►
Description
A
A
Okay
on
to
the
CPC
and
board
meeting
updates
on
the
board
front,
the
meeting,
if
I'm
remembering
correctly,
is
this
Friday
I
didn't
have
any
issues
on
the
board
last
time,
I
checked
so
I.
Don't
think,
there's
anything
that
we're
waiting
on
the
board
for
if
there's
anything
that
there
is
just.
Let
me
know
on
the
CPC
front,
I
guess
a
couple
things
to
call
out
might
be
the
I
think
like
last
week.
B
We
continue
to
flesh
out
the
project
progression.
Stuff
we've
had
a
couple
projects
come
in
inputted
applications
they're
in
a
silent
period.
Right
now,
that's
when
we
have
internal
discussion,
you
know
if
anyone
from
this
group
or
who's
involved
with
the
foundation.
If
you're
listening
wants
to
be
involved
in
that
process,
then
you
know
please
start
attending
the
CPC
meetings
and
nominate
yourself
as
a
regular
member.
You
could,
you
know,
become
part
of
this
process.
The
more
folks
we
have
helping
out
my
personal
opinion,
the
better
okay.
So.
A
A
B
So
I've
got
a
new
pull
request,
that's
open,
it
has
seven
day
bill,
and
so
that's
great
it
was
you
know
actually
right.
There
was
only
one
more
patch
that
we
needed
how
the
team
got
it
extremely
close.
We
were
almost
there
that
PR
I
was
having
some
issues
with
Windows
I'll,
follow
up
with
build
on
that,
because
I
think
they
were
infrastructure
related,
that's
Sam
by
the
way
needs
to
be
promoted.
Oh
thank
you.
The
only
thing
I
guess
on
here
is
it
is
semver
major.
When
was
the
cutoff
for
13,
okay,.
A
B
A
B
A
C
B
C
Planning
to
do
an
Aussie
today,
so
I
guess
just
once
it
landed
in
muster
each
time.
I
refresh
the
Aussies
I
know
they're,
not
really
Elsie's,
but
there
will
be
one
a
week
from
now
on,
a
Tuesday
are
just
ping
TSE
on
the
list
of
new
majors
they've
gone
into
master
and
asked.
If
there's
any
objections.
A
A
B
A
A
B
Great
yeah
and
the
patches
I
think
there
was
only
like
two
factors
or
something
like
that.
We
needed
to
float
to
remove
language
features.
So,
overall
you
know
it's
pretty
straight
forward
and
I
was
building
a
test
with
you
know
the
newer
version
of
clang
with
7.9
and
it
seemed
to
work
fine.
So
we
should
be.
You
know
good
the
foreseeable
future
with
that
update.
Okay,.
B
A
Okay,
so
it
was
Matteo
who
added
it
to
the
agenda.
We
don't
have
Matteo.
Today,
though,
I
think
the
discussion
you
know,
the
the
sense
was
that
you
know
of
the
in
terms
of
the
TSC
people.
Who'd
weight
in
the
preference
was
to
make
it
opt
in.
My
read
of
the
feedback
from
the
tooling
team
was
that
it
was
also
for
opt-in.
A
A
A
A
G
D
D
A
Okay,
so
moving
on
to
the
next
one,
the
next
thing
is
ongoing
list
of
actions
for
Python
3
compatibility,
number
25,
789
Klaus
is
out
in
adding
in
a
list
of
the
key
issues
which
are
sort
of
the
next
steps.
I
think
Sam,
you've
been
fairly.
On
top
of
that,
do
you
want
to
add
anything
cover?
What
he's
got
there.
A
H
Okay,
so
so
thanks
for
a
lot
of
help
from
from
Sam
and
all
the
others,
we
made
some
nice
progress.
This
week
we
moved
the
dependency
repo
from
outside
of
our
get
up
or
into
the
inside
of
our
goetaborg.
So
we
now
have
a
new
repo
called
tap
to
j-unit.
Tap
to
J
unit
is
something
that
we
use
in
the
Jenkins
process,
and
it
basically
was
blocking
our
ability
to
run
tests
in
Jenkins.
H
So
we've
now
brought
that
inside
we've
made
it
Python
3
compatible,
that's
landed
there,
that
is
in
review
and
once
that
lands
I'll
push
it
back
up
to
the
Python
packaging
index
so
that
we
can.
We
can
continue
with
Jenkins
tests,
so
I
think
that
that
one
has
really
made
beautiful
progress.
Thank
you,
Sam
and
then
the
second
one
is
that
we
we
sort
of
have
a
we
sort
of
have
a
strange
relationship
with
gif.
H
H
We
have
to
publish
gift
so
that
NPM
picks
it
up
so
that
we
can
bring
it
back
in
it's
one
of
our
dependencies
and
that
that
is
blocked
by
a
Mac
bug
that
I
need
to
fix
and
I
now
triangulated
on
it,
and
we
know
how
to
replicate
it
in
our
testing
so
that
one
I'll
fix
and
then
that
will
kick
off
a
chain
of
events
where
we'll
publish
a
new
get
and
PM
will
pick
up
that
gift
and
then
we'll
bring
it
back
in
this.
Whatever
so,
I
think
those
two
things
are
probably.
G
E
H
Ones
we
reckon
yeah
the
one
where
I
could
really
use
help
is
if
somebody
could
go
into
Travis
CI
and
just
create
a
new
job
run
for
testing
against
ninja
I've
tried
that
a
couple
times
and
not
been
successful.
It's
probably
something
quite
simple
for
somebody
who
has
the
right
perspective,
but
I
don't
so
all
we
want
to
do
is
be
able
to
test
configure
against
configure
a
ninja
inside
inside
Travis
CI.
E
But
in
this
particular
case
system,
there's
an
ongoing
conversation
about
whether
or
not
there's
a
lot
of
configure
options
and
we
don't
test
test
them
all
so,
where
we've
been
doing
we've
kind
of
on
an
ad
hoc,
what
always
gonna
be
at
hoc
basis,
you're
testing,
some
of
them
in
the
docker
containers
inside
the
what,
with
the
docker
container
building
VI,
what
I'm
not
sure
that
we
any
other
there's
an
ongoing
discussion.
It
basically
said
what
came
up
with
like
what's
special
about
ninja.
What
about
all
the
other
options?
E
H
E
Yeah
Travis,
is
it
simple,
I
guess:
stick
with
the
encryption
but
I've,
maybe
I'm
wrong,
but
I
think
our
main
CI.
We
use
it
more
as
a
smoke
test.
Our
main
CI
is
still
yep
and
no
tip
itself.
I,
don't
know
if
you
I'm
just
catching
up
on
my
notifications
this
morning.
Maybe
we
don't
have
to
talk
about
the
slides
and
a
fancy
meeting,
but
it
sounds
like
those
between
the
reports
of
no
chip
is
working
with
Python
3.
E
H
A
H
H
E
A
E
We've
been
able,
we
figured
out
a
way
to
actually
we
haven't,
figured
out
an
approach
or
even
really
tried
too
hard
to
testing
Python
through
NCI.
That's
still
the
question
of
how
many
we've
talked
about
this
I
mean
Michael
suggested
that
maybe
our
nightly
should
use
Python
threes,
but
the
CI
is
during
the
day
on
PRS.
She
is
Python
2
that
might
get
us
decent,
decent
coverage
of
Python
3,
but
not
burden.
Every
single
PR
build
with
a
full
Python
3
and
Python
2.
E
A
E
No
chip
itself
needs
to
be
republished.
Okay,
I
might
have
missed
a
an
update
about
this
since
yesterday,
but
yeah
no
chip
master
works,
okay
for
Python
three
okay,
because
it
works
on
Linux,
but
it
sounds
like
it
doesn't
work
on
Windows
and
Mac,
but
anyhow
it
needs
to
be
republished
so
that
it
can
get
incorporated
to
the
NPM
and
circle
back
into
into
node.
E
E
And
there's
both
have
the
quality
that
they
they
involve.
Interaction
with
teams
that
aren't
particularly
active
there
are
now
Christian
is
a
maintainer
of
top
to
j-unit.
So
we
don't
have
that
problem
well
with
no
chip,
you
know
so
maybe
maybe
this
is
relevant
for
the
TFC
in
general,
like
anybody
who
has
any
interest
or
time
to
help
support
note
chip,
it
would
be
well
worth
it
we
to
make
it
super
clear
in
this
meeting.
Well,
we've
done
to
make
no
chip.
Work
is
best
categorized
as
a
hack.
E
It's
like
there's
enough
little
changes
that
have
gone
in
to
make
it
work
with
Python,
3
or
close
to
pretty
sure,
but
it
doesn't
really
address
the
bigger
question
of
how
to
maintain,
in
the
long
term,
there's
a
there's,
a
new
for
tip
3
of
no
chip
which
replica
was
supporting
and
looked
like.
It
was
developing
really
well
and
there
was
I
think
we're
close
to
switching
to
that
fork.
But
then
you
know
it's
not
as
active
in
the
project
anymore.
E
H
And
just
so
that
you
understand,
my
long-term
goal
is
to
is
to
sort
of
do
what
we
do
with
this.
This
new
you
little
utility
tap
to
G
and
and
do
that
same
thing
with
gift
that
is,
make
it
a
package
that
we
can
publish
to
pi
pi
and
then
import
it
instead
of
vendor
it
into
our
software
at
three
or
four
times
you.
E
Know
yeah
that
question
I
suspect
that's
not
going
to
happen,
but
that's
a
longer
conversation
there's
some
there's
some
strong
reasons
to
stick
with
rendering
so
I
come,
but
but
to
get
better
maintenance
and
to
move
it
into
the
node,
the
node
jazz
organization.
So
we
can
control
the
publishing
life
cycles
triggered
and
faster
would
be
fabulous.
D
E
I
mean
if
it
would
be
I
mean
it's
impossible,
like
you
can
imagine
the
javascript
file
that
would
read
the
binding
jib
rewrite
it
into
a
scene,
make
text
file
and
then
run
see
make,
but
then
installing
see
make
would
become
a
prerequisite
for
using
NPM.
But
you
know:
NPM
has
other
pre
works.
If
you're
gonna
build,
you
know
binary
add-ons,
you
already
have
to
have
full
compiler
Xcode
installed
or
the
correct
compilers
in
the
correct
Python.
So
it's
not
impossible
to
say
you
will
now
need
this
extra
third-party
tool
installed
before
running
in
panels.
E
A
E
And
it's
it's
also
worth
mentioning
that
you
know
again
there's
in
a
way
there's
two
issues
so
the
chip
we
with
some
amount
of
work
we
could
probably
build.
We
could
absolutely
build
nodes
with
C
make
or
any
of
the
myriad
build
tools
out
there.
So
node
itself.
We
have
a
fair
amount
of
ability
to
compete
with
some
amount
of
work.
We
can
just
change
to
a
new
build
tool
and
that's
our
business.
E
It's
our
code,
Betty's
having
builders,
it's
our
business,
but
that's
even
if
we
moved
away
from
chip
for
node
that
doesn't
get
rid
of
the
the
the
ecosystem
problem.
So
there's
it
would
just
it
would
just
mean
that
we
diverge
from
the
ecosystem
and
now
we're
kind
of
maintaining
two
systems,
but
you
know
they're,
so
they
that's
a
quick
recap
of
the
conversation
every
time
it
goes
around.
A
E
A
E
Dead-End,
it's
like
we're
out
of
Google
Maps
driving
down
a
road,
but
it's
got
the
wrong
date
and
who
knows
when
we're
gonna
that
some
it
remains
a
problem.
We
don't
really
have
a
strong
strategy
and
what
we're
doing
with
Python
3
is
gonna,
hopefully,
is
gonna
get
that
soon.
We
hope
you
know
Python
3
support,
but
it
doesn't
mean
that
we
saw
digit
problem
and
there
we
have
a
long
term
solution.
H
So,
just
to
wrap
up
well,
while
we've
been
talking,
I've,
been
updating
the
current
ongoing
list
and
I
defined
something
called
round
trip,
which
is
basically
this.
The
short-term
near-term
notion
of
what
we
need
to
do
and
I
put
down
three
action
items
that
I'm
going
to
be
attacking
in
that
order
to
try
to
execute
the
kit.
They
get
Brown
trip.
A
A
A
A
E
A
Think
the
only
like
I
don't
see
it
as
a
problem.
Necessarily
the
only
argument
Ike
it's
one
of
the
arguments
I
could
see
playing
devil's
advocate
is
that
that
it
provides
a
little
bit
more
direct
link
to
the
TSC
in
terms
of
if
there's
problems
or
if
there's
help
needed
it
would
give
you
that,
but
I
mean
I,
think
the
moderation
team
has
been.
You
know
pretty
good
about
pulling
in
the
TSC
and
the
calm
calm
when
needed.
A
B
I'm
not
personally
concerned
about
it
and
I
also
think
you
know
if
we
do
find
that
we
want
to
be
more
hands-on
at
any
point.
Any
of
us
can
come
here
and
get
involved.
So
I
don't
think.
Like
I
I
would
be
concerned,
if
falls
in
there
with
some
sort
of
barrier
to
entry,
but
it
seems
like
mostly
it's
just
about
timing
and
prioritization
I,
don't
see
it
as
an
issue.
A
A
B
They're
moving
forward,
although
we've
hit
a
tiny
bit
of
a
hiccup,
hopefully
we're
gonna
get
a
proposal
together
this
week,
but
essentially
the
flag
caused
like
the
way
that
we
set
up
the
flag
meant
that,
like
nodes
of
bootstrap
without
the
flag,
is
different
than
nodes.
Bootstrap
I
know
the
bootstrap
with
the
flag
is
different
than
the
bootstrap
without
the
flag,
which
means
that
removing
the
flag
changed
the
bootstrap
and
further
there
is
like.
Essentially
all
the
unit
tests
in
node
core
are
all
expecting
the
bootstrap,
the
way
in
which
it
happens
without
the
like.
B
You
know,
without
including
the
flag
or
without
removing
the
flag.
So
when
we
did
some
initial
test
and
removing
the
flag,
we
identified
some
unexpected
regressions
that
were
chasing
down
most
of
most
of
the
regressions
were
like
a
non-issue,
but
it's
not
really
something
that
we
can
make
kind
of.
Like
lazy
loaded,
it
does
change
the
Buddha.
The
biggest
issue
that
was
noticed
was
we've
needed
to
proxy,
is
in
place
for
built-in
modules,
specifically
to
allow
them
to
stay
in
sync
between
common
Jas
and
ESM.
B
B
You
know
like
coming
up
so
soon.
My
gut
is
that
we
are
not
going
to
necessarily
be
in
a
place
to
remove
the
flag
before
we
move
into
LTS,
but
I
think
the
hope
would
be
to
land
anything
that
we
may
need
to
land
to
get
like
performance
characteristics
consistent,
so
that
you
know
if
we
do
remove
the
flag
at
some
point
during
LTS
we're
not
all
of
a
sudden
seeing
a
major
performance
regression
on
the
line.
B
Now
with
that
being
said-
and
this
is
a
little
tangential-
we've
identified
some
pretty
major
performance
regressions
that
have
happened
in
12
independent
of
this.
So
my
hope
would
be
that
we're
gonna,
hopefully
see
like
a
15
to
20
percent
increase
in
performance
when
we
fix
these
regressions.
So
even
if
we
introduce
regressions
by
removing
this
flag
like
the
overall
performance
story
of
12,
will
be
net
positive
by
the
time
these
changes
land.
A
B
F
A
A
F
F
A
F
I
Might
be
currently,
the
biggest
problem
is
probably
a
web
platform
tests.
I
know
the
joy
is
been
looking
into.
Those
and
I
can
try
to
get
them
to
work,
but
that
seems
that
the
execution
model
for
the
web
platform
tests
for
web
crypto
isn't
really
working
for
the
node
environment
right
now,
so
I
haven't
really
looked
into
anything
else.
Yet.