►
From YouTube: Node.js Diagnostics WG 2018-10-17
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
Yes,
his
live
okay,
yeah
I,
don't
know
like
Michael.
If
you
wanted
to
I'm
sure.
B
I
mean
basically,
we
did
an
intro
to
some
of
the
the
key
things
we've
been
working
on
just
to
give
people
who
were
there.
There
was
a
number
of
people
who
I
think
weren't
as
involved,
so
it
was
like
a
heads
up,
hopefully
getting
people
interested.
Then
there
was
the
deep
dive
you
gave
into
the
sort
of
formal
async
terminology,
formalization
I'd
say
the
punchline
from
that.
B
C
So
I
summarized
the
common
issue
in
the
Diagnostics
repository,
but
in
summary,
we
are
going
to
try
to
use
the
unhindered
rejection
events.
That's
already
available
note
and
try
to
read
TRO
from
that,
as
well
as
introduced
some
extra
information
on
the
ever
object
in
v8.
So
we
are
going
to
open
a
request,
pH
to
add
more
data
to
the
site,
traces
so
that
we
can
inspect
that
on
another
node
or
or
another
to
write.
D
C
A
That
needs
to
happen.
There's
some
additional
work
that
needs
to
happen
there,
but
I
think
they
sort
of
you
know-
and
you
know
I
think
with
you
know
full
awareness
that
what
we
have
now
is
you
know
nothing
more
than
a
set
of
slides,
but
I
think
there's
potential
here
sort
of
raises
the
question
of
well:
what
is
a
sink
cooks
and
is
a
stand?
Cooks,
necessary
and
I
think
there's
probably
some
use
cases
that
require
a
sink
hooks
to
still
be
there,
but
I,
not
necessarily
net.
A
For
a
sync
contacts,
tracking
and
so
I
think,
given
that
and
I
think,
given
the
uncertainty
around
it,
sort
of
it's
the
kind
of
thing
that
it
shouldn't
yet
come
out
of
experimental
on,
but
we
can
talk
more
about
the
I.
Think
there's
an
open
issue
on
the
agenda.
I
just
wanted
to
clarify
I
say
that
there's
no
confusion.
A
So,
moving
on
the
Diagnostics
best
practices
guide
to
11
I
mic
from
node
source
that
he
actually
had
written
some
stuff
up
on
this
I,
don't
know
if
he's
on
the
call
on
one
thing
so
on,
but
the
I
think
we
need
to.
We
need
to
find
a
way
to
move
this
forward
and
I
think
the
last
time.
What
we
said
is
that
you
know
we
just
need
some
point
have
a
set,
maybe
a
separate
meeting
to
kind
of
kick
it
off.
Yeah.
B
B
A
So
the
deep
dies
I
think
you
know
we
did
a
deep
dive
on
the
async
context,
stuff
I'm
willing
to
do
that
again
on
a
hangouts
or
something
I
think
it
probably
useful
for
folks
who
aren't
there.
So
you
know
look
for
that.
I!
Think
sometime,
maybe
in
the
next
few
weeks
I
could
probably
do
it.
I
don't
know
if
there's
any
other
deep
dive
topics
that
anybody
wants
discuss.
C
A
A
F
A
G
D
A
H
Think
it
would
depend
on
who's
interested
in
participating
it
the
way
I
interpreted.
It
was
that
it
doesn't
necessarily
have
to
be
this
precise
lot
that
we
use
for
the
diagnostic
work.
You
know
it
would
be
dependent
on
the
topic
and
who
is
interested
in
participating
and
whether
we
can
find
a
mutually
agreeable
time
working
on
certain.
A
A
I'd,
certainly
like
your
feedback
on
the
proposal
and
so
I
think
when
we
get
to
scheduling
that
and
like
thinking
about
my
calendar
between
now
and
like
the
next
two
weeks,
I
don't
feel
like
it's
gonna
happen
until
then,
but
maybe
next
next
meeting
two
weeks
from
now,
we
can
talk
about
what
the
like
we
can
get
that
scheduled
on
and
we
can
find
a
time
that
sort
of
works
for
folks
in
various
various
places.
So
does
that
sound
fair?
A
A
Okay,
so.
B
B
So
I
guess
for
today,
unless
there's
you
know,
I
wouldn't
say:
if
somebody
has
something
they
want
to
talk
to
talk
about
or
a
specific
one.
We
could
do
that.
Otherwise,
it's
it.
It's
a
matter
of
us
figuring
out
what
what's
the
next
things
we
want
to
do
to
help
support
that
you
know.
Is
there
more
testing
we
need
to
get
in
the
CI
or
or
so.
I
A
Sounds
like
there's
really
no
sort
of
no
clear
updates
here,
and
we
should
just
also
knit
again
next
time:
I'm
here,
yep,
okay,
so
a
sing
cooks
table
API
tracking
issue
number
124,
Allie
you've
been
speaking
to
this
one
you
want
to
chime
in.
H
I,
don't
think
there's
any
update
this
week.
There
was
I
mean
the
latest
I
hear
is
that
I
know
is
that
there
was
some
discussion
of
reviving
the
they
think
resource
availability
on
the
I
think
cooks,
API
is
this,
was
materials
PR
and
given
that
the
HTTP
parser
may
be
may
no
longer
be
blocking
I,
don't
know,
but
andreas
has
more
to
say
about
this:
oh
yeah,
that's
the
only
update,
I
think.
G
G
G
G
I
G
Capture
when
it's
gaps
collected
yeah,
so
on
that
note,
I
made
recently
a
pull
request
about
removing
the
promise
from
the
promise
wrap.
So
that's
a
resource
object.
You
get,
and
this
is
something
we
talk
about
a
long
time
ago.
Just
got
to
it
now,
and
the
idea
here
is
that
we'll
make
like
escape
analysis
each
in
the
future,
for
example.
So
that
could
be
like
a
different
solution
to
try
and
solve
some
of
the
the
expensive
area
destroyed,
looks
for
promises.
G
H
Yes,
so
one
of
the
things
we
should
do
is
we
had
we've
been
saying
this:
we
had
that
Matteo
we
good
to
have
him
and
both
young
in
this
meeting
and
actually
sort
of
getting
together
to
talk
about
in
more
depth.
I
think
and
I
think
this
at
I'm,
not
a
doesn't
work,
so
maybe
the
best
thing
we
can
do
is
schedule
like
just
like
a
deep
dive
and
make
sure
the
matteo
and
young
can
be
there,
and
there
is
so
I'll.
Take
an
action
item
to
get
this
scheduled.
H
A
A
B
I
say
I'm
gonna
ask
to
reach.
If
he
can,
can
you
open
a
doodle
to
see
you
find
it
better
because
I
know
he
has
trouble
making
this
time
so
I
just
asked
if
he
can
create
a
doodle
to
see
if
we
can
find
a
time
that
or
get
maybe
not
be
better
for
everybody,
but
find
one
that
gets
a
larger
percentage.
People
be
able
to
make
it.
F
Related
to
the
destroy
hook
there,
there
was
a
bunch
of
conversation
at
the
summit
just
about
like
the
the
fact
that
Asian
cooks
currently
like
models
make
their
resource
concept
as
like
both
handles
and
other
thing
to
not
consumable
resources
which
is
confusing,
and
that
sort
of
results
in
like
this
destroy.
Okay,
resisting
for
promises
when
it
doesn't
actually
logically
really
make
sense
he's
like
the
only
concept
of
destroy
it
would
potentially
apply
to
his
garbage
collection,
which
we
don't
actually
care
like
it's.
At
least
a
PM
doesn't
actually
care
about
that.
F
We
just
care
about
the
data.
We're
attaching
to
the
thing
disappears
eventually,
so
we
can
just
stick
it
on
the
object
and
let
it
be
garbage
collected
on
its
own,
which
is
kind
of
why?
There's
that
current
resource
request
in
the
first
place
like
if
we
could
better
detach
these
concepts
of
the
tracking
of
handles
from
cracking
of
the
asynchronous
context
that
we
wouldn't
actually
need
destroy
for
context
tracking
at
all.
A
A
A
Natalie
paying
me
yesterday
about
trying
to
sort
of
take
some
of
those
ideas
and
get
an
implementation
like
a
monkey
patch
implementation
for
node,
with
the
caveats
of
where
that's
been
not
going
to
work
and
so
I've
been
sort
of
trading,
some
mails
with
her
and
Brian.
Is
it
and
and
just
kind
of
like
hey?
This
is
maybe
what
the
shape
of
these
things
look
like.
I,
don't
know
if
you've
had
any
more
thinking
on
that
Natalie
and
or
Brian.
J
K
A
A
Properties
over
from
that
original
function
so
that
you
can
expose
them
like
if
you
were
to
just
sort
of
look
at
all
of
the
properties
on
the
thing
and
then
just
copy
them
for
now.
Maybe
that's
not
exactly
the
right
thing
to
do,
but
I
think
at
a
minimum.
It
would
unblock
things
so
that
people
are
able
to
be
like
okay,
I
kind
of
see
how
these
things
are
fitting
together,
and
you
know
it
can
be
called
out
as
like:
hey
wait,
a
minute.
This
doesn't
quite
work
and.
A
K
A
And
I
wonder,
though
it's
like,
if
you
you
know
it's
at
some
point
like
and
so
I
think
the
question
here
is
like:
if
you,
if
you
say
that
you
know
you
continue
a
5
a
function,
app
is
the
result
of
that
triple
equal
to
X
and
I.
Think
the
answer
that
is
no
right,
yeah
like
it
should
be
no
and
so
and
I-
think
that's
okay,
right,
I!
A
K
A
Yes,
interesting
yeah,
okay,
yeah,
like
you
know
like
if
there's
anything
you
know
like
blocking
you,
there
might
be
like
an
interesting
thing
to
get
like
whatever
code
you
guys
are.
Writing
like
on
you
know,
github
somewhere,
so
that
people
can
actually
kind
of
see
it
and
poke
holes
in
it.
The
other
thing
I,
don't
know
how
we
would
do
with
just
like
if
you
were
to
try
and
solve
it
from
monkey-patching,
is
get
that
hole,
ready
contact,
stuff
set
up
because
I
think
you
need
some.
A
J
We
have
various
the
holes
of
promise
instrumentation
that
we've
worked
with
in
the
past,
so
we
can
just
sort
of
take
one
of
those
strip
out
our
own
logic
and
just
try
and
paste
it
in
I
think
we
can
get
pretty
darn
close
a
VM
level,
one
definitely
better,
but
at
the
very
least
we
can
get
close.
Okay.
A
Sweet
yeah
also,
let
us
know
like
how
it
goes
I
think
you
know
it
would
be
great
to
see
like
progress
on
this
I.
Think
a
couple
of
other
like
points
here
I
think
we're
gonna.
Do
we're
gonna
sort
of
socialize
this
more
with
some
folks
on
that
the
a
team
I've
got
to
socialize
it
more
with
a
couple
people
here
on
chakra
team
on,
and
so
you
know
like
in
particular
curious
as
to
lately
I
want
to
talk
to
you
later
argent
expert,
you
know:
have
him
poke
holes
in
it.
A
A
A
H
H
H
There
were
some
race
conditions
that
are
fixed
in
the
past
couple
of
weeks,
so
yeah,
so
the
participants
is
getting
more.
Mature.
James
is
working
on
adding
more
instrumentation
in
various
places,
but
one
of
the
main
things
we
want
to
need
at
this
point
is
sort
of
more
experimentation
from
folks
in
the
community
on
what
what
should
be
the
default
set
off,
trace
events
that
you'll
be
emitted
by
a
node
like
what
should
be
the
baseline.
H
What
what
should
you
get
if
you
don't
specify
any
specific
categories,
and
one
of
the
other
things
that
that
would
be
night
should
have
at
this
point
is
maybe
a
benchmark
that
basically
we
can
that
make
sure
we
don't
regress
the
default
trace
event
performance.
The
idea
would
be
that
tracer
man
should
be
usable
in
production
and
it
should
be
a
finite
like
pre
deterrent,
like
we
should
have
an
idea
of
what
overhead
we
want
to
ask
when
traces,
an
event
is
enabled.
H
A
It
seems
like,
though,
like
the
benchmarking
thing,
do
you
have
any
insight
into
what's
going
on
with
the
benchmarking
thing,
because
I
thought
like
I
know
that
there
are
some
folks
talking
about
it
in
Vancouver
and
I
kind
of
feel,
like
there's
a
pretty
big
gap
in
terms
of
where
current
benchmarking,
you
know
the
current
benchmarks
available
for
node
and
where
I
think
people
would
want
to
be,
and
also
I'm,
not
sure
like
if
there's
any
sort
of
ongoing
like
perf
runs
that
happen
like
are
there?
Yes,.
B
I
mean
working
Michael
can
talk
about
it.
Yeah
I
mean
that
there
is
a
working
group.
It's
not
had
a
huge
amount
of
you
know,
there's
been
a
few
people
participating,
so
it's
like
you
say
we
certainly
don't
have
the
coverage
we'd
like
we
do
run
Acme
air
and
DCE
is
along
with
a
couple
of
other
benchmarks
nightly,
and
we
have
the
charts
on
benchmarking,
no
jested
org.
They
have
on
occasion
found
a
few
regressions,
but
I
wouldn't
be
confident
that
they
would
cover
all
regressions.
Certainly
there's
no
version
of
them
that
turns
on
tracing.
B
Know
if
we
thought
that,
like
turning
tracing
on
and
running
Acme
air
was
one
of
those
or
DCIS
or
you
know,
one
I'm
not
sure
we
can
turn
on
all
the
benchmarks
and
have
a
duplicate
set.
But
if
there
was
one
we
thought
would
have
reasonable
coverage.
That
might
be
a
possibility,
or
maybe
we
need
something
completely
different.
B
H
B
A
A
B
B
Is
Jamie
Davis
we
might
have
seen
at
the
conference
Wow?
It
sounds
like
he's
planning
to
to
spend
the
reasonable
amount
of
time.
I'm.
Also
talking
with
the
group
at
UNB
that
I
work
with,
and
they
have
somebody
coming
in
in
January,
so
I
might
be
able
to
get
them
to
work
with
Jamie,
but
I.
Don't
think
we're
gonna
solve
that
bigger
problem
overnight,
but
if
you're
interested,
if
anybody's
interested
in
helping
we'd,
certainly
like
more
help
on
that
front,.
B
B
A
B
A
B
A
B
A
That's
the
thing
is
like
sometimes
in
the
past,
it
was
like
we
said.
Well,
it's
11:00
a.m.
yes
and
when
we,
when
he
has
to
eat
shifts,
then
everybody
else
shifts
whether
or
not
you
do
the
daylight
savings
thing
or
not.
And
then
in
some
cases
it
was
like.
Well,
it
was
11:00
a.m.
PST
and
then
something
shifted
and
it
moved
to
like
whatever
noon
and
whatever
was
right
and.