►
From YouTube: 2022-12-01 meeting
Description
Open cncf-opentelemetry-meeting-3@cncf.io's Personal Meeting Room
A
A
B
B
C
B
So
we
can
be
a
pair
of
minutes
more
someone
added.
B
One
of
the
the
second
topic
was
added
me:
I.
D
B
It
yeah
okay,
okay,
great
yeah,
I
I
saw
the
document
today
and
I
already
saw
a
topic
there.
So
yeah,
that's
great
yeah.
D
Yeah
but
but
I
think
we
we
can
start
with
the
review
of
last
meeting
because
I
think
Bryce
wasn't
there
so.
A
D
Actually,
we
can
start
with
the
easier
one.
The
last
one
updates
to
the
event
specification.
B
D
So
I
forgot
did
I
write
the
details.
There
were
two
changes
there,
okay,
so
this
even
specification
is
actually
actively
evolving.
There
is
a.
There
are
a
lot
of
opinions
on
that.
You
know
we're
doing
active.
You
know
discussion,
it
is
you
know,
good
thing
is
I
think
it
is
only
incremental
there
is.
There
is
one
you
know,
kind
of
a
breaking
change.
You
know
that
was
made,
but
other
discussions
are
mostly
you
know
incremental.
So
we
might
not
be
that
bad.
D
The
domain,
the
event
so
every
event
has
two
attributes
set
event
or
domain
and
even
dot
name
and
the
event.domain
used
to
be
at
the
scope
level.
It's
been
moved
to
the
log
record
level,
it's
been
yeah,
it's
moved
to
the
log
records
attributes
now,
along
with
the
you
know,
name,
so
they
are
both
together.
D
Domain
is
not
part
of
the
log
record,
yeah
correct
and
the
second
change
is
the
the
logging
API
was
was
never
had
a
strong.
You
know,
support
people
always
wanted
to.
You
know
avoid
having
a
login
API,
because
all
languages
have
already
logging
apis
and
a
flocking
API,
so
they
didn't
want
to
introduce
yet
another.
D
But
since
we
wanted
an
event
API,
you
know
that
kind
of
forced
locking
API
to
be
there
too,
but
the
latest
development
is
that
the
you
know,
since
the
event
specification
is
still
evolving.
The
events
is
taken
out
and
built
as
a
wrapper,
so
there
will
be
a
logger
API.
The
ribbon
events,
API
and
events
API
uses
logger
API.
D
So
this
is
yeah
events.
Api
will
wrap
the
logger
API
correct
log
API,
okay,
those
are
the
two
changes
and
the
things
that
are
being
discussed
currently
I.
Think
that
also
I
think
you
know
a
good
write
in
the
note,
so
one
is
I
think
the
main
thing
is
there
is
a
parallel
specification
called
Cloud
events
that
we
came
across
recently
and
it
is
also
a
cncf
project
and
apparently
it's
already
very
popular
among
Cloud
providers,
and
it
is
from
what
I
understand
so
far.
D
It
is
used
to
exchange
events
across
Cloud
providers,
especially
for
serverless
handling.
D
D
What
is
being
you
know
currently
under
discussion,
so
it
won't
change
much
as
far
as
I
know,
I
think,
a
few
things
that
they
have
is
they
have
separated
the
actual
event
payload
from
the
from
what
they
call
as
context
attributes
context.
Attributes
are
those
that
enable
routing
and
high
level
decisions.
Let's
say
you
know
a
pointer
to
the
schema
for
the
payload
and
some
name
and
domain
like
equivalence,
you
know
that
can
be
used
for
routing
events
to
you
know
appropriate
pipelines
on
the
back
end.
D
So
we
might
adopt
you
know
their
terminology
on.
You
know
their
structure
but
end
of
the
day.
You
know
if
we
were
to
build
events
on
top
of
log
records.
We
can
only
do
so
much,
which
is
put
everything
only
within
attributes,
because
the
high
level
Fields
they're
very
you
know
log
Centric,
application,
log
Centric.
So
that's
what
is
is
going
on
there.
C
D
To
the
I
think
that
is
up
to
us,
okay,
I
think
the
the
only
requirement
is
that
you
register
it
so
that
you
know
if
somebody
else
comes
there
after
us,
you
know
they
will
have
to
use
something
else.
So
it's
an
enum,
so
we
register
something
there
and
the
reason
I
think
device
was
okay
was
because
it's
already
used
as
the
namespace,
for
you
know,
device
level
attributes
for
iOS
for
the
mobile
apps,
so
devices,
although
are
not
uncommon.
On
the
server
side.
D
D
So
so
iOS
mobile
I
mean
iOS
phone
and
Android
phones
and
tablets
and
any
potentially
any
other.
You
know
things
in
in
consumer
homes.
Like
you
know
all
your
iot
devices
Alexa
or
you
know
television
sets
anything.
That's
not
on
the
cloud.
That's
not
on
the
you
know,
Enterprises,
let's
say
at
the
end
user
places.
You
know
those
I
think
will
fall
under
the
rum
category,
so
I
think
those
we
can
safely
use
the
term
device.
Okay,.
D
So
I
noticed
you
have
a
PR
for
the
logging
API
login
events
API,
so
that
one
change
you
could
make,
which
is
first
build.
The
logging
API
take
out
just
the
events,
part
of
it,
okay
and
and
then
have
another
separate
layer
called
events
API,
which
takes
an
instance
of
the
logger
I,
think
I
think
they're.
The
the
specification
you
know
tells
you
how
to
build
the
interfaces.
A
A
D
Yeah,
so
that's
that
and
I
think
last
week
we
also
went
over
a
few
resource
attributes.
A
D
And
we.
B
Could
I
I
shared
this
with
the
ram
team
about
your
document
and
the
and
yeah
in
the
meeting
I
invited
them
to
this
meeting,
also
okay,
to
come
to
this
meeting
if
they
want,
but
I
I
didn't
receive
feedback
on
that,
so
so
I,
don't
think
I
don't
know
if
they
actually
added
something
to
the
document
I
but
I.
Don't
they
didn't
say
me
anything
about
that.
So,
probably
now.
A
D
But
for
the
purpose
of
like,
as
in
since
price,
wasn't
there
I
I
want
to
bring
up?
One
document
can
I
share
my
screen
for
a
second
yeah.
E
D
Share
my
screen
share.
Okay,
so
the
you
know
there
are
a
couple
of
you
know,
ideas,
I,
think
one
is.
You
know
we
wanted
to
kind
of
list
down
the
attributes
for
each
type
of
event.
Like
you
know,
in
the
case
of
resource,
you
know
these
are
the
expected
attributes
in
the
case
of
Network
request
span.
You
know
these
are
the
attributes
in
the
case
of
you,
know
a
network
change
like
a
Wi-Fi,
two
cellular.
D
You
know
what
parameters
to
expect
I
think
today,
Splunk
is
sending
you
know
these
things.
Similarly,
an
error
and
a
crash
report.
We
listed
a
few
things
and
the
the
idea
is,
you
know
we.
We
want
to
make
this
a
standard
between
both
Android
and
iOS
and
SPL
and
then
put
it
on
the
you
know
in
in
let's
say,
I
think
in
the
documentation.
So
you
know
on
on
this
website
in
the
docs.
You
know
along
the
lines
of
what
Splunk
has
so
the
Splunk.
D
You
know
this
documentation
page,
so
they
have
listed
for
each
type
of
event.
What
attributes
to
expect.
D
This
will
enable
the
receivers
the
back
ends
to
to
build
any
validations.
You
know,
since
most
of
the
ram
data
is
going
to
come
unauthenticated,
so
anybody
can
post
data.
We
don't
want
people
to
post
garbage,
so
we
have
seen
that
in
our
case
I
mean
it
won't
prevent.
D
You
know
people
from
sending,
you
know
valid
data,
but
that's
not
genuine
right.
It
like
they
can.
They
can
still
fake
it,
but
at
least
we
don't
want
any
garbage.
We
wanted
to
catch
errors.
You
know
there's
another
another
where
to
say
it.
So
it's
good
to
list
this
down
both
initially
as
a
human
readable,
but
ultimately
as
a
machine
readable,
so
that
you
know
it
can
be
used
for
validation,
so
yeah
I
think
if
you
folks
can
also
review
these
documents.
I
mean
these
attributes
at
your
convenience.
C
D
See,
okay,
no
I
think
you
can
just
add
a
you
know,
comment
or
maybe
easier
might
be
just
add
to
the
notes,
and
then
we
can
talk
next
time
and
I
I
blindly
only
picked
up
from
from
what
these
guys
had.
I
I
haven't
discussed
internally
at
in
in
our
company
as
well.
What.
A
D
D
D
Yeah
yeah
I
think
I
have
spoken
to
sirath
I,
think
she's,
the
PM,
but
I
have
never.
I
have
spoken
to
her
only
about
Android
I,
don't
know
if
they
are
also
using
this
open,
Telemetry
iOS
or
the
open
telemetric
Swift
repo
foreign.
B
But
the
network
instrumentation
they
are
using
is
different
from
the
public
accessible
one.
They
implemented,
they,
they
don't
okay,
but
on
the
basic
they
are
they're
using
the
same
I,
don't
know
which
version
they
are
using.
We
had
one
report
from
one
from
a
user
that
had
an
issue
with
with
Sal
with
their
SDK
that
was
crossing
at
open
Telemetry,
something
like
that
so
yeah
they
are
using
it,
but
don't
know
the
exactly
the
version
they
are.
They
are
Distributing
but
yeah.
Okay,.
D
So
we
know
the
could
you
have
Sirata
or
somebody
you
know
from
Splunk
engineering.
You
know
join
this
meeting.
D
Okay,
both
diverse
and
Android
folks,
you
know,
will
be
great.
B
Yeah
we
have
there
in
the
in
the
document
on
the
at
the
part
we
have
people
connected
from
from
a
Splunk
previous
to
be
not
been
in.
This
plank,
I
was
I,
can't,
say.
B
D
Yeah
and
on
the
resource
attributes.
D
Bring
that
up
now
so
on
the
specifics
you
know
we
I,
don't
know
if
that
was
the
same
thing
yeah,
so
I
created
this
ticket,
it's
not
assigned
to
anyone,
but
can
you
take
a
look
at
this.
C
Yeah
we
can,
we
can
remove
that
and
it
was
just
kind
of
a
decision
because
it
fit
there
a
little
bit
but
yeah
yeah.
Well,
so,
what's
the?
What
is
the
the
intention
for
the
service
name
namespace?
C
So
that's
noted
in
those
links.
D
There
yeah
yeah,
it's
basically
namespace
is
is
the
name
for
the
collection
of
things
right.
Okay,
so
here
the
collection
could
be
your
IOS
app
and
Android
app.
You
know
if,
if
they
are
both,
you
know.
Typically,
customers
have
multiple
apps.
So
if
you
want
to
group
identify
you
know
which
to
which
few
apps
belong
to
one
name
space,
you
know
they
can
be
grouped
together.
Okay,.
A
C
B
So
probably
on
configuration
when
creating
or
the
visual
part,
I
I
I
know
we
do
for
some
environment,
violence,
I,
don't
know
if
this
one
can
be
read
from
those
environment
variables
that
we
have
so
it
can
be
automatically
configured
for
for.
C
It's
not
required.
Is
it
the
problem
with
environmental
variables?
Is
they
only
work
when
you
build
from
xcode
I.
B
Mean
when
you
are
I
mean,
if
you
you
can
read
them
also
in
the
process
if
they
feed
they
are
there.
If
you
run
the
process
with
those
environment
variables
first,
it
should
be
red.
C
C
If
you
release
a
device,
you
know
you.
B
Download
from
the
App
Store
you're
not
going
to
get
it
there,
yeah
yeah,
that's
right!
You
can
also,
for
example,
I.
Do
that
read
it
from
the
P
list
file.
B
You
can
release
with
a
police
file
and
read
the
value
from
there.
I
do
that
also,
so
you
can
modify
things
from
the
previous
screen
file.
C
B
I
can
add
that,
to
my
to
my
so
I
can
bring
what
I
have
to
open
to
to
that.
One
I
use
read
the
same
that
we
read
from
the
environment
variable,
but
from
the
previous
file
of
the
main
application
that
that
that
can
be
useful
to
for
developers
to
configure
things
instead
of
yeah.
B
D
So
yeah,
so
this
I
think
maybe
I
don't
know
how
folks
work,
but
you
you
can.
You
can
assign
it
to
one
of
yourself.
E
You
had
that
spreadsheet
right
that
that
you're
showing
can
you
add
that
in
the
document
can
you
add
that
link
for
the
spreadsheet.
C
It's
there,
it's
a
semantic
conventions.
There.
D
D
B
D
There
is
also
the
other
thing
from
last
week
was:
you
know,
we'll
also
need
a
standard
way
to
identify
the
hybrid
agents
but
I
think
I
think
we
can
wait
on
this
too.
I,
don't
know
whether
this
is
really
needed
right
away
or
we
can
wait,
but
but
basically.
D
React
native
Factor
when,
when
we
use
them
I
think
it'll
be
good
to
indicate
that
in
in
one
attribute.
C
Have
you
considered
like
what
is
the
what
is
like
the
the
request
for
the
support
for
these
sort
of
libraries?
Because
have
you
considered
like
spinning
up
new
cigs
for
them,
because,
in
my
experience,
they're
so
Divergent
from
the
based
platform?
And
they
move
so
quickly?
Just
in
terms
of
like
the
features
that
they
add,
remove
or
tweak
that
it's
not
feasible
to
support
them
like
on
top
of
like
a
swift,
SDK
or
an
Android
SDK?
C
They
almost
need
their
own,
their
own
libraries
built
for
them.
D
I
I
think
I'm
not
too
familiar
with
them,
but
I'm
told
that,
like
you
said
you
know
they
are
built
on
top
of
the
you
know
the
the
base
agent
on,
but
if
you're
saying
they
move,
you
know
in
a
different
style,
then
are
you
so
you
is
it
typically
a
different
set
of
folks
that
work
on
it
and
you
guys
don't
get
into
it?
Well,.
C
Yeah,
so
the
the
idea
is
at
least
yeah,
so
I've
done
work
with
xamarin
and
react
native,
and
the
idea
is
is
that
you
can
have
Engineers
who
are
familiar
with
with
like
c-sharp
for
in
xamarin's
case
or
JavaScript
in
react,
native's
case
or
reacts
in
react.
Native's
case
you
know
Engineers,
who
are
comfortable
with
those
languages
working
on
products
that
you
can
ship
to
IOS
and
Android.
C
People
really
want
it
to
work
and
they'll
put
a
lot
of
effort
into
making
it
work,
but
I
I've
tried
multiple
times
a
New
Relic
to
to
get
a
react
native
like
what
my
brain's
not
working
an
agent.
That's
the
word
I'm
looking
for
an
agent
working
for
react
native
to
actual
like
viable.
You
know
usable
status
and
by
the
time
something
gets
implemented
for
some
version
of
react
native.
They
release
like
react.
Native
itself
is
a
beta
right
like
it's,
not
they.
C
They
reserve
the
right
to
break
any
feature
that
is
existing
in
it
and
I.
Don't
know
if
that's
changed
in
the
past
like
a
year
or
two,
but
the
last
I
looked
at
it.
It
was
a
beta
and
they
definitely
stuck
to
that.
Like
you
know,
by
the
time
we
got
a
solution
for
one
version
of
react
native,
they
would
have
bumped
the
version
and
basically
everything
we
were
using
to
instrument
it
would
change
like
it
wouldn't
be
viable
anymore,
so
we'd
have
to
basically
start
from
the
ground.
D
D
Is
it
the
same
with
the
xamarin
and
flutter
as
well.
C
Yeah
they're
difficult
to
work
with,
okay
and
and
attacking
it
from
like
from
a
native
side
is,
is,
is
really
difficult.
I
think
that
they
needed
they
need
specific.
C
D
Fair
enough,
so
that
maybe
you
know,
let's
just
differ
talking
about
it
until,
like
someone,
you
know,
starts
working
on
it
practically
at
this
point,
it's
only
you
know,
Theory.
E
Yeah
I
think
this.
Essentially,
you
need
to
collect
some
of
the
stuff
from
the
build
tool
chain
right,
because
yeah.
A
E
What
you're
really
interested
in
is
the
version
and
what
are
the
things
that
are
specific
to
xamarin
or
react
native
right,
as
opposed
to
the
platform
projects
that
they
built
right
at
the
end
of
the
build
tool
chain:
correct,
correct,
oh
yeah,
so
just
having
a
device.
You
know
the
agent
type
that
is
useful,
that
it
is
you
know,
project
that
is
an
iOS
project
that
is
built
from
you
know,
laminarin
tool
chain,
but
beyond
that
you're
not
getting
anything
right.
So
yeah.
C
E
And
same
examine,
I
have
maintained,
it
maintained
this
government
binding
for
almost
four
years,
and
you
know
every
time
customer
would
use
the
new
version.
The
old
one
would
not
work
so
then
they
would
come
back
and
I
would
rebuild
the
binding
and
so
yeah
I
did
that
for
a
couple.
You
know
almost
four
years
right.
D
Okay,
I
actually
put
that
note
in
the
last
minute
last
week's
notes,
but
so
that
it's
all
in
one
place,
Okay.
B
B
And
one
one
thing
more
that
that
it
was
talked
last
week
that
maybe
also
Bryce
cannot
his
point
of
view
is
the
purpose
of
the
login
API
that
we
we
were
talking
a
bit
about
it,
so
that
I.
B
The
login
API,
but
correct
me
if
I
am
wrong
price,
is
that
it
can
be
used
to
send
logs
a
standard
locks
to
the
to
any
back
end
right.
Yeah.
A
B
Not
only
okay,
yeah,
okay,
so
we
also
have
mac
applications
that
can
be
used
in
open
Telemetry
to
send
a
standard
standard
logs
for
for
other
uses.
D
So
how
does
it
work,
though,
I'm
just
curious,
so
like,
for
example,
in
in
Java,
the
lock
for
J
and
slf4j
the
popular
logging
libraries?
They
support
something
called
log
upender.
So
it's
a
it's
a
back-end.
D
You
know
behind
the
scenes
configuration
so
when
the
developers
write
log
statements,
you
know
they
they
obtain
a
logger
and
then
they
say
log.debug,
log
dot
info
and
all
but
the
when
they
obtain
the
logger
that
logger
at
runtime
it
it.
You
know
behind
the
scenes.
You
know
you
could
configure
it.
You
know
where
the
logs
get
sent.
D
You
could
configure
it
to
send
the
logs
to
multiple
destinations.
You
know
to
the
yesterday
out
or
to
files
or
to
you
know
some
HTTP
destination.
So
in
the
case
of
Swift,
what
I
was
you
know
what?
What
is
your?
How
does
it
work.
C
B
I,
don't
think
it's
feasible.
Oh.
B
Low
themselves
are
very
private,
I'll,
say
because
I
I
have
been
working
on
that
for
a
long
time
in
the
past,
I
talked
about
it.
They
have
previously
to
the
OS
log.
They
have
ASL
logs
that
those
were
configurable,
but
they
always
looks
now
are
handled
privately
by
the
operating
system.
D
B
It
I
hope
you
can
now,
but
my
my
experience
with
it
is
that
you
cannot
get
that
automatically.
What
what
really
exists
out.
There
is
a
library
from
Apple
that
it's
called
sweet
logs
that
that's
a
real
log
framework
that
allows
you
to
save
your
loss
to
different
places.
You
can
capture
from
there
and
we
should
maybe
yeah.
We
should
work
for
that.
Yeah.
C
B
Not
thought
for
that
at
least
it
was
not
two
years
ago
and
I.
Don't
think
that
has
changed,
but
maybe
it
has.
We.
B
Support,
for
example,
that's
something
that
I
do
in
my
product
is
capturing
a
standard
output
and
a
standard
error
from
the
process
and
send
that
as
part
of
of
the
logs
in
the
user,
configures
it
or
just
I
mean
we
can
provide
that
those
are
as
libraries.
A
B
The
user
can
just
can
just
join
or
or
link
with
them
and
enable,
if,
if
desired
or
suddenly,
that.
B
Maybe
we
I
mean,
probably
you
can
also
have
in
that
system,
something
that
is
really
in
the
standard
logs
or
the
logs
outside
that
and
sending
with
another
system
yeah,
but
yeah
I
mean
we
can
do
things
like
that
that
that's
doable
but
I
think
you
always
lock
itself.
It's
a
it's
a.
He
said.
B
It's
how
to
say:
I
know,
exit
strip,
I
would
say
yeah.
B
B
I
think,
as
Bryce
said,
was
having
a
log
support
as
one
of
the
pillars
of
the
observability
yeah.
E
So
so
I
think
what
what
Santosh?
What
is
he
saying
is
in
case
of
Java,
or
you
know,
a
different
language.
The
the
open
Telemetry
logs
can
also
be
you
know
it
can
use
the
log
furniture
or
something
as
the
the
output
format.
E
So
interesting,
whatever
yeah.
D
The
application
logs
normal
logs
and
then
you
you
build
something
called
log
upender
that
is
then
configured
to
to
take
the
logs.
So
what
that
log
upender
does
is
it
takes
the
normal
logs
from
the
application
and
then
converts
them
into
otlp,
Hotel
logs
and
then
sends
it
to
the
hotel
destination.
D
E
Okay,
the
offender
would
actually
emit
to
TLP,
okay
and,
and
then
the
Java,
the
otlp
library
will
have
somewhere
to
emitted
to
whichever
endpoint
yeah.
D
E
Okay,
yeah
all
right,
let's
make
sense
yeah
if
you
cannot
yeah,
so
you
have
to
use
something
that
is,
you
know
you
can
expose,
not
not
something
like
not
something
that
is
private
right.
Yeah
log4j
is
obviously
writing
into
some.
You
know
it
is
writing
into
logs
that
is
available
right
for
the
whatever
privileges
of
the
application
has.
Also
output
will
have
yeah.
D
B
Yeah
that
yeah
that's
the
primary
use,
probably
was
for
mobile
Sometimes.
Some
errors,
maybe
are
better
I
mean
you
might
want
to
handle
your
not
only
your
events
but
your
logs
as
a
a
way
to
signal
errors
that
are
critical.
Probably
you
can
do
that
faster
that
way,
that,
by
using
events
and
things
like
that,
yeah
I
think
that
being.
B
It's
much
much
better,
yeah,
okay,
so
going
to
the
topics
of
today,
I
mean
we
have
been
talking
a
lot
yeah.
D
Oh
I
see
okay,
so
yeah,
so
yeah
I
think
I
was
internally,
I
was
talking
with
Stephen
I,
don't
know
he
we
have
another
meeting.
He
didn't
join.
E
D
He
mentioned
that
the
on
a
span
as
a
tribute
you
know
today,
you
know
there
is
no
attribute
to
you
know
to
represent
the
request
and
the
response.
Content
length
and
I
think
that
we
needed
in
our
product
and.
B
D
B
If,
if
he
can
create
a
pair
of
creative,
he
doesn't
find
a
way,
I
mean
I.
Think
it's.
There
is
it's
quite
easy.
Yeah
yeah.
B
B
D
B
D
That
that's
all
I
had
yeah.
B
Okay,
and
the
next
topic
is
the
one
that
we
have
told
me
for
it.
Just
was
just
to
validate
this
approach.
It
will
be
if
there
is
no
environment
variable
for
that
check.
Your
P
list
for
a
value
which
has
the
same
key
as
the
environment
variable,
and
it
hits
their
use
that
one
using
the
the
P
list
of
the
the
infopy
list
of
the
main
bundle
of
the
application
building.
That's.
B
Try
to
to
add
it
this
week,
cool,
okay,
great,
it's
really
very
easy.
B
You
want
yeah,
that's
right,
I,
don't
know
if
we
are
already
using
environment.
D
Oh,
so
this
depends
on
this
change
that
we
just
discussed
yeah.
You
could
fuss
this
dude.
C
I
think
that
I
think
that
the
I
think
that
the
the
the
the
resource,
the
SD
resource,
SDK
like
instrumentation,
that
I
added,
might
be
grabbing
that
as
well.
So
we
can
just
remove
that.
Okay,
I,
don't
think
that
we
need
to
actually
send
the
bundle
ID
it
just
I
just
felt
like
it
fit
in
that
spot,
but
I
guess
it
doesn't.
So,
let's
just
pick
it
out.
D
Yeah
only
thing
is,
you
know
I,
it
might
be
a
breaking
change
like
if,
if
people
are
expecting
it,
you
know
your
back-end
teams,
you
might
want
to.
Let
them
know.
A
D
B
Or
we
can
do
another
I
mean
we
can
use
that
as
a
fallback.
If
there
are
no.
B
A
D
Think
maybe
then
you
know
if
you
could
assign
this
to
an
you
know,
one
of
you.
You
know.
D
E
E
C
B
B
That
was
like
a
generic
class.
I
mean
it's
good,
but
could
it
be
used
something
like
that
further
for
the
other
or.
C
C
Yeah
I
know
which
one
you're
talking
about
the
oh
I,
can't
remember
what
the
name
of
it
is
so
I
I
use
that
from
the
Java
SDK.
What
is
it?
What
is
it
called?
It's
called
the
component.
C
Yeah,
so
this
the
spec
requires
a
schema
URL,
which
we
don't
use
anywhere,
but
maybe
we
should
start
respecting
it,
and
so,
if
we
wanted
to
update
the
other,
the
other
components
that
trace
and
the
metric
stuff
to
also
respect
the
schema
URL,
then
we
might
or
yeah
we
might
want
to
use
this
component
registry.
Because
yeah,
the
the
the
lookup
for
a
scope
is
like
the
combination
of
all
three
values,
the
name,
the
version
and
the
schema
and
yeah.
A
C
Exactly
yeah,
so
so,
if
you
use
some
combination
of
those
values,
it
is
a
different
lookup,
because
it's
a
required
like
the
you
know.
Nil
also
is
a
a
unique.
You
know.
What
is
it
set
theory
right,
so
it's
the
the
empty
set
is
its
own
unique
set.
C
So
if
you,
if
you
reference
a
component
by
name
and
version
but
not
schema
but
then
later
come
back
and
reference
name
version
and
schema,
that's
a
different
component
that
you're
looking
at
so
this
is.
This
is
kind
of
the
solution
that
the
Java,
the
Java
sck
came
up
with
and
I.
Think
it's
pretty
apt.
C
You
could
probably
do
it
in
other
ways
that
might
be
a
little
less
heavy
on
the
on
the
the
array
structures
but
I.
Don't
think
that
you
know
this
is
probably
the
most
easy
to
understand
version
yeah.
B
B
If
it
will
make
sense
to
be
to
be
to
be
in
the
common
folder
instead,.
B
It
to
the
common
folder
yeah,
it's
so
I
mean
yeah.
It
makes
a
lot
of
sense
to
simplify
that
that
stuff,
so
maybe
move
in
there.
For
the
rest.
From
my
side,
everything
is
is
perfect,
I
mean
it's
clear.
Is
it
it
looks
functional
so
yeah
just
missing
test.
B
Yeah,
but
we
should
say
that
it's
very
unfortunately
that
yeah
that's
right,
so,
okay,
yeah
for
the
rest,
I
really
I,
really
like
it.
It
follows
the
rest
of
the
of
the
API
in
other
areas
of
the
framework.
So
it's
it's
really
great.
B
Okay,
okay,
so
no
more
topics
from
my
site.
If
you
have
any
other
topic,
do
you
do
you
know
I,
do
you
know
Bryce
I
fixed
the
memory
that
was
not
being
freed?
Oh.
C
B
B
Should
be
I
mean
I
think
that
it
should
be
not
finding
active
contests
that
it
shouldn't
so
in
a
child
context
at
the
span
ended.
It
won't
find
it
because
the
Espana
was
ended,
yeah,
sometimes
in
another
response.
Even
when
it
had
ended,
it
will
still
continue
finding
that
in
other
contexts,.
B
That
is
seen
anything
that
parent
context
cool,
because
before
as
it
was
alive,
it
will
be
found
about
the
by
this
child
context.
So
some
of
the
things
I
had
to
change
because
they
were
based
on
the
behavior
that
shouldn't
happen.
I
mean.
A
B
C
D
B
Ended
so
it
was
not
correct
to
find
it.
B
B
Failing
for
you,
for
example,
but
take
that
that's
not
the
case.
Some
of
the
tests
in
the
project
I
had
to
update
them,
also
because
we're
behaving
on
the
wrong
waiting
for
it
to
be
alive
when
it
shouldn't
so
I
had
to
make
some
synchronization
to
validate
that.
Okay.
A
B
Great,
if
there
are
no
more
topics,
we
are
almost
on
time,
I
mean.
B
Okay,
then
so
we
can
next
week,
I
I
won't
be
joining
it's
about
holidays
here
in
Spain.
B
That's
that's
in
my
office.
That's
my
engineering
director
I.