►
From YouTube: Node.js Foundation Build WG Meeting 2016-08-30
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).
A
Yeah,
we're
live
okay.
So
a
warm
welcome
to
the
node.js
foundation
build
working
group
meeting,
and
this
is
the
30th
of
August
2016
and
we
have
a
new
participation
record.
So
here
with
us
today
is
me
Johan
we
have
Phillip,
we
have
broad
miles:
Michael,
Joelle,
Hans
and
Gibson
and
in
normal
fashion,
we'll
start
with
a
stand
up,
and
this
time
we
actually
have
Hans
first,
so
take
it
away.
Hans,
okay,.
B
B
It's
very
work
in
progress,
but
I've
put
up
a
pull
request,
just
to
start
getting
feedback
and
incorporating
that,
as
we
go
also
been
looking
into
adding
offline
reasons
to
Jenkins
alerts.
I've
noticed
that
Jenkins
has
some
labels
for
various
offline
reasons,
and
then
it
could
be
nice
to
add
that
to
their
IRC
messages
at
the
emails
and
also
been
talking
with
John
about
moving
the
junkies
monitored,
the
node
build
working
infra,
and
it's
currently
on
my
personal
detail.
C
Hi
I'm
sold
the
last
three
weeks.
There
was
about
one
week,
I
think
maybe
two
weeks
ago,
that
Windows
machine
started
failing
I,
never
understood
the
reason.
Some
completely
shut
down
some
Jenkins
just
bit,
and
this
lasted
for
about
a
week
and
stopped
happening
as
mysteriously
as
it
had
started.
So
during
that
week,
I
had
to
restart
the
machines
manually
a
lot
of
times
and
I
could
never
understand.
The
reason
for
this
and
I
have
also
been
working
on
a
release
machine
with
the
sauce
2015
for
windows.
It's
done
good
job.
C
A
D
A
A
A
Yeah,
I
think
we
have
a
nice
little
topic
about
platform,
support
and
official
platform
supported
yeah.
That's
gonna
be
interesting
anyway,
I've
been
doing
some
more
work
in
my
refactor,
it's
being
a
bit
slower
than
I
wanted
to,
but
if
you
guys
are
keen,
you
can
always
check
it
out.
The
good
part
is
now
that
we
support
pretty
much
all
posts
we
have
in
the
Bill
of
working
group.
A
So
if
you
want
to
have
a
sage
config
with
all
the
hosts
in
there
feel
free
to
try
it
out,
I
own
boarded
both
Phillip
and
Hans
has
been
with
us
for
a
while,
but
I
guess.
I
I
collected
the
information
while
working
with
Philip
and
try
to
apply
it
on
a
monster.
He
already
knows
most
of
the
stuff,
so
I
was
good
and
it
ended
up
in
a
document
that
more
than
me
hopefully
can
use
to
onboard
other
people.
A
A
We
have
a
semi
native,
a
Triton
host
on
joy
and
running
alpine,
we're
still
kind
of
talking
about
whether
we
should
use
that
or
use
that
you
meant
to
docker
container
for
testing
alpine,
because
you
know
that
would
probably
be
the
most
common
deployment.
So
in
the
end
we
might
want
to
do
both
I'm,
currently
working
with
a
PR
to
fix
nightlies
generation
from
linux.
A
We
landed
a
thing
in
node
two
months
ago
that
made
it
break
and
no
one
saw
it
until
now,
so
it's
classic
I
also
doing
some
work
with
Michael
to
redeploy
PPC
machines.
We
are
kind
of
moving
from
an
old
cluster
which
had
a
lot
of
problems
with
it
to
a
new
cluster
at
OSU
and
hopefully
it'll
go
much
smoother
from
there
on
and
oh
yeah
I'm
doing
some
commit
message:
validation,
testing
in
Jenkins,
with
even
Evan
Lucas.
A
So
the
the
thought
is
that
we
at
some
point
can
integrated
as
part
of
landing
a
PR
to
make
sure
we
don't
miss
out
on
commit
validation
and
I'm.
Finally
doing
some
get
out,
but
integration
work
Jenkins
when
Philip
and
hopefully
they
get
up
group
to
get
our
pod
bill-
will
help
out
in
an
adding
functionality
to
it.
So
that's
me.
D
D
There
were
some
test
issues,
basically
surfaced
by
the
fact
that
you
know
we're
now
running
them
much
more
often
and
in
a
higher
performance
mode,
and
that
these
new
ones
have
more
threads
and
can
build
faster
and
all
that
so
working
through
those
issues,
think
that
sits
back
in
and
not
green,
but
yellow
so
only
flaky
tests,
bailing
working
on
the
configuration
edition
is
you
had
mentioned
of
new
kpc
machines.
So
hopefully
this
will
be
our
final
production
cluster.
D
We
just
need
to
recreate
the
machines,
configure
them,
get
them
into
tight
and
sort
of
do
the
transition
from
the
old
ones,
also
working
on
the
addition
of
the
Linux
one
release
machine.
So
again,
with
your
hand,
you
know
adding
that
in
testing
it
out
to
making
sure
all
the
keys
and
everything
that
are
needed
it
going
to
sniff
test
the
binaries
with
the
goal
of
getting
it
into
the
standard
release.
Job
I
started
at
Knightley's,
but
now
that
six
is
actually
v.v
five
dot
one
landed
in
six.
D
We
can
actually
do
that
for
linux,
one
in
six
as
well,
so
I'll
be
looking
to
get
those
into
the
the
releases
that
go
up
for
six
and
for
the
night
least,
also
working
on
a
presentation
for
note,
interactive
EU.
So
thanks
for
everybody's
comments
and
help
on
that,
I
think
it's
looking
pretty
good.
I
spent
a
little
bit
of
time
fixing
up
the
benchmarking
jobs.
D
There
was
like
a
something
that
caused
some
dependency
issues
in
terms
of
MongoDB,
so
I
just
had
to
help
out
in
terms
of
updating
that,
and
just
very
recently
you
know,
I
had
some
discussions
with
one
of
our
SVT
team
members
who's.
Looking
to
help
move
the
code
coverage
forward
in
terms
of
getting
it
published
in
the
you
know
like
we
do,
for
benchmarking.
D
E
Yeah,
if
I
can
get
my
cell
phone
muted,
so
I
did
a
heap
of
work
on
the
cluster
in
my
garage
over
the
weekend.
I
reorganized
everything
in
there
irie
provisioned,
a
bunch
of
machines
that
were
like
had
altered
up
so
everything's
been
refreshed.
It's
all
nicely
laid
out.
It's
running
really
well
I,
also
added
two
new
machines
that
I
had
sitting
around
for
a
while
that
I
couldn't
get
going.
E
Dave
glass
donated
some
Oh,
droidx
use
and
I
hadn't
been
out
to
get
them
running,
but
I
think
it
again
with
them
and
managed
to
get
them
going
so
I
added
them
in
so
we've
got
30
droids
in
there
I
added
some
more
raspberry
PI's
that
I
had
delivered
and
yeah
that's
going
really
well.
What's
a
fun.
Do.
E
Two
more
during
the
week
as
well,
so
we
have
I'm
still
concerned
about
the
PI
ones.
We
have
12
of
them,
but
we're
two
of
them
are
used
for
release,
so
I
ordered
two
more
so
we
should
have
12
for
bills
and
two
for
release
and
then
that
sort
of
finishes
it
out.
The
only
discrepancy
is,
we
only
have
we
only
have
11
PI
threes,
but
I've
got
12
of
the
others,
but
that's
not
really
a
big
deal.
I,
don't
think.
Unless.
E
E
What
should
I
do
I
did
something
there's
something
else
was
going
to
report
on
that
I
did
have
forgotten
all
right.
I
did
want
to
comment
on
my
break
too.
If
you've
commented
on
Alpine
I
want
to
ask
hence
Christian
what,
if
there's
any
difference
between
your
alpine
work
and
the
pea
I
have
sitting
at
437,
which
was
that
was
to
add
it
add
alpine
support
to
see
I,
for
that
was
just
in
the
docker
container.
That's
running
in
CI
now
of
the
doctor
alpine
machine
running
on
digital
ocean
is
this
different?
Is
this?
E
A
Yeah
yeah,
but
I
guess
its
first
men
mentioning
that
I'm
talking
to
Christopher
Hall
over
at
joy
and
about
adding
a
kvm
instance
just
to
avoid
the
emulation
layer,
I
mean
we
still
have
to
figure
this
out
pine
thing
out,
but
in
the
long
term,
I
think
building
on
a
native
muscle
tool
chain
and
testing
in
both
dr.
native
would
be
the
optimal
scenario,
at
least
for
me.
E
E
By
a
follow
up
on
the
OSX
thing,
so
I
expect
to
have
movement
on
OSX
this
week,
so
I
am
in
the
process
of
exhausting
the
options
for
getting
a
sponsor
to
pay
for
that
and
if
that
doesn't
come
through
this
week,
then
I
will
follow
up
with
the
foundation
about
paying
for
that.
So
we
should
get
that
sorted
out
this
week.
Yeah.
A
F
So
this
is
my
first
stand
up
with
you
guys
so
hi
everyone
and
I'm
off
to
walking
my
dog
I'll,
probably
forget
a
lot
of
things
as
well,
but
first
of
all,
I
got
on
boarded
with
Johan.
That
was
great,
especially
the
ansible
part
of
it,
because
I
got
zero,
ansible
experience
before
hand,
so
I
really
needed.
G
A
Think
you
were
doing
at
least
we
talked
about
because
you
had
ansible
working
with
vagrants
and,
like
all
deploying
debian
aight.
I
think
you
know.
If
you
continue
to
you,
you
said
you
were
thinking
about
doing
like
a
readme
or
something,
and
I
think
that
would
be
awesome
for,
for
me
at
least
just
to
have
playing
around
locally.
F
H
We
have
that
smoke,
tester
job
for
ABI
breakages
that
I
had
working
just
before
I
left,
that's
running
off
of
you
know
a
forked
branch
of
carrying
the
gold
mine
so
I
have
some
work
to
do
and
getting
that
Forge
branch
I
get
some
tests
in
there
and
get
that
ready
to
get
landed
on
master
Chao
and
I
were
talking
about
a
canary
in
the
gold
mine
support
for
windows.
I
recently
got
a
windows
box
will
be
digging
into
that
outside
of
those
two
main
things.
There
is
not
a
ton
with
build
that.
A
Cool
keep
sandy
wanna
join
as
well
yep.
G
G
So
particular
having
a
look
at
ansible,
because
that's
completely
new
to
me
and
also
been
looking
at
GM
on
windows
and
we've
been
running
it
for
a
while
now,
but
recently
became
the
Syrian
issue
with
one
of
the
more
recent
surge
in
versions.
That's
causing
freak
adjust
for
awesome
windows,
so
I've
been
investigating
that,
and
also
looking
at
trying
to
get
a
test.
Mpm
updating
the
test,
10pm
scripts
and
working
windows
as
well
as
I,
could
have
progressed
out
in
node.
To
do
that,
that's
pretty
much!
It.
A
Okay
cool,
so
that
brings
us
to
the
agenda
and
first
on
that
is
issue
for
48.
How
can
I
project
qualify
for
access
to
the
building
for
structure?
This
guy
has
been
with
us
for
a
while
Raji
I'm.
Anything
you
want
to
add
I
know
we
have
one
new
comment.
No
one
has
replied
to
I
said
yesterday,
so
I
feel
a
bit
bad
I'm
regards
to
that.
But
you
have
anything
you
want
on
I,
bring
out.
E
D
Not
directly
related
to
the
discussion,
but
I
did
have
a
fairly
good
discussion
with
recon
bought
at
at
the
recent
node
live
and
I.
Think
you
know,
there's
going
to
be
proposal
to
move
node
serial
port
under
nodejs
like
github
nodejs,
which
may
make
it
a
little
easier
and
Lisa,
not
specific
instance,
to.
D
D
E
D
I
guess
do
we
have
any
cases
where
people
are
saying
hey
we
want
to,
but
we
don't
see
that
that's
a
reasonable
path,
like
you
know,
for
note
serial
port.
It's
like
it
seems
like
everybody's
receptive
to
that.
We
have
a
case
where
there's
somebody
saying
we
want
to
use
your
infra,
but
we
don't
think
that
would
ever
happen.
E
Not
explicitly
I
mean
I
had
discussions
with
some
people
and
the
leveldb
ecosystem
about
this
I
mean,
like
that's
a
I,
mean
because
I
I
work
over
there
as
well
and
do
stuff
there
and,
and
you
know,
doing
working
F
with
hair
and
then
switching
between
the
infrastructure.
We
have
and
then
going
back
over
there
and
having
to
use
Travis
and
a
player,
and
that's
it
it's
frustrating,
but
but
they
don't
fall
in
within
our
responsibility
at
all.
So
the
only
connection
is
simply
that
I
do
stuff
there
and
I.
Don't
think.
D
E
A
responsibility
I
think
it
has
to
be
much
more
fallen,
that
for
now
anyway,
unless
we
get
more
security
things
in
place,
then
I
think
that
the
limitation
has
to
be
that
the
tsc
has
some
responsibility
over
it,
so
not
just
the
ctc
by
the
tsc,
so
the
hot
of
the
whole
log
side,
whether
it
comes
in
in
some
form.
Otherwise
we
just
can't
support
it,
because
we
don't
have
the
security
infrastructure
in
place.
H
E
D
A
D
A
E
Guess
that's
a
matter
of
wording
then.
So
it's
not
that
we
are.
The
tsc
is
responsible
for
it,
but
it's
that
it's
I
guess
it's
been
authorized
by.
Maybe
it's
either
the
TSA
is
responsible
for
it
or
a
project.
Underneath
the
organization
needs
testing
of
some
project
for
shipping
its
own
product
so
like.
H
Within
reason,
yeah
cake
is,
for
example,
another
project
that
I've
talked
to
that
was
interested
in
our
resources
was
nvm,
and
one
could
argue
that
because
the
number
of
people
who
are
consuming
node
via
nvm
is
not
minor,
that
us
testing
it
on
infrastructure
as
value.
But
you
know
now
we
kind
of
get
into
you
know
like.
Where
do
we
draw
that
line?
I.
E
Think
we
can,
we
can
probably
add
something
to
them.
That
says
that
there
has
to
be
a
negotiation
we
have
so
the
tsc
and/or
CTC
can
negotiate
with
the
bill
working
group
to
add
a
project
that
falls
outside
of
the
scope,
and
then
we
could
look
at
those
trust
links,
and
so
we
could
figure
out
what
are
the
relationships
we
have
with
nvm
that
make
that
something
where
we,
as
the
build
working
group,
can
sign
off
on
the
security
and
the
ctc
or
tsc
can
justify
wanting
to
have
that
those
resources
used.
E
A
Okay,
moving
on
then
next
issue
is:
do
x.x,
build
boats,
see
I
call
to
action.
Issue
number
367.
This
guy
has
also
been
with
us
for
a
while
and
since
rod
just
mentioned
that
the
latest
update
here
is
essentially
that
will
give
exhausting
sponsors
another
week
and
after
that,
we'll
probably
move
to
our
commercial
option.
I,
don't
think,
there's
too
much
dad
besides
that
or
do
you
have
anything
you
want
on
road.
E
No,
but
I
think
this
this
doesn't
need
to
move
off
the
agenda
before
we
have
another
meeting.
We
should
have
this
resolved
by
then
it's
just
like
we
had
a.
We
had
a
problem
last
week
where
I
was
fixing
it
over
sex
machines
and
one
of
them
didn't
come
back
up
after
a
reboot
Thank.
Thankfully
it
was
like
I
realized.
We
could
actually
ssh
into
the
Machine
and
do
a
reboot
that
way,
and
was
it
was
just
the
screen
sharing
that
wasn't
working,
so
something
to
do
with
that
boot
process.
E
That's
happened
before,
but
if,
if
we
had
a
problem
where
it
didn't
come
back
up
at
all
and
we
had
to
reach
out
to
somebody
I,
but
we
haven't
got
a
contact
there
at
voxel
that
is
responsive
enough,
that
you
know
we
could
just
heat
up
and
say:
hey,
can
you
go
and
just
pull
the
plug
and
plug
it
back
in
Ian?
Please?
So
it's
a
real
problem.
It
gotta
be
resolved
by
the
next
meeting.
E
A
Yeah
and
I
guess
the
other
question
that
relates
to
that
is
also
the
platform
thing,
but
let's,
let's
not
jump
ahead
of
the
agenda.
So
the
next
thing
on
our
agenda
is
the
rsync
endpoint
to
marry
the
releases.
This
is
issue
number
55
and
the
latest
update
here
last
time
we
spoke
about
this
I
was
to
send
Michael
some
notes
about
this
and
he
would
help
out
in
writing
something
a
bit
more
formal
week.
A
Poston
ojs,
org
I
sent
him
something
earlier
today,
so
I've
been
holding
that
up,
but
the
idea
here
is
to
do
a
trial
run
somewhere
in
after
we
cut
the
new
LTS
over
a
weekend
or
something
and
then
a
month
later
we
switch
nodejs
org
to
strict
HSC
s
and
prior
to
this
we
just
do
we
communicate
this
publicly.
So
that's
kind
of
the
plan
and
I
guess
we'll
try
and
collaborate
about
this
in
public
before
we
proceed.
A
A
A
A
yeah
now
you
do
it.
Let
me
just
start
by
saying
so.
This
is
a
documentation
required
by
the
tsc
to
the
github
boat
repo
and
as
a
result
of
that,
the
discussion
kind
of
came
up
whether
or
not
this
should
be
an
independent
working
group
or
if
it
perhaps
should
be
a
subgroup
of
the
bill
group
for
now.
B
Yes,
so
William
and
added
a
pull
request
with
them
required
a
governess
documents
where
he
had
proposed
to
make
it
and
a
separate
working
group,
while
others
on
the
build
working
workmen,
they
have
proposed
the
boat
working
group
to
be
under
the
build
working
group,
essentially
yeah
I
guess
that's
the
main
main
topic
for
this.
This
issue
is
whether
or
not
this
should
be
under
the
standalone
working
group
directly
under
the
TC
or
under
the
build
work.
A
Yeah,
so
my
thoughts
on
this
matter
is
at
least
that
it's
kind
of
in
its
inception
and
I
think
what
we
should
focus
on
is
just
no
progress
and
just
living
under
the
bill
group.
Where
we
have
you
know
you
have
people
in
both
groups
in
both
hands
and
Philip,
which
are
active
contributors
through
the
blood
it's
part
of
in
both
groups
and
I'm
kind
of
joining.
That
group,
as
well
so
I
think
for
me
at
least
having
it
as
a
subgroup
to
the
bill
group.
B
Yeah
and
since
William
is
not
Harold
and
r
afraid
our
say
his
argument,
he
he
is
talking
about
the
broader
collaboration
aspect
of
the
boat,
which
he
he
says
is
is
larger
than
just
a
technical,
built
kind
of
with
ad
posting
and
pull
request
status
and
so
on
and
where
he
says
it's,
it's
important
for
this
to
be
as
standalone
as
possible
to
to
give
it
a
broad,
except
in
the
organism
in
the
node
organization.
Basically
so
that's
kind
of
the
other
side.
Oh
there
are
women's.
A
Yeah
so
I
I
think
we
suggested
that
the
other
working
group
or
that
the
getup
bought
group.
We
have
some
kind
of
brainstorming
session,
so
we
tried
to
map
out,
like
all
the
things
that
people
want
to
achieve
with
it.
So
I
think
like.
Unless
someone
disagrees
in
the
bill
group,
we
could
at
least
say
hey
you're
welcome.
F
A
Okay,
if,
by
the
way,
if
someone
could
help
out
taking
notes
in
the
document
on
multitasking
tabs,
here,
it's
not
going
too
well
anyway.
So
let's
just
agree
that
we
have
a
tentative
welcome
if
they
want
to
come
and
and
then
we'll
see
what
what
the
next
steps
from
that
group
is
before
we
make
any
form
of
decisions,
then,
and
that
means
we're
moving
on
to
the
next
topic.
A
A
Oh
sorry,
rich,
isn't
here
sorry
my
bad
I
just
assumed
it
would
be
okay,
so
so
this
this
is
essentially
about
and
we
I
guess:
we've
been
on
the
subject
a
few
times
already.
We
need
to
at
some
point
say
which
support
which
platforms
are
operating
systems
no
Jess
will
officially
support
to
run
on
and
I.
Guess.
We've
been
pushing
this
ahead
of
us
for
a
good
while,
and
we
just
haven't
found
a
good
way
of
you
know
defining
this.
What
rules
do
we
apply
to
say
that
we
support?
A
E
Feel
free,
so
I
think
I
think
that
it
should
be
up
to
the
CTC
to
decide
what
they
want
to
officially
support,
but
I
also
think
that
the
build
working
group
should
be
free
to
tinker
with
hardware
we're
adding
new
things
as
it
sees
fit,
not
not
not
under
the
assumption
that
they
would
be
officially
supported.
So
an
example
of
this
would
be
our
pine,
where
we
did
we're
doing
work
to
mess
around
with
the
tsc.
Has
it
the
CTC
hasn't
necessarily
said?
Yes,
we're
going
to
officially
support
our
pine.
E
That
was
just
sort
of
an
hour
hoc
thing
that
the
only
problem
with
that
is
that
a
you
know
we
at
the
moment
anyway,
we
don't,
we
can't
really
add
infrastructure
and
then-
and
it
not
be
part
of
the
official
build
set.
Because
then,
when
you
run
the
tests
and
it
fails,
then
it's
failed.
So
I'm
wondering.
E
Is
there
a
Jenkins
set
up
somewhere
that,
where
we
can
add
hard
and
say
failures
of
this
hardware,
don't
you
know
should
come
up
as
a
yellow
rather
than
a
red,
and
they
shouldn't
fail
the
whole
the
whole
thing,
because
if
we
could
do
that,
then
we
could
add
new
hardware
like
proposed
or
experimental
hardware,
and
we
could
we
can
have
the
you
know
the
dynamism
that
we
want
within
this
group,
but
the
CTC
can
also
nail
down
what
they
care
about
and
not
actually
have
to
be
constantly
shifted
by
what
we're
tinkering
with.
Oh.
A
Yeah,
that's
that's
a
good
way
of
at
least
introducing
new
platforms
and
hardware.
How
many
what
what
we
could
do
is
you
could
add
all
of
this
hardware
to
the
current
Jenkins
set
up
and
then
just
add
a
label
like
staging
or
experimental
or
whatever
levels
we
want
to
introduce,
and
then
we
could
expose
that
to
the
PR
job
or
we
just
copy
that
job
to
a
new
job
and
say:
hey
run
over
this
label
said
instead,
I.
E
E
We
just
have
to
make
it
make
it
a
really
clear
signal
for
any
any
collaborator
and
then
in
the
noches,
like
in
the
node
repo,
when
they're
looking
at
Jenkins,
make
it
clear
to
them
that
this
thing
doesn't
necessarily
matter.
So
don't
go
chasing
your
tail.
If
this
thing
fails-
and
maybe
we
do
that
with
the
name
of
the
job
or
the
sub
job.
A
Yeah
I
mean
winks,
clearly
our
communication
effort
and
we
can
care
to
collaborate
or
read
me.
We
can
add
it
to
our
read
me.
We
can
probably
make
a
note
in
Jenkees.
We
can
hide
it
from
the
Jenkins
no.js
tab.
This
is
the
different
ways
we
can
with
the
sold
out,
but
I
mean
like
that's
it
I
guess.
If
you
really
want
to
push
everything
to
see
to
see
that
that's
the
simple
response
from
our
group
done
right.
D
I
just
think
it
was
this
concern
was
also
but
like.
No,
it's
more
like
you
want
to
be
all
add
something
to
the
the
domain
test.
P
are
potentially
without
having
it
make
it
go
red
because
I
think
that
was
the
biggest
complaint
that
there
were
things
you
know.
Aix
wasn't
quite
stable
enough
and
it
caused
more
red
than
was
desirable.
D
E
D
E
A
Know
so
what
you
can
do
is
like
put
the
we
have
this
plugin
for
like
handling
post
post
test
jobs.
It's
basically
where
we
pick
up
the
tap
file
and
do
a
few
other
things,
for
instance
post
to
the
github,
but
so
in
there
you
can
actually
do
like
branching
based
on
scenarios.
So
you
can
say
if
this
job
failed
or
went
unstable
or
went
I.
A
Think
there's
like
five
different
statuses
of
Jenkins,
so
they
like
make
a
difference
between
a
narrower,
build
and
a
failed
build
because
of
tests,
and
you
can
essentially
just
branch
based
on
this
and
then
even
update
the
post
starters.
So
you
could
say
this
job
succeeded,
even
though
it
failed,
but.
A
D
A
So,
like
I
think
we
need
to
check
what
build
statuses.
Jenkins
supports
I
mean
we
have
success,
we
have
failure,
we
have
error
and
we
have
unstable
and
then
there's
something
else
so
like
we
can
choose
between
those
five
in
terms
of
visualizing
how
the
build
went
and
what
I'm
saying
is
that,
like
regarding
how
the
build
actually
went,
we
can
modify
post
that
level,
so
we
can
kind
of
manipulate
it.
But
at
the
same
time
we
don't
have
a
lot
of
options
to
choose
from
so
I.
A
A
You
know
we
just
removed
fedora,
20
and
21,
just
because
we
thought
it
was
old
because
of
upstream,
but
you
know,
do
we
supported
only
supported
so
like
we've
kind
of
been
no
just
doing
our
own
thing
like
growing
our
own
shop
here
and
that's
why
I
think
just
having
TTC
just
create
a
list
to
get
wood
together
together
with
us,
I
guess
and
then
try
to
treat
that
a
bit
more
formal.
Now
that
brings
up
something.
E
Else
that
I
need
to
mention
before
I
forget,
which
is
that
the
latest
version
6
actually
fails
on
clang,
3.4
and
we'd
only
noticed
that
when
we
tried
to
build
for
the
Debian
packages
for
wheezy
on
the
the
download
source
distributions,
so
we
haven't
got
wheezy
packages
for
640,
because
clang
3.4
doesn't
work
like
there's
something
in
there,
something
in
the
compile
in
the
new
code
for
the
v8.
That
doesn't
work
with
that
and
we
never
picked
that
up
as
a
bill
working
group,
because
we
don't
do
any
claim
slam.
This.
A
E
A
H
A
Okay,
well
so
at
some
point
we
compile
a
list
or
do
we
let
the
CTC
do
that
and
then
we
try
to
cater
those
permutations.
E
E
A
E
A
A
Yeah,
because
that
could
actually
drive
other
people
too
to
help
out.
You
know
if
we
had
a
like
this
is
a
wish
list
of,
for
instance,
alpine
34.
We
would
be
pretty
like
high
up
there
in
terms
of
what
we
should
be
doing
right
now,
so
you
know
having
that
fully
visible
for
those
people
to
come
and
say
hey,
I
tried
this
or
try
that
and
you
can
do
this
nut
would
be
good.
A
Okay,
anyway,
so
I'll
start
with
that
and
then
we'll
move
from
there,
then.
A
A
We
at
least
suspect
that
one
of
the
big
biggest
reasons
to
why
we're
spraying
so
much
time
and
cpu
time
on
Jenkins
is
that
we
use
a
tap
plug-in,
which
is
essentially
every
time
we
run
our
tests.
We
export
these
tests
in
a
tab
format,
and
then
we
have
a
plugin
for
Jenkins
that
consumes
this
and
then
embedded
into
build.xml.
So
for
every
time,
Jenkin
runs
a
job.
It
creates
like
a
xml
file
which
contains
information
about
the
building.
A
So
what
happens
here
is
that
the
tap
plugin
embeds
the
information
in
the
build
and
then
for
every
time
it
actually
reads:
XML
file,
it
chooses
to
reparse
it,
which
yeah
consumes
a
lot
of
memory
and
then
once
run
out
of
memory,
we
get
a
lot
of
GC
and
yeah.
Then
we
sit
there
and
wait
for
a
while.
So
we've
had
a
few
suggestions
from
how
to
solve
this.
First
of
all,
we
could
jump
into
Java
land
and
try
to
to
modify
this
plug
into
our
well.
A
A
E
H
E
H
Thing
as
well-
and
this
is
kind
of
a
unexpected
benefit-
that
the
UI
for
diving
into
breakages
for
the
j
unit
reporter
is
just
far
better
than
the
tap
one
you're
able
to
see,
especially
if
you
have
failures
from
multiple
jobs,
all
the
failures
propagate
to
the
top
job,
you
can
start
doing
a
dive
in
without
actually
having
to
click
through
like
five
layers.
So
I
have
not.
I've
noticed
not
just
a
performance,
enhancement
and
time-saving
in
using
the
application,
but
actually
for
myself,
just
like
digging
through
results.
It's
been
a
lot
better.
Okay,.
E
So
I
guess
we
have
some
candidate
options
here
and
it
really
just
requires
somebody
doing
the
work
I
did
propose
like
when
you
and
I
were
talking
about
as
Johan
last
week.
Doing
it
in
Java
actually
wouldn't
be
too
hard
because
the
codes
already
there,
the
like
the
tap
program,
plugin
for
Jenkins,
does
this
already.
So
it's
just
a
matter
of
pulling
that
out
and
writing
it,
and
we
have
Java
olive
al
Jenkins
modes.
E
So
it
wouldn't
be
that
big
a
deal
to
do
it
that
way,
I
mean
I
could
do
it
in
Java,
but
that
would
mean
that
I
have
to
actually
spend
time
on
it
and
I
I'm
wary
of
committing
to
doing
more
stuff,
because
I
know
I
don't
get
the
stuff
down
I'm
supposed
to
do
it.
Now.
We
have
expertise
on,
though
only
in
this
group
for
Python,
even
JavaScript,
so
I
guess
it's
just
a
matter
of
somebody
putting
up
their
hand
to
do
it.
E
H
Willing
to
raise
my
hand,
to
write
a
note
utility
to
do
this,
since
I've
already
played
with
a
good
chunk
of
like
actually
generating
parsable
j
unit.
Xml
I
already
know
how
to
get
that
up.
The
off
the
and
going
and
parsing
tap
is
actually
not
all
that
difficult.
There's
a
tap
parser,
that's
inside
of
tap
that
is
event-based
and
can
do
this
pretty
quickly.
H
E
D
A
A
Friendly,
no,
no,
so
we
we
can
write
to
output
and
then
it
bites
to
an
output
file
in
the
build
directory.
And
then
we
tell
the
script.
That's
part
of
the
post
build
action
to
consume
it,
and
then
we
just
disable
the
type
plug-in
completely
because
we're
writing
a
build
XML.
And
then
we
tell
Jenkins
to
pick
up
the
XML
SJ
unit.
So
like
I,
think
we
have
a
pretty
clear,
or
at
least
I
think
I
have
a
pretty
clear
path
forward
in
terms
of
how
to
treat
this.
A
E
A
E
Okay,
that's
actually
sounds
like
a
door.
I
guess
that's
an
easy
way
to
do
it
then
so
yeah
miles.
If
you
don't
think
that
that's
a
ton
of
work,
then
then
we
could
experiment
with
that.
Maybe
you
could
just
write
like
a
something
really
simple:
it's
just
like
a
proof
of
concept
that
doesn't
do
the
full
thing
that
may
be
pronounced,
oom
to
tap
and
spit
out
like
a
really
basic
j
unit.
We
could
then
use
to
verify
what
your
hunts
suggesting
here
is
possible.
Yeah.
A
You
know
I'm
pretty
certain
and
we
can
get
this
work
as
like
by
running
it
on
the
host
at
least
I
played
around
with
it
and
I
could
at
least
like
process
things
on
the
house.
So
it
shouldn't
be
a
problem
if
we
install
node
and
this
utility.
H
Yeah
so
I'm
willing
to
go
through
sub
stack
as
a
tap
parser
that
I'm
a
I
think
I
can
have
a
proof
of
concept
together,
pretty
quickly
and
yon.
If
I
could
get
you,
you
know
like
an
example,
tap
file
in
just
a
little
binary
to
run,
and
then
you
could
play
with
having
it
run
on
the
Jenkins
host
and
the
week
you
decide
if
we
want
to
go
full
force
with
it.
Yep.
H
A
A
A
D
A
Cool,
so
just
to
finish
this
off,
the
next
meeting
will
be
set
timer
20th
at
the
same
time,
I'll
try
and
we
should
try
and
get
an
email
out.
This
felt
like
a
week
ahead.
Just
so
we
don't
forget.
So
we
get
like
an
invitation
stuff
and
also
it
would
be
great
if
I
could
get
some
help
with
setting
up
all
the
YouTube
and
hang.
E
D
E
E
I
think
we
might
go
to
the
conference.
In
fact,
that
might
be
the
easiest
way.
Then
we
don't
do
any
setup.
We
just
don't
we
just
connect
in,
but
we
can
discuss
that.
But
I
don't
know
I
don't
know.
The
timeline
is
for
the
hangouts
on
the
air
to
go
away,
but
I'll
get
you
at
it
doesn't
have
been
for
that
that
you
can
do
this
without
having
to
rely
on
Michael,
okay,.
A
Cool
sweet,
so
thank
you
very
much
for
attending
today
and
I'll
see
you're
on
ok.