.
So
now,
if
you
check
the
update
PR,
when
you
post
a
PR
job,
you
will
actually
see
github
getting
updates
from
Jenkins.
So
that's
pretty
awesome.
B
We
have
a
few
things
we're
going
to
add
to
that
shortly,
so
we
can,
for
instance,
if
it's
just
one
failed
test.
We
can
probably
like
just
mention
which
test
it
is
just
straight
in
the
test:
output
and
a
few
other
things
like
that:
I've
updated
both
CI
release
and
CI
interns
to
have
latest
Jenkins
versions.
B
Ci
is
still
on
1x,
but
I'm
also
preparing
to
do
an
upgrade
on
that,
but
that
requires
a
bit
of
coordination
and
everything
and
I'm
also
doing
a
lot
a
bit
of
local
testing
with
20,
with
the
same
plugins
as
in
CI
and
roughly
the
same
amount
of
hosts.
But
we
need
to
test
it
against
our
larger
data
set.
Then
I
think
especially
the
tap
stuff,
which
is
the
kind
of
our
biggest
issues
in
terms
of
performance.
A
Okay,
I
guess
we'll
go
in
the
order,
the
list
I'm
next
then
so
Kiki
and
I
I've
been
keeping
or
any
PPC
machines.
The
good
news
is,
they
seem
much
more
stable.
Now,
I
haven't
had
an
issue
in
the
last
week
or
more
I've
been
working
on
adding
the
new
AIX
machines
that
we're
getting
from
us
usl,
along
with
updating
the
manual
instructions,
if
they're
slightly
different,
you
know
different
version
of
urs
or
whatever
and
I'm
finding
things
that
I
can
get
right.
The
first
time.
A
B
A
C
C
We
should
be
able
to
use
this
to
see
if
there's
a
VI
breakages
that
specifically
meant
to
test
a
v8
5.1
on
veut
on
node
v6
before
we
move
into
LTS
so
that
we
hopefully
can
you
know,
get
a
more
modern
v8
in
there
before
we
go
into
the
LTS
cycle,
could
use
some
suggestions
on
native
modules
that
we
should
be
testing.
So
you
know
I'll
be
opening
an
issue
soon
about
that.
C
C
D
D
A
D
B
Okay
moving
into
the
minutes,
then
most
of
these
are
actually
from
last
meeting
and
I.
Think
mostly
because
we
haven't
really
closed
any
of
them
and
some
of
them
were
kind
of
open
discussion.
Hopefully
we
get
some
feedback
going,
but
there
hasn't
been
a
lot
of
activity,
but
I
guess
we
should
just
run
them
through
and
if
anything's
new,
we'll
just
bring
it
up,
otherwise,
we'll
just
gonna
have
to
hopefully
get
more
feedback
until
the
next
meeting.
B
C
B
A
B
Let's
do
that,
so
the
next
issue
is
run
and
deployed
no
just
get
out
but
number
404.
So
what
was
holding
that
up
was
pretty
much
I
guess:
I
raced
a
few
things
with
regards
to
security.
I
cannot
want
to
make
sure
that
we
keep
our
github
tokens
and
access,
as
you
know,
the
vector
as
small
as
possible.
What
we've
done
since
is
that
we've
moved
the
boat
to
Hardware
provider
or
VMs
provided
by
the
bill
group,
so
we
we
now
control
it
from
within
the
bill
group,
and
it's
pretty
much
me
and
Phillip.
B
B
B
So
when
we
start
a
job
in
Jenkins,
we
send
a
an
update,
saying,
hey
this
job
started
and
what
this
button
does
is
essentially
talk
to
github
and
just
posts
it
on,
which
means
that
it's
an
open
relay,
if
you
know
the
endpoints
yeah
so
on
top
of
using
tokens,
are
similar.
We
should
just
whitelist
our
Jenkins
eyepiece
just
too
yeah
to
avoid
any
further
security
problems.
B
Ok,
so
moving
on
then
OSX
buildbot
see
I
call
to
action.
It's
been
around
for
a
while,
and,
unfortunately,
I
don't
have
much
to
report
there.
Either
I've
been
in
trying
to
get
in
touch
with
the
red
fan
and
I've
also
had
not
looked
at
mack
stadium,
which
is
what
we've
previously
identify,
as
probably
the
most
soothing
being
provided
for
us
and
I.
B
Think
we
should
pretty
soon
try
and
push
this
to
ctc
or
similar
to
or
the
foundation
just
to
make
a
decision
that
we
should
go
with
that
we're
still
in
a
situation
where,
if
we
lose
the
boxer
machines,
we're
going
to
have
a
pretty
hard
time.
So
we
should
really
escalate
this
and
make
sure
we
get
access
to
resources.
B
B
B
Yeah
at
least
I
had
a
quick
shock
even
yesterday,
and
it
was
pretty
much
nothing
to
report
some
and
I
mean
will
again.
If
we,
if
we
sign
up,
for
you,
know
a
housing
provider,
it
doesn't
mean
that
we
have
to
say
there
forever
either
and
if
we
get
an
option,
meanwhile,
we
could
always
get
explorer
moving.
So.
A
I
think,
from
my
perspective,
the
next
step
would
be
to
ask
the
foundation
for
money
cuz.
I
don't
think
like
in
terms
of
raising
it
at
the
ctc.
I
don't
know
that
that's
necessary
because
you
know
build
stuff
has
been
delegated
to
this
work
group.
This
doesn't
seem
like
it's
out
of
you
know,
out
of
the
ordinary
or
anything
like
that
right.
A
A
C
So
one
thing
that
may
be
worth
taking
a
look
at
would
be
when
we
have
the
discussions
with
Google
as
the
V
V,
a
team
and
their
cloud
team
we're
talking
about
maintaining
v8.
It
was
mention
that
they
may
be
able
to
let
its
infrastructure
to
help
with
the
testing
for
v8
and
stuff.
So
in
I
don't
know
what
infrastructure
they
have,
but
I'm,
assuming
that
they're
likely
testing
chrome
on
all
of
the
deployment
operating
systems
so
may
be
possible.
A
I
wouldn't
delay
ourselves
too
much
on
that,
because
I
think
it
was
kind
of
an
abstract.
Okay,
we
might
be
able
to
give
you
some
horses.
He
as
opposed
to
yeah
that
so
I
I
I'd
say.
Maybe
we
can
ask.
We
can
look
the
things
in
parallel,
like
you
should
ask
Ali
just
in
case,
but
I
wouldn't
want
to
necessarily
have
that
delay
our
next
step.
That.
B
A
A
B
Ok,
no
problem,
ok,
so
moving
on
rsync
endpoint
to
mirror
the
releases
so
we've
since
last
meeting
at
least
have
I've
gotten
a
con
job
in
place.
So
the
only
thing
that
is
really
missing
now
is
just
ansible
scripts
to
set
it
up.
But
the
bigger
thing
here
is
really
a
bit
more
coordinated
effort
on
you
know,
perhaps
communicating
setting
a
public
deadline
on
when
we
will
switch
to
strict
and
tell
us
and
imperil
also
introducing
this
machine
for
those
that
needed.
B
So
that
requires
a
bit
of
you
know,
p
or
writing,
and
just
agreeing
on
a
date.
I
had
a
quick
chat
with
wrote
about
this,
and
this
is
pretty
overloaded
at
the
moment,
so
it's
somewhere
between
us
but
probably
on
my
desk.
So
if
anyone
wants
to
help
with
that-
and
with
that
I
mean
like
it,
all
news
item
know
about
this,
and
why
and
and
and
also
what
the
alternative
would
be
and
feel
free.
A
B
I
mean
I
think
we
we
through
nvm
already
have
had
a
similar
problem
for
a
long
time
where
the
child
is
having
like
a
super
old
version
of
curl
or
whatever
it
was
so
yeah
I
had
a
quick
chat
with
LJ
about
this,
and
and
he
was
supposed
to
do
some
testing
with
unencrypted
nodejs
org.
So
we'll
see
here
we
go
okay,
but
we
still
need
to
make
some
kind
of
blog
post
and
set
a
date
to
when
we
will
migrate
to
strict
Tillis
right.
A
B
Pretty
much
in
the
air
to
be
honest,
I'm
happy
to
put
something
in
wherever
it
needs
to
be.
I
can
give
you
all
the
information
about
the
unencrypted
so
ever
alright
and
then
for
the
ssl
ik.
We
might,
I
think
I
have
a
stubborn,
too
and
I
think
road
might
have
something.
So
let's
try
and
collect
that.
But
if
you
wanna,
you
want
to
hold
the
flag
on
this
one
that'd
be
awesome.
Yeah.
A
B
Okay,
so,
lastly,
we
have
Alpine
linux
/
dr.,
build
issue,
nomas,
75
and
I.
Don't
have
much
to
add
there.
It's
been
kind
of
the
problem
here
is
essentially
figuring
out.
The
lib
see
stuff
from
a
muscle
point
of
view
and
I
think
rod
at
the
moment
is
probably
in
the
best
position
to
to
drive
this,
unless
you
guys
have
done
some
homework.
I.
B
A
B
A
A
B
B
A
C
C
A
We
can
talk
about
some
of
the
feedback
too.
Okay,
so
assume
title
page,
not
too
controversial
or
myself
build
team.
I
thought
it
was
good
to
have
you
know.
Pictures
of
everybody
is
best
I
could
to
show
who
was
all
involved.
Make
sure
we
give
lots
of
credit
to
all
the
people
are
contributing
yep
our
approach.
So
you
know
our
preference
for
community
donations
anything
else
in
terms
of
like
our
fundamental
way.
We
do
things
we
should
be
calling
out
there.
B
C
A
A
A
A
Graphics
on
that
slide,
so
when
they
have
to
convert
that,
but
I'll
see
what
I
can
do,
but
we
want
to
make
sure
we
give
credit
to
everybody
who's,
contributing
both
the
infrastructure
providers
and
then
our
community
donations
and
exercising
that
hey,
you
know,
individuals,
you
can,
you
can
donate,
it
doesn't
have
to
just
be
companies,
yep
infrastructure.
So
then
talk
a
little
bit
about
a
very
structurally.
You
know
the
key
thing
that
we
have
weaves
Jenkins.
We
have
a
separate,
build
and
release
instances.
A
B
A
A
Not
so,
then
you
don't
talk
a
bit
about
our
building
release
infrastructures.
You
know
the
next
slide
is
just
a
point
out
that,
even
though
you
know
linux
is
a
single
west
list
there,
we
actually
you
know
test.
A
good
variety
of
releases,
also
go
into
a
little
bit.
How
that
you
know
we
test
the
different
versions,
but
we
end
up
building
on
centos
so
that
it
will
run
on
all
the
other
later
releases.
A
In
terms
of
the
bills,
just
showing
the
different
builds
that
we
have,
you
know
the
node
test
PR,
which
then
starts
the
test,
commit
and
sort
of
go
into
a
bit
of
detail
on.
You
know
what
you
see
when
you
come
to
the
builds.
Also
that
we
have
you
know
quite
a
lot
of
quite
a
lot
of
runs
of
those
builds.
So
you
know
you
can
see
that
they,
you
know
the
durations
about
20
minutes
and
that
we,
you
know,
are
running
quite
often,
you
know
several
times
an
hour.
B
B
A
Okay,
yeah
I'll
add
something
on
that.
In
terms
of
you
know,
you
can
maybe
putting
a
snapshot
of
the
the
make
file
itself,
or
at
least
mentioning
that,
to
that
extent
thing
hated,
you
can
see
we're
running
make
there,
but
there's
a
lot
of
actual
good
information
in
there
c'mon
delve
into
the
details.
Yeah.
A
A
B
A
In
terms
of
our
machines,
you
know
they're
mostly
virtual
machines,
except
for
our
arm
machines
and
our
bare
metal
bench
marking
machine
bill
did
mention
about.
Should
we
mention
that
the
arm
one
source
are
in
I,
don't
know
if
I
want
to
say
they're
in
rods
basement,
because
you
say
they're
hosted
at
posted
a
note
source
or
something
like
that.
That
makes
sense.
I
mean.
B
A
Okay,
so
yeah
I'll
grab
it
thats
great
idea.
I'll
get
the
picture
from,
hence
asos
today
not
source
a
lot
of
machines
to
manage.
So
how
do
we
possibly
keep
up
with
all
of
that
pull
the
answers?
Ansible,
you
know
it's
it's
on
the
next
page,
its
agents
agent
lists,
so
you
don't
need
to
install
anything
any
of
the
machines
it
connects
or
eight
SSH
and
really
there's
three
main
files
which
we
use
to
drive.
A
You
know
so
the
ansible
inventory,
which
tells
us
what
machines
we
have
to
configure.
You
know
that,
basically,
the
entries
are
this
header,
which
is
you
know
the
machine
type
and
then
the
instances
the
machines
themselves,
the
ansible
bars,
which
then
fills
in
some
specifics.
In
terms
of
you
know
what
the
ansible
bars
is
more
a
list
of
configurable
things
like
say,
packages
where
you
say
on
a
particular
machine.
You
want
to
install
this
set
of
packages,
refuse
name,
and
you
know,
for
example,
on
big
andy
and
even
there
you
know.
A
There's
this
example
of
you
need
multi
lips.
So
basically
it
lets
us,
you
know,
provide
data
to
the
plate,
the
main
playbook,
which
can
be
more
lists
oriented
and
we
can
use
to
actually
do
before
configuration
and
then
finally,
the
ansible
playbook,
which
is
really
the
file
that
defines
the
steps
that
we're
going
to
go
through
to
configure
any
particular
machine.
And
you
know
we
go
through
sorry,
go
ahead.
B
A
Yeah,
so
we
are
basically
the
playbook
worry.
You
know
we
have
a
number
of
steps,
and
so
we
basically
go
from
a
bear
machine
to
the
configuration
that
we
need
to
connect
it
up
to
the
CIA
and
basically
run
everything
that
needs
it
includes.
Steps
to.
You
know,
update
machine,
install
all
the
packages
that
we
need,
configure
the
I/o
Jess
user,
that
the
Jenkins
agent
runs
under
synchronized
time
date.
A
You
know
install
scripts
that
will
start
the
Jenkins
agent
on
startup
and
so
forth.
So
you
know
by
looking
at
that
script,
you
can
see
all
the
things
that
we
do
figure
and
it's
really
great,
because
you
know
you
really
just
pointed
at
one
or
more
machines
and
then
take
care
takes
care
of
doing
all
the
configuration
for
you
make
sure
that
we
stay
in
the
state
for
the
most
part
that
our
machines
are
pretty
much
disposable
or
that
you
can
just
dump
them
and
create
a
new
one
whenever
you
eat
yep
its
input.
A
The
next
thing
you
know
so
is
how
we
match
machines.
One
of
the
important
things,
of
course,
is
that
we
want
to
stay
nimble,
and
so
that
means
that
we
need
to
be
able
to
build
relatively
quickly.
So
we
do
mean
you
know
we
have.
We
do.
We
are
able
to
build
and
test
across
all
the
platforms
in
less
than
30
minutes,
that's
possible
because
we
implement
CC
cash,
which
basically
lets
us
avoid
recompiling
everything
every
time
it.
A
Basically,
you
know
if
the
compilation
would
end
up
with
the
exact
same
binary,
will
we
won't
do
the
compilation
and
just
use
the
cached
artifact?
We
use,
fan
job
so
in
cases
what
we
just
can't
make
individual
jobs
fast
enough
will
fan
them
out,
so
those
we're
actually
running
multiple
jobs
on
multiple
machines.
At
the
same
time,
and
in
particular,
that's
used
for
the
arm
and
windows
machines
to
go
speed
them
up
so
that
we
can
get
the
required
tests
across
all
the
different.
A
You
know,
arm
six-armed,
7
and
Windows
versions
fast
enough
to
be
able
to
meet
our
our
quick
turnaround
and
then,
finally,
you
know:
there's
been
a
fair
amount
of
work
in
the
test.
Scripts
themselves
that
come
with
node
2
now
enable
parallel
execution
so
that,
when
we're
running
tests,
even
if
it's
on
a
particular
machine
will
take
advantage
of
multiple
threads
to
run
tests
that
can
be
running
parallel.
All
at
the
same
time,
I.
B
That
would
be
on
on
poplar
operating
system,
/
argued
architecture
combos.
We
actually
have
multiple
machines.
So,
like
you
bunch
of
14,
we
have
like
two
of
each
and
I
think
that
the
benefit
of
adding,
that
is
that
it
makes
a
distinction
between
what
a
fine
job
is
vs
rice
running,
you
know
a
job
12
the
same
job,
twice
on
two
different
machines
episode.
B
A
A
You're
something
that
cuz
it's
doing
more
than
just
that
sure
yeah
exactly
I'll,
add
that
in
so
we
focused
a
lot
on
build
building
our
binaries,
which
of
course
is
very
important,
and
a
lot
of
the
infrastructure
is
dedicated
to
that.
But
the
build
workgroup
is
just
a
building
and
testing
those
binaries.
A
B
A
B
A
Okay,
I'll
add
in
the
backups
nursing
mirrors
is,
as
other
things
just
to
show
that
yeah,
it's
like
you
know
you
get
to
do
a
pretty
very
you
know
very
interfacing,
with
the
os's
and
infrastructure
and
stuff
with
that.
The
other
thing
is:
that's
that
you
know
the
built
as
important
to
build
workgroup
is
that
you
know
we
understand
that
collaborators
and
contributors
may
not
all
have
may
not
access
to
test
on
all
the
platformer
OS
combinations,
particularly
if
you
know
you
submit
a
PR
and
there's
a
regression.
A
So
there
is,
you
know
one
of
the
things
we
have
a
love
for
us
for
people
to
get
access
to
the
build
infrastructure,
to
test
them,
investigate
issues,
and
you
know
you
can
go
to
that
particular
document
and
the
repo
to
read
more
about
how
you
might
gain
access
to
do
that.
You
meet
a
few
things.
We're
currently
working
on,
so
I
guess:
OSX
support
you.
A
No
we're
still
looking
for
a
good
way
to
get
OSX
machines
of
all
the
different
levels
that
we
want
to
support
because
it's
been
harder
to
get
older,
OS,
x
versions,
and
it's
also
harder
to
get
you
know
them
hosted
so
we're
you
know
currently
looking
at
different
options
for
that
we
may
I
guess
have
it
solved
by
the
time,
but
I
think
it's
still
worth
mentioning.
Just
in
case
we
can
get
contributions
versus
having
to
pay
for
it
right
and
then
you're
saying
Jenkins
security
up.
B
A
A
B
Any
other
slogan,
the
reason
I
didn't
want
to
talk
specifically
about
that
is
that
you
know
it
doesn't
look
great
on
look
great
for
the
jenkins
guys
right
on
there
yeah
oh
yeah.
I
was
thinking
about
one
thing
as
well
like
when
I
mentioned
the
security
stuff
in
general,
like
if
it's
interesting
I
mean
some
people
might
ask
I
guess
it
might
be
on
more
relevant
question
to
us,
but
just
how
you
know,
how
do
we?
How
do
we
get
access
to
machines?
How
do
we
share
access?
B
So
I
think
we
have
a
pretty
good
blog
system
in
place
and
it
might
be
worth
mentioning.
You
know
how
we
are
we
delegate
and
how,
more
importantly,
how
like?
How
does
someone
join
our
organization
and
how
do
we
build
trust
and
like
how
do
you
activate
within
the
different
types
of
the
bill
group,
a
parts
of
the
Duggar
brother
right.
A
A
B
I
was
more
thinking,
you
know
as
an
organization
and
that's
an
open
source
organization.
It's
like
a
pretty
common
problem
is
delegating
trusts
and
yeah.
You
know,
if
there's
anything,
we
can
do
as
an
open
source
product
to
you
know.
Okay,
we
sold
it
this
way
and
we're
pretty
happy,
but
here's
the
challenge
or
whatever
kids
at
Lisa,
pretty
interesting
discussion.
If
it
pops
up
so
I
guess.
A
B
And
also
I
mean
the
trust
of
between
us
and
organizations,
and
you
know
who
has
access
to
release
me.
There's
a
lot
of
interesting
stuff
should
work
with
regards
to
trust
yeah.
It's
just
like
it.
You
can
get
pretty.
You
know
we
can
be
a
deep
dive
pretty
quickly.
So
I'm
not
sure
if,
if
we
want
to
talk
too
much
about
it,
but
yeah
have.
B
A
A
B
A
B
B
A
B
Okay,
so
unless
anyone
else
have
anything
to
add
our
keynotes
to
review,
I
think
we're
done
for
today.
So
the
next
meeting
will
be
at
the
end
of
august
august
of
30th
30th
of
August,
which
should
be
three
weeks
at
8pm.
You
two
seem
so
normal
time.
Yeah,
hopefully
we'll
made
some
more
progress
than
and
you
will
have
had
the
talk
as
well.
Then
right,
Michael,
I,.