►
From YouTube: 2020-08-28 Java SIG
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
C
B
C
C
On
the
the
java,
the
open,
telemetry
java
crew
yeah-
we're
excited,
I
know
john-
has
been
excited
for
some
help
on
the
the
java
sdk
project
for
a
while.
C
Now
so
I'm
sure
we'll
find
lots
of
good
stuff
for
you.
C
So
we
had
so
usually
zoe.
We
on
this,
the
six
o'clock
meeting
here
or
the
friendly
for
tokyo
and
now
you're
in
australia.
C
Sydney
all
right
so
sydney
and
tokyo
friendly
time.
We
just
chat
about
things
that
we
talked
about
in
the
thursday
morning,
sig
meeting,
which
covers
java
instrumentation
project
and
then
also
chat
about
the
java
sdk
project,
which
is
on
friday
mornings,
so
yeah,
so
john,
primarily
involved
in
the
java
sdk
project.
C
So
I'll
start
and
see
what
see
if
the
morning
sig
notes
remind
me
of
what
we
chatted
about.
I
was
not
very
organized
this
morning,
but
luckily
we
found
stuff
to
chat
about
people
had
stuff
to
chat
about.
C
So
the
the
sad
news
was
that
we
closed
zero,
p1
issues,
I'll
wait
for
honorag,
because
he'll
be
sad.
No,
it
happens.
We
closed
zero
p1
issues.
Last
week
we
got
two
new
p1
issues,
so.
B
C
B
Build
it
themselves.
So
again,
I
don't
know
whether
it's
something
we
want
to
think
about
supporting
as
an
option,
but
I
don't
know
how
you
mean
you
have
to
you
have
to
build
your
grpc
messages
to
include
your
teddy's
vanity,
and
then
you
have
to
manually
figure
out
how
to
stitch
things
together,
because
you
don't
have
the
at
least
the
way
that
the
the
grpc
instrumentation
that
we
have
today
is
written.
Doesn't
it
just
doesn't
work
the
same
way
so
did.
C
B
A
I
have
a
hardware
mic
mute
button.
It
didn't
work
so
yeah.
I
met
super
duper
senior
senior
engineers
who
have
this
principle
like
everything
has
to
be
in
the
proto.
That's
just
how
it
is,
and
that
might
be
the
case
here.
I
don't
know,
but
I
don't
think
that's
a
good
reason,
but
I
mean
at
the
same
time,
manual
propagation
should
work
right.
So
I
haven't
looked
at
this
issue
since
they
sent
the
repo,
but
I'm
going
to
take
a
look
also,
even
because
they
were
doing
everything
by
hand
right
well.
B
B
A
A
B
Oh
yeah,
the
other
thing
we
talked
about
was
planning
for
the
0.8.0
sdk
api
release,
which
should
happen
early
next
week.
Is
the
plan
we're
not
going
to
include
the
span
context,
rework
for
strings
that'll
go
into
0.9.0
just
because
it's
a
little
short
notice
and
it
will
be
a
large
breaking
change,
I'm
pretty
sure
for
the
the
auto
instrumentation,
maybe
maybe
not,
but
but
nikita
was
like
no.
No!
No
don't
do
it
yet.
Wait
until
wait.
Give
us
some
time.
C
B
Yeah,
which
I'm
totally
fine,
we
can
wait
until
tuesday
or
whatever
to
merge
that,
assuming
that
honorable
is
satisfied
with
my
updates
and
the
and
so
the
plan
then
for
o.9.0
is
to
really
have
the
the
tracing
api
nailed
down,
and
I
would
say
I
would
say,
release
candidate
level,
the
only
the
only
hitch
in
that
plan-
and
all
of
this
is
basically
the
agenda
for
tomorrow
morning's
meeting.
B
C
A
A
B
A
Just
yesterday,
like
that
issue,
I
filed
the
trust
looked
at
actually,
the
root
like
I
was
playing
with
things,
but
the
first
bug
I
ran
into
was
a
verifier
cannot
assign
thread
local
storage
to
grpc
contact,
storage
or
something
like
that
and
that
started
a
rabbit
hole.
So
I'm
imagining
grpc
context
is
biting
me
again.
Yeah.
C
B
Yeah,
so
the
only
other,
I
think,
there's
there's
five
open
p1
issues
for
ga.
One
of
them
is
grpc
context.
One
of
them
is
implementing
baggage.
One
of
them
is
just
the
honor
of
your
issue
on
too
many
wrapper
objects
which
the
span
context
fix,
plus
potentially
attribute
value.
I
put
in
a
little
draft
pr
to
demonstrate
my
proposal
with
the
type
keys
today,
so
it
would
be
great
if
people
could
take
a
look
at
that.
B
A
B
A
B
But
my
guess
is
it's
not
a
release
process
thing.
It's
probably
something
about
the
way
that
bin
tray
is
publishing
to
maven
central,
that's,
not
working
or
something.
I
don't.
That's.
My
guess
is
it's
not
actually
our
publishing
like
where
I
think
people
have
seen
that
it
shows
up
properly
in
bin
tray,
but
it
doesn't
show
up
in
david
central.
So
it's
that
is
the
linkage.
That's
probably
the
problem
anyway.
I
I
am
not.
B
A
A
A
C
B
So
anyway,
I
think
those
are
the
five
open
issues
yeah.
This
is,
I
don't
know
if
we
want
to
talk
about
this
now,
but
this
little
prototype
is
rough
draft
of
what
the
kind
of
I
did
a
really
simple
implementation
of
only
three
key
types,
just
to
demonstrate
kind
of
how
the
api
would
work
and
what
it
would
look
like
to
define
symmetric
attributes
and
what
it
would
like
look
like
to
use
them
to
generate
some
attributes,
you
know
and
to
consume
them,
because
that
was
the
part.
B
I'd
never
implemented
the
consumption
part
before,
but
luckily
all
of
that
work
we
did
on
getting
rid
of
the
attributes.
I
I
knew
what
options
were
and
I
implemented
the
one
we
kind
of
all
decided
we
liked
so.
A
B
C
Oh,
I
see
string
array
gotcha
I
mean
you
could
do
consume
string
array
consume
boolean
array.
B
Yeah,
I
mean,
I
guess,
that's
the
what
either
one
would
be
easy.
You
know
change
between
them.
A
C
A
B
I
think
it
makes
a
nicer
like
the
api,
seems
a
little
well
as
I
as
mentioned.
It's
slightly
fewer
allocations.
One
fewer
allocation
in
the
string
case
same
number
of
allocations
in
the
other
cases,
but
I
think
it
makes
a
nicer
api
and
I
think
that
makes
the
semantic
attributes
you
don't
have
to
have
like
a
separate
set
of
symmetric
attributes
just
to
find
your
keys
and
you're
done.
C
Yeah,
I
really
like
the
tie-in
to
the
semantic
attributes.
One
thing
I
guess
there
was
one
other
thing
that
I
was
thinking
about
with
it,
which
was
like
you
mentioned.
The
the
rappers
like
the
I
mean
boolean
is
fine,
but
more,
and
the
reason
I
mentioned
is
because
there's
a
data
dog.
C
Pr
coming
at
some
point
soon
did
it
just
happen:
yeah
they
they
optimized
away
the
integer
boxed
integers,
and
for
because,
like
ports
and
http
status,
codes
are
outside
of
the
the
def,
like
the
I
think,
by
default,
java
will
store
like
minus
128
to
128,
or
something
like
that.
You
don't
pay
any
you're
very
little
cost
for
boxing
those.
C
B
Never
heard
of
that
data
structure
before
so
interesting
yeah.
I
was
pondering
with
all
this.
Like
the
storage
on
my
little
prototype
is
not
supposed
to
be
the
final
storage.
It
was
just
something
super
simple
that
I've
I
cranked
out
just
as
to
demonstrate
I'll
get
storing
in
a
map
right
now,
so
there's
a
whole
bunch
of
extra
allocations
with
map
with
map
entries,
but
I
was
pondering
that
we
could
potentially
have
different
storage
for
each
of
the
types.
B
Although
ordering
is
a
little
ordering
would
be
weird,
and
I
think
there's
also,
I
don't
even
know
how
I
haven't.
I
just
realized
there's
a
weird
issue
like
if
you
use
the
same
string
key
on
two
different
types,
I'm
not
sure
like
if
you
use
the
same
like
http
status
and
you
had
both
a
string
version
and
a
long
version
of
it
like
you
have
to
figure
out
how
to
deal
with
eliminating
one
or
I
don't
know
that
just
realized,
that's
a
little
weird
corner
that
would
need
to
be
figured
out.
B
A
B
But
I'm
more
concerned
actually
with
the
apis,
like
I'm
less
concerned
with
the
storage.
With
this
implementation,
we
can
figure
that
out.
We
can
optimize
the
credit
out
of
that
as
we
need
to
I'm
much
more
concerned
with
the
api
from
the
user
perspective
like
the
actual
instrumentation
perspective,
and
then
the
consumption
perspective
for
exporters,
etcetera
and
what's.
B
C
From
a
user
perspective,
I
like
it
because
you
can
chain
the
set
attribute
cause
with
the
and
use
the
semantic
attributes.
B
Right
exactly
that's
the
that
was
the
main
point
was
to
that
was
one
of
the
main
points
was
to
make
sure
you
didn't
have
to
have
both
keys,
like
keys,
defined
as
strings
and
separately
a
symmetric
attribute
value
or
whatever
those
things
are
called
that
you
can
just
do
it
all
in
one
place,
one
set
nice
and
easy.
B
It
is
it
it
is
in
the
sdk
yeah
okay,
I
just
put
it.
I
just
put
this
all
in
one
nice
little
compact
place
to
play
around
with,
but
yeah
that'll.
That
will
be
something
that's
in
the
sdk
for
sure.
B
Yeah
events
and
links
and
yeah-
and
this
would
that
keyed
attributes
would
replace
the
existing
attributes.
I
just
needed
a
unique
name
for
it,
so
it
would
become
right.
It
replaced
the
existing
attribution
class.
C
Right
so
I'm
just
thinking
from
from
a
bridge
instrumentation
bridge
perspective
for
the
span
builder
and
span
set
attributes.
That's
easy,
because
we
just
push
those
straight
over
to
our
shaded
keyed
attributes.
We
would
the
keys.
We
can
keep
internal
copy
in
there
of
our
shaded
key,
since
those
are
only
constructed
rarely
and
then
I'm
trying
to
think
for
events
and
links
where
the
user
passes
us.
The
whole
keyed
attribute.
B
B
B
C
So
if
it
was
an
abstract
class
and
then-
and
you
had
your
own
subclass,
then
with
a
factory
method
in
the
the
base
class,
then
we
could
hijack
the
factory
method
return.
C
C
C
Yeah,
I
was
okay
with
the
inner
nested
inner
class,
but
not
being
able
to
ever
add
a
new
method
to
an
interface
is
seems
like
a
deal
breaker.
B
Yeah
not
having
default,
although
if
we
can
go
to
java
8
post
ga
as
donald
trump
used
to
be
okay
with,
although
maybe
not
in
the
api,
though
not
clear
about
api
post
release,
but
that
would
be
that
would
be
super
nice
like
right.
I
mean
that
so
it
seems
like
such
a
small
little
feature,
but
those
defaults.
B
C
A
C
A
A
So
that
is
one
option.
If
we
stick
with
interfaces
yeah
I
mean
that
sense
aiming
for
that.
Rather
than
optimizing,
everything
for
java
7
forever
is
probably
a
better
idea,
so
we
could
use
interfaces
as
long
as
we
know
that
any
type
when
we
add
a
method,
that's
when
we
also
update
to
java
8
baseline.
B
C
Those
nested
factory
classes,
we
could
just
deprecate
them
and
add
the
factory
methods,
new
factory
methods.
A
A
B
Anyway,
we
don't
need
to
talk
about
this
anymore.
I
don't
want
to
take
up
the
instrumentation
time.
Okay,.
C
No,
no,
there
was
some.
What
was
I
thinking
default
factory?
Oh,
I
know
what
I
was
gonna
ask
do
you
know?
If
so,
if
we
started
with
abstract
and
factory
methods,
can
we
convert
that
to
an
interface?
Is
that
binary
compatible.
B
Unless
somebody
has
their
own
custom
extension
of
it
already.
A
A
A
A
C
A
C
A
C
All
right
so
john
you're,
you
are
making
great
progress
here,
really
psyched
about
the
span
context
and
the
attribute
stuff.
That's
going
to
make
our
bridge
much
more
efficient,
so
yeah
just
keep
pinging
me
and
I'll
try
to
well
the
span
context.
I
could
probably
is
this
in
a
good
enough
shape
right
now
that
I
could
try
it
out.
B
C
I'll
go
ahead
and
try
it
out
in
the
instrumentation
and
yeah:
that's
the
only
way,
I'll
I'll
be
able
to
provide
super
good
feedback,
except
for
that
I
mean
I,
like
the.
I
like
the
idea
of
it.
A
My
flags
are
always
going
so
if
w3c
was
to
release
a
new
version,
what
happens
to
the
trace
span
context?
We
just
add
more
methods
or
something
like
that.
Probably
right.
A
B
B
C
B
C
B
Even
mention
flags
here
doesn't
well,
the
trace
date
would
be
the
other
one
right
trace
date
is
the
one
that's
old,
maybe
a
little
weirder.
Just
because
you
know
the
this
tray
stays
a
fairly
complex
beast.
B
B
Yeah
well
before
it
used
to
have
entries
and
the
entry
there
was
an
entry
key
and
an
entry
value
yeah
inside
them,
and
I
at
least
stripped
that
stuff
strip
the
key
out.
So
it's
just
strings,
although
I
mean
yeah,
we
should
revert
that
we
might
need.
I
think
it
was
like
lazy
keys
or
something
I
don't.
B
A
C
So
if
entry
is
just
a
key
value,
can
we
use
map
dot
entry.
B
That's
a
very
good
question.
I
don't
know
what
I
haven't
looked
at
the
internals
that
map
that
entry
for
a
long
time,
but
I
assume
it's.
C
The
only
reason
I
ask
is
because
if
it's
a
java
object,
we
don't
have
to
bridge
it.
B
Yeah
I
hear
you
I'm
not
opposed,
I'm
I'm.
I
would
be
totally.
B
What
about
trade
well?
Do
we
want
to
also
think
about
trey
state?
Oh
there
we
are.
This:
is
trade
state,
so
there's
two
issues,
there's
kind
of
the
trace
flag.
Can
we
just
push
that
into
the
spain
context
and
not
have
to
keep
the
whole
extra
wrapper
around
a
single
byte.
B
A
B
B
B
No
you
actually,
if
you,
if
you
exceed
the
total
number
of
valid
entries,
you
have
to
start
dropping
before
propagating
and
that's
and
you
have
to
drop
the
oldest,
and
so
there
is
an
ordering.
There
is
a
explicit
ordering
that
needs
to
be
preserved.
A
And
and
linked
hash
map
is
too
inefficient
or
something
yeah
I
mean
you
could
probably
use
linkedin.
A
C
A
A
C
We
may
be
able
to
I'm
going
to
look
at
that
span
context.
Api
span
context
change,
because
if
we
back
the
span
context
by
if
we
just
do
a
delegate
to
our
internal
spam
context,
oh,
but
you
already
have
the.
C
B
C
Yeah
the
key
it
tends
to
be
like
like
with
trace
id
if
we
need
to
pass
it
to
somewhere,
but
we
could
pass
the
list
of
entries
right.
C
Zoe,
when
did
you
start
at
google,
or
are
you
just
or
is
this
a
new
project
at
google
that
you're
starting
on.
D
No,
I
actually
started
so
this
is
my
fourth
week
but
yeah.
Officially,
I
think
yeah
officially
my
fourth
week
and
the
but
yeah
for
the
first
week.
I
didn't
get
access
to
a
lot
of.
C
And
which,
what
kind,
what
team
do
you
work
on
there.
D
Yeah,
and
also
niall
so
niall,
what's
his
last
name,
islam.
D
C
Any
well,
we
had
some
folks.
We
had
some
interns
this
summer.
C
Spawn
yeah
there
were
a
few
interns
munir
who
has
been
very
active
in
the
instrumentation
project,
and
I
forgot
there
were
a
couple
of
met.
There
were
a
couple
of
mentors
google
employees
who
were
mentoring
them.
C
Do
you
happen
to
have
any
in
your
four
weeks
there
have
you
met.
A
C
D
C
I've
been
pinging,
morgan
and
some
other
folks,
yeah
cool.
Have
you
had
a
chance
to
look
at
any
of
the
open
telemetry
stuff?
Yet?
Do
you
have
any
questions
for
us
that
we
could
help
with
your
onboarding.
D
I've
looked
at
the
the
google
cloud
trace
side
of
it,
I
suppose
the
exporter
and
yes,
I've
started
looking
into
currently
I'm
working
on
the
adding
a
metric,
I'm
adding
metric
to
the
cloud.
Google
cloud
monitoring,
operation,
exporter
so
yeah,
so
I
haven't
looked
too
much
into
the
the
you
know:
the
open,
telemetry,
sdk
or
yeah
the
I
suppose,
non-google
side
of
things
yet
yeah.
C
The
metric,
the
metric
exporter,
you
were
doing
open,
telemetry
metric
exporter
to
google
cloud.
D
Yeah
to
google
cloud
monitoring
and
yeah.
I've
just
started
working
on
that
now.
Looking
into
that
cool.
C
Well,
john
knows
all
about
the
metrics
stuff,
so
any
metric
questions.
C
B
I
don't
know
if
you
saw
any
of
the
josh's
metric
sdk
specification.
I
shouldn't
be
too
snarky.
B
That,
basically,
is
hey.
You
want
to
implement
metrics
sdk,
do
it
the
way
go
did
and
that's
you're
kind
of
left
with
that.
So
I
think
it
is
going
to
mean
unless
we
can
get
that
spec
rewritten
in
a
way,
that's
more
flexible,
I'm
going
to
have
to
rewrite
the
metrics
sdk
for
java,
which
I
am
not
excited
about.
C
B
Yeah,
so
so
in
the
go
and
the
the
c
plus,
because
they
copy
go
implementations,
they
have
so
at
the
so.
The
the
sdk
implements
the
api
and
the
sdk
itself
has
a
component
called
the
accumulator
that
manages
all
of
the
rook
metric
recordings.
So
it's
one
component
that
measures
manages
all
of
the
metric
recordings.
B
So
basically
it's
all
funneled
funneled
down
into
this
one
class,
one
instance
that
handles
all
metric
recordings
and
then
sends
it
like
two
aggregators
as
appropriate,
but
there's
one
central
thing
in
the
way
that
the
java
with
bogdan-
and
I
built
the
java
one-
is
each
instrument.
Implementation
has
its
own
accumulator
that
accumulates
the
data,
as
recordings
are
made
on
it.
B
So
we've
done
it
much
more
object-oriented
for
good
or
for
ill,
and
the
go
has
been
basically
just
there's
one
big
mo
like
a
monolithic
component
that
that
records
all
of
the
all
of
the
data,
so
it
would
be.
Does.
C
That
affect
have
any
effect,
for
implication
for
users.
B
It
should
not,
I
mean
this
is
kind
of,
so
what
I
want
is
I
want
a
specification
that
explains
here's
the
functionality
that
that
an
sdk
needs
to
provide
it
needs
to
be
able
to
have
plugable
exporters
needs
to
be
able
to
have
plugable
aggregators.
It
needs
to
be
able
to.
You
know,
specify
for
these
things,
what
the
behavior
is
going
to
be
with
this
particular
instrument.
It
needs
to
be
able
to
handle
timed
or
pull-based
exports.
B
B
It
must
have
this
component,
it
must
have
this
component
and
it
must
have
this
component
and
they
must
be
wired
together
this
way,
so
I'm
much
more
interested
in
a
functional
specification-
and
it
sounds
like
josh
and
cjo
are
much
more
interested
in
like
here-
are
the
actual
components
that
you
must
implement
as
a
sdk
kind
of
the
way
that
the
tracing
sdk,
like
you
have
span,
processors,
and
you
have
exporters
and
you've
got
like
all
of
that
is
very
well
defined
and
all
of
the
languages
kind
of
conform
to
it.
B
But
since
we
haven't
had
that
for
metrics
at
all,
each
of
the
metric
transformations
has
gone
its
own
way.
Well,
bogdan.
Did
it
one
way
and
josh
did
it
another
way,
and
then
everybody
copied
josh's,
because
they,
because
it
was
what
was
in
go
and
so
login's
implementation
doesn't
match
what
josh
or
cjo,
because
he
copied
what
josh
had
done
want,
and
I
don't
want
to
have
to
rewrite.
I
don't
care
that
much.
I
just
don't
want
to
have
to
rewrite
the
entire
metrics.
That's
the
game
or
ga.
B
Because
if
we
have
to
rewrite
the
entire
metrics
sdk
before
ga,
I
am
not
going
to
be
able
to
deliver
that
by
november.
I
don't.
A
B
B
C
Yeah
zoe,
do
you
know
what
kind
of
metrics
for
google
cloud
google
cloud
that
are
going
to
be
collected
or
like
what?
Because
like
on
the
instrumentation
side,
we
haven't,
we
haven't
wired
up
any
instrumentation
yet
to
the
open,
telemetry
metrics
api.
D
I'm
not
completely
sure
sorry,
yes,
I'm
still
kind
of
looking
into
what
I'm
yeah.
I
guess
still
get
gathering
context.
Okay,
yeah.
C
D
I've
just
been
looking
at
the
python
one,
the
metric
implementation
for
python
for
the
cloud
monitoring
connection,
so
yeah
and
kind
of
trying
to
base
it
off
that
a
little
bit
yeah.
C
C
Yeah
well
yeah.
Let
us
know,
I
know
we're
there's.
Oh
john
somebody
from
new
relic,
graham
pinged
me
about
he
was
gonna
start
working
on
generating
metrics.
I
think
from
requests
or
something
I.
B
Think
http,
I
think
he
was
going
to
try
to
tackle
http
request
and
server
metrics
I'll,
probably
start
with
one
of
the
other,
but
because
he's
been
he
is
authored.
The
spec
that
is
maybe
very
close
to
being
merged
for
what
the
semantic
inventions
are
for
http
metrics.
B
C
Oh,
I
know
he
said
he
was
going
to
be
out
today
and
invited
him
to
this
meeting
to
chad,
because
I
had
suggested
when
you
say
couldn't
make
this
me.
I
suggested
opening
an
issue
to
get.
It
kicked
off.
C
B
C
Yes,
alex
I'll
tell
him
that
john
assigned
it
to
you
yeah.
C
A
C
A
A
A
And
I
I
mean
it
must
be
something
related
to
my
dependencies
or
shading
or
something
like
when
you
mention
the
shading
thing.
That
was
because
first,
I
suspected,
because
I
was
using
an
implementation
dependency
on
one
of
these
modules,
the
aws
support
module.
That
might
be
a
problem,
so
I
moved
it
to
be
a
runtime
dependency
only
in
the
agent
rather
than
sort
of
this
transitive
thing.
A
I
didn't
move
the
shading,
though
so,
once
I
moved
the
shading,
I
got
back
to
this
same
error
so
that
didn't
do
anything
and
yeah
the
other
thing,
I'm
looking
through
the
content
of
the
zip
file
and
yeah.
I
mean
this
looks
very
suspicious
in
that.
For
some
reason,
this
thread
to
local
context,
storage
class
is
all
by
itself
here
like
it's.
Not
this
doesn't
seem
to
be
the
whole
grpc
context
package
for
some.
C
A
A
C
A
C
A
This
I
can
zoom,
I
think
so.
I
found
the
shaded
version
and
this
looks
correct
and
it's
also
as
you'd
expect
thread.
Local
context,
storage
and
context.
Storage
are
in
the
same
place,
but
this
error.
Sorry,
I
showed
you
on
the
stack
trace,
but
this
error
is
happening
in
the
exporter,
so
this
is
the
exporter
class
loader.
A
A
C
Question
I
would
assume
the
safest
thing
is
for
it
to
use.
I
actually
think
I
don't
think
it
matters,
but
right
so
this.
A
C
A
C
It
matter,
but.
C
To
have
the
context
in
the
needs
to
be
under
the
inst
directory
which,
where
it's
not
do
you
want
to
pop
open
the
official
like
zero,
seven,
zero
build
and
see.
A
C
A
A
A
A
A
C
A
Yeah
and
it's
there
like,
so
I'm
assuming
it's
some
sort
of
dependency,
how
like
I'm
getting
the
wrong
version,
because
we
also
have
context
storage
here,
but
we're.
A
A
C
Yeah,
the
type
is
not
assignable,
so
I'm
wondering,
if
thread
local
context,
storage
there.
Oh
no,
but
why
would
it
have?
C
Maybe
that's
I'm
just
gonna
check
the
shading
here.
D
D
B
Great
to
meet
you
zoey
hope
see
you
around
in
the
project
more.
A
A
C
C
A
You
did
okay,
so
it
might
be
related
to
that
change.
We
did
recently
changing
the
shading
strategy
a
little
bit.
Let's
see
so
grpc
con
yeah
yeah.
I
don't.
A
A
C
A
A
C
Nice
same
smoke,
test
harness.
A
A
A
A
A
Generation
hardcoded
in
the
runtime
and
yeah
we're
we
might
be
sort
of
blocked
yeah.
They
said
they
don't
want
to
add
anything
to
the
run
time,
but
we
can't
find
any
way
around
it
outside
the
run
time,
so
we're
going
to
have
to
push
harder
because
otherwise,
x-rays
just
not
compatible
basically-
and
that's
probably
not
the
intent.
So
we'll
see
the
zip
code.
A
A
Yeah,
exactly
if
open
tell
me,
you
could
generate
trace
ids
with
that
bucket,
just
by
default
copy
pasting
that
did
not
help
so
yeah.
I'm
just
gonna
have
to
keep
on
digging.
A
A
A
A
C
A
Yeah
right
now,
like
this
provider,
factory,
was
much
more
complicated
before
now.
It's
only
using
these
fours
yeah.
Technically,
I
don't
need
much.
C
All
right
ping
me
on
slack,
yep
cool
anything
else,
any
other
questions
you
have
zoe
about
the
project.
D
C
Yeah,
so
this
usually
it's
a
recap
of
the
morning
the
meeting
that
happened
this
morning,
pacific
time
and
then
also
john,
so
that
the
thursday
morning
meeting
is
instrumentation
related
and
then
john
started
joining
also
this
to
chat
with
honorag,
so
that
we
could,
since
honorable
can't
make
the
friday
morning
meeting
either
so
we've
kind
of
broadened
the
scope
to
be
just
anything
java,
open,
telemetry,
java,
related
yeah,
and
sometimes
we
invite-
or
I
mean
everybody's
always
invited.
C
D
So
pretty
much
like
kind
of
debugging
or
problems
or
whatever
that's
related
to
open
telemetry
java
talk
about
it
here
for
the
I
guess,
the
apac
people.
C
Yeah
or
even
non
apac
folks,
I
invited
I
invite
people
here
and
to
join
this
one,
just
as
a
venue
for
troubleshooting
and
then
have
you
seen
the
public
calendar
and
the
the
meeting
notes
from
so
there's
meeting
notes
from
the
thursday
from
both
instrumentation
and
the
java
sdk
meetings.
C
Meeting
notes,
let
me
put
in
chat
here
in
the
community
repo
you'll
find
links
to
the
meeting
notes,
so
those
can
be
useful,
but
also,
if
you
have
something
that
you
want
to
since
those
times
probably
aren't
going
to
work
for
you.
C
This
is
a
good
venue,
for
you
know
escalating
any
thing
that
you
are
not.
You
know
any
questions
that
you
have
that
aren't
getting
addressed
like
in
the.
D
A
C
Typically
and
my
headset
might
die
any
second,
so
then
it'll
take
me
a
second
to
switch
over
the
on
your
wednesday
midday.
I
guess
your
wednesday
midday
my
tuesday
night
on
nikita's
wednesday
morning.
C
C
Yes,
yes,
so
that
was
on,
I
haven't
turned
that
into
a
recurring
one
yet,
but
I
was
going
to
plan
on
turning
it
into
a
recurring
meeting,
and
so,
if
you
want
to,
you
know
always
join
that
you
can
join
that
anytime.
And
if
you
want
to
have
questions,
we
can
always
cover
those
first
and
then
we
can
let
you
know
sort
of
the
meeting
topics
and
you
can
decide
if
there's
no,
no,
no,
we
won't
be
hurt.
If
you
say
that's
not
terribly
interesting,.
C
Sometimes
we
yeah
the
instrumentation
stuff
can
be
very
different
than
the
sdk
stuff,
but
it's
great
when
I'm
like
this
week
was
great
awesome
because
john
joined,
so
we
had
a
good
sort
of
melding
of
both
sdk
topics
and
instrumentation
topics.
Yeah.
D
Okay,
and
is
there
any
good
resources
you
would
recommend
on
kind
of
onboarding
myself
into
all
of
us,
or
is
it
basically
the
just
the
open
telemetry
website
and
the?
I
guess,
the
github
pretty
neat
one.
C
Really
read
too
much,
oh
no,
not
for
reading,
but
for
asking
questions.
D
C
And
for
onboarding,
what
I
would
suggest
is
asking
john
for
a
like
a
couple
of
issues
like
an
issue
recommendation
or
honorable
honorable
spends.
I
think
enough
time
in
the
sdk
repo
to
maybe
have
some
ideas.
A
I
mean
there's
some
help
wanted
issues.
Definitely
if
you
can
take
a
crack
on
any
of
them,
I'm
sure
they'd
be
really
helpful.
C
Yeah,
it
may
be
worth
asking,
I
don't
know
honor
would
that
be
like
to
post
to
get
her
or
post
an
issue
or
just
really
do
you
think
the
help
wanted
ones
are
in
good
enough
shape
to
just
because
I
know,
for
example,
when
I
was
when
I
was
onboarding
helen
into
the
instrumentation
repo,
you
know
I
I
kind
of
picked
very
specific
issues.
That
sort
of
you
know.
A
C
A
A
C
Get
some
good
sort
of
filter
some
a
cup,
an
issue
in
order
to
or
from
people
who
know
these
issues?
Who
can
predict
that
they're
not
going
to
take
you
down
some
rabbit
hole
but
give
you
you
know
a
chance
to
you
know:
try
out
some
stuff,
learn
the
build
environment.
Learn
you
know
all
that
stuff
and
then
kind
of
continue
along
there,
and
you
know
it
it's
a
win-win.
You
know,
for
I'm
sure
john.
C
Would
you
know
for
the
maintainers
also
to
to
kind
of
find
good
onboarding
issues
for
somebody
who
you
know
is
going
to
especially
going
to
be
part
of
the
project.
A
So
let
me
so
I
I
found
a
couple
issues.
One
of
them
is
assigned
to
nikita.
I
don't
think
he
started
so
I'll.
Ask
him
about
it
and
then
probably
assign
it
to
you
and
it's
a
good
starting
issue
to
be
able
to
disable
resource
providers
right
now
we
take
resource
information
from
anything
on
the
class
path,
but
users
might
want
to
disable
it.
So
it's
adding
a
flag
to
be
able
to
do
that.
I
think
that.
D
D
D
C
All
right,
well,
hey!
It
was
great
to
meet
you
zoe
yeah.
I
look
forward
to
seeing
you
around
in
the
project
honorable
pleasure
always
and
let
me
know
how
the
shading
goes.
A
C
All
right,
yeah
have
a
good
have
a
good
friday.
Both
of
you.