►
From YouTube: 2020-11-13 meeting
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).
C
A
B
Me
show
me
so
the
class
has
javadoc,
and
yet,
when
honorard
looks
at
it,
he
sees
no
javadoc.
C
A
B
C
B
A
B
B
B
So
if
I
click
over
here,
it
gives
me
the
actual,
like
output
from
github
over
here,
which
is
pretty
cool,
and
then
I
can
click
checkout
and
I
can
give
it
a
name.
I
always
prefix
with
pr.
So
I
can
remember
to
delete
them
and
then
it'll
just
check
it
out
locally
and
I
can
check
out
the
pr
or,
if
I
don't
want
to
I
mean
checking
it
out
is
nice,
but
I
can
also
view
all
the
diffs
here
and
make.
C
C
Can
you
like
control
click
or
whatever
on
these
names,
and
it
goes
to
the
code
that
would
be
sweet?
B
D
A
Awesome,
yes,
yeah.
I
don't
know
what
I
feel
like.
We
need
to
send
you
some
stickers
or
something
for
this
productivity,
major
productivity
tip
for
maintainers.
B
B
Oh
they're
there
just
there's
a
lot
of
it
takes
a
long
time,
yeah
yeah,
it's
really
pretty
rad.
It
is
getting
better
each
release
like
they
really
put
it
in.
I
think
like
I
want
to
say
nine
months
ago.
Finally,
and
it's
been
getting
better
with
every
release,
so
it's
pretty
yeah,
it's
pretty
cool,
and
I
love
even
looking
at
here
or
sorry.
B
Looking
at
it's
a
little
wider
looking
at
this
view,
you
can
actually
go
and
do
just
like
a
command
d
to
look
at
the
diff
side
by
side,
which
I
find
way
easier
to
look
at
than
the
github
diffu
yeah.
You
can
step
through
the
changes.
B
I
think
we
know
what
the
diff
window
looks
like
yeah
right
now.
You
can
step
through
the
changes
and
you
can
like
hit
f7
repeatedly,
it'll,
go
to
every
change
and
then
jump
to
the
next
file
go
through
every
change,
all
right,
yeah.
It's
really
it's
pretty
slick.
It's
not
perfect.
There's
like
I
have
some
issues,
but
it's
pretty
darn
good
so
to
try
this
down
here.
B
C
E
C
C
C
A
Yeah
because
I
noticed
it
because
it
was
remember,
it
was
broken
and
I
would.
A
I
pinged
you
on
slack
to
approve
one
of
my
prs
to
fix
master
because
we
had
submerged.
Oh.
E
A
C
A
B
Well,
we
definitely
have
master
builds
on
java
because
they
keep
failing
with
stupid,
stupid
seg
faults.
You
still
see
this
oh
yeah
tons
of
them.
I
probably
saw
like
eight
or
nine
today
alone.
What.
A
Oh,
did
you
get?
I
saw
you
had
a
oh,
an
issue
to
do
the
cat
thing
I
haven't
gotten.
B
D
B
C
C
A
C
A
Let
me
I
will
find
our
notes
from
this
morning.
We
talked
a
lot
about
spi's.
A
So,
let's
see
here
share
yeah,
so
I
think
we
had
for
the
bite
buddy
one.
I
think
we
agreed
that
it
made
some
made
sense
to
give
some
vendors
would
want
access
to
bite
buddy
directly.
C
A
The
particular
use
case
for
it
was
not
super
convincing
because
it's
a
use
case
that
we
already
have
a
hook
for,
but
the
hook
was
broken
and
but
pavel
was
pretty
insistent
that
he
didn't
feel
comfortable.
A
Not
having
that
hook
to
be
able
to
do
other
things
short
of
you
know
waiting
for
a
new
release
from
the
java
agent
and
it
seems
reasonable,
like
if
you
want,
if
you
have
some
existing
bite,
buddy
stuff,
for
example,
that
you
want
to
add.
A
A
A
Yeah,
just
that
you
need
to
be
able
to
have
access
to
some
of
the
to
repair.
I
guess
the
he
was
he's,
convinced
that
that
he
may
need.
C
A
B
C
C
Like
so
in
my
comment
I
had
mentioned
outside
in
versus
inside
out,
like
it
seems
like
we're
going
from
the
outside
in
which
is
not
common
in
libraries.
Maybe
it's
fine,
but
if
we're
going
to
add
an
sp
at
every
single
code
path
on
the
flip
side,
just
having
builders
and
letting
people
wire
together
in
their
own
main
function,
that's
also
an
option
that
could
be
more
idiomatic.
I
don't.
C
C
Like
I
think
nikita
is
a
distro
didn't
even
have
a
custom
pre-made,
which
is
interesting
like
I
have
it
just
for
the
heck
of
it.
It's
just
a
one
line
that
delegates,
so
it's
not
really
doing
anything,
but
that
could
easily
be
calling
a
builder
and
configure
stuff
programmatically
instead
of
spi.
I
think
so.
A
Yeah,
I
saw
your
comment
and
let's
find
that,
do
you
remember
which.
A
I
mean
I
hit
space
right
here.
There's
yeah!
I
think
that's
a
really
interesting.
I
did
not
make
that
connection.
C
A
Our
do
we
have
so
some
of
the
problem
is
we
need
stuff
to
run
inside
of
the
agent
class
loader,
so
we
want
those
classes
to
be
loaded,
but
I
mean
potentially
we
could
give
in
the
builder.
We
could
give
names
of
those
classes
to
load
in
the
agent
class
loader.
C
Eventually,
yeah
because
the
class
loader
stuff
could
make
that
idea,
I
mean
that
might
be
a
reason
not
to
do
it,
but
this
is
an
idea
that
came
up
like
if
you
have
agent
builder
dot,
add
bite
body,
transformer
dot,
add
instrumentation
module
dot
dot
ad
instead
of
spi's
for
everything
it
can
work.
It
seems.
D
C
A
Yeah,
let
me
at
least.
A
E
D
A
Yeah
yeah,
I
think
that's
a
good
thing
to
think
about
at
least
because
we
are
facing
spi
explosion.
I
think
we
have
it
probably
at
least
three
or
four
more
spi's
that
we're
going
to.
B
B
B
A
We
that
a
good
on
this
other
spi,
what
it
sort
of
boiled
down
to
is
pavel,
wanted
to
do
something
similar
to
the
open,
telemetry
api
sdk,
where
we
put
the
api
in
the
bootstrap
class
loader.
But
we
put
the
sdk
implementation
in
the
agent
class
loader,
and
so
he
has
his
own
sort.
B
A
A
So
actually,
I
did
this
one
actually
documented
our
decision.
Our
discussion
so
he's
gonna
kind
of
explore
this,
making
it
like
a
component
installer
which
would
sort
of
generalize
the
idea
of
a
the
tracer
installer.
A
But
also,
then
you
could
have
your
own
custom
component
installer
and
it
sort
of
needs
this
before
and
after
hook
of
before
we
start
the
bite
buddy
agent
and
after
because,
before
the
age
before
we
start
the
bite,
buddy
agent.
We
have
this
hacky
thing
in
our
agent
class
loader
to
force
sort
of
the
spi
to
get
resolved
correctly
right
away.
A
So
I
was
trying
to
think
with
how,
with
a
builder,
I
mean
it's
still
the
kind
of
a
similar
thing
right
I
mean.
Basically
it
would
be
a
consolidating,
I
mean
it
sort
of
doesn't
have
to
be
consistent.
I
mean
sorry
with
a
builder.
A
We
can
just
have
all
these
things
added,
just
instead
of
an
spi
yeah,
so
we
could
sort
of
just
conglomerate
everything
into
one.
A
Which
would
be
yeah
yeah?
The
spi
explosion
is
a
little.
A
A
These
don't
I
don't
know
why,
but
this
is
nice.
This
is
some
nice.
A
couple
nice
diagrams
pavel
wrote
up
to
help
sort
of
explain
our
initialization
sequence
and
then
sort
of
how
the
class
loaders,
what
we
put
in
different
class
loaders
and
adding
that
to.
A
Our
docs,
which
helped
to
to
sort
of
this
that
helped
this
discussion
sort
of
helped.
I
think
other
people
sort
of
visualize
as
we
were
discussing.
A
Example
of
oh
yeah
nikita's
example,
so
a
couple
extra
things
that
I
imagine
we'll
do
later
of
how
to
use
muzzle
in
the
custom
distro.
I
think
these
were
things
that
pavel
was
thinking
of
that
he's
done
in
his
distro
and,
of
course
everybody
sort
of
wants
this.
But
it's
no,
but
it's
a.
I
think
it's
painful.
A
Yeah,
so
this
is
gonna,
be
I
th
this
I'm
really
happy
how
this
turned
out
and
and
even
if
we
completely
change,
you
know
the
the
spi's
and
whatnot.
It's
still.
It's
so
good.
To
have
an
example.
There.
A
And
then
I
think
we
skipped
over
other
things,
because
this
took
like
40
minutes
and
so
then
sergey
was
asking
about
config
file,
so
he
was
mentioning.
We
don't
publish,
we
don't
doc,
so
we
do
have
it.
It's
config
file
properties,
config
file
support,
but
we
don't
document
it
so
he
said
he
would
submit
a
pr
for
that.
A
Because
so
right
now
it
will
only
you
have
to
provide
a
location
to
read
the
the
config
file,
and
so
in
my
in
my
experience
or
my
what
I've
done
in
the
past
is
the
default
location
is
in
like
relative
to
the
it's
in
the
same
directory
as
the
jar,
the
agent
jar,
and
I
know
new
relic,
I
think,
does
the
same,
and
but
somebody.
A
Was
saying
that
I
can't
remember
who
now
were
you
on
the
call
this
morning?
John?
A
He
was
saying
his
his
feeling
was
that
having
a
default
location
was
more,
can
cause
more
confused
confusion
and
I
think
the
new
relic
folks,
both
you
know,
agreed
and
but
also
you
know,
so
it's
also
helpful,
but
I
don't
think
there's
any.
Nobody
really
seems
to
care
too
much
about
this.
A
People
cool
we
did
discuss
because
pavel
so
pablo
has
was
mentioned.
He
has
his
sort
of
own
config
and
that
we're
discussing
the
precedence
and
sort
of
how,
if
I
think
I
asked
him,
because
if
you
wanted
to
hit
the
precedence
of
that
property
source,
the
vendor
specific
config
to
override
like
the
open,
telemetry
config.
So
so
people
can
still
do
open,
telemetry,
config,
environment
variables
and
that
will
still
take
precedence
over
your
vendor
config
story.
A
And
so,
for
example,
I
for
me
I'm
I've
kind
of
hidden
the
entire
open
telemetry
configuration,
because
I
don't
want
to
expose
all
of
those
configuration
properties.
I
want
just
a
really
limited
number
of
config,
which
nikita
doesn't
understand.
He
was
a.
A
Why
would
you
know
he
understands
not
having
minimal
configuration,
but
why
would
you
not
let
users
do
more
configuration
stuff
which
I
get
also,
but
for
my
we
didn't
go
into
it,
but
from
my
each
knob
that
people
can
turn
is
something
else
that
people
can
open
support
issues
about.
So
that's
sort
of
why
I've
kept
it
minimal.
A
So
pavel
and
I
are
both
potentially
support-
an
option
to
suppress
the
open.
Telemetry
agent's
configuration
properties,
so
maybe
at
some
point
we
would
add
that
to
like
the
property
source
or
something
as
a
way
to
say,
take
this
property
source
and
only
this
property
source.
B
That's
funny,
this
is
the
second
time
today.
I've
heard
a
similar
sentiment,
so
in
the
metrics
saved
today,
there's
a
there
was
somebody
who,
from
a
company,
I
can't
remember
the
name
of
who's
been
doing.
Metrics,
ingest
and
metrics
work
for
a
decade
and
they've
been
like
ingesting
histograms
for
a
decade.
They
were
talking
about
histogram
support
for
open
telemetry
and
what
the
histogram
should
look
like
and
how
they
should
be
configured,
and
he
said
after
10
years
of
supporting
users
with
histograms.
B
He
has
decided
emphatically
that
he
should
allow
exactly
zero
configuration,
because
users
will
always
do
it
wrong,
and
then
they
will
complain
that
it's
not
working
the
way
they
expect
it,
and
so
his
his
his
opinion
now
is
that
everyone
should
standardize
on
one
single
histogram
that
serves
99
of
the
use
cases
and
make
it
not
configurable
ever.
B
It
was
a
histogram
implementation
that
he
wrote
like
he
had
written
a
pretty
good
academic
paper
about
the
implementation
of
how
it
works,
and
it
has
something
like
his
the
default
that
that
he
publishes
and
uses
for
his
company
has
something
something
crazy
like
37
000
buckets
in
it,
but
it's
also
sparse.
So
if
you
don't
populate
them,
it
doesn't
use
any
memory
and
he
actually
has
used
it
on
embedded
systems
like
that.
B
Don't
actually
have
like
they
don't
have
floating
point
arithmetic
on
it
on
it
and
it
still
works
just
fine
in
all
sorts
of
crazy
use
cases
across
wide
varieties
of
different
different
hardware
and
uses
anyway.
The
point
is
like:
don't
let
people
configure
stuff
because
they'll
always
configure
it
wrong,
is
what
the
what
the
the
lesson
was.
A
Yeah,
even
if
they
configure
it
yeah
or
that
there's
just
it
raises
more
questions,
it
raises
more
support
tickets.
It
does
bring
in
more
customers
also.
So
you
know
it's
a
like
I,
as
opposed
to
saying
no,
you
can't
configure
this.
Those
customers
will
go
away,
but
you
know
you
have
to
be.
Your
organization
has
to
value
that
balance
and
say
yes,
it's
worth
the
support
cost
to
retain.
A
You
know
the
extra
number
of
customers,
and
my
experience
at
least
where
I'm
currently,
is
that
it's
it's
not
funded
at
that
level
and
it's
cheaper
for
us.
It's
easy,
you
know.
So
it's
it's.
I
don't
know
cheaper
easier,
better
to
say.
No
sorry,
we
don't
have
that
feature
and
there's
other
vendors.
That
would
be
happy
to
help
you.
B
Yeah
I
have
worked
at
companies
where
they've
gone
the
opposite
route
and
accepted
every
possible
feature
and
guess
what
happens
a
decade
in
you
end
up
supporting
the
decade
worth
of
features
that
maybe
one
customer
uses
each
and
it's
incredibly
expensive.
C
Yep,
that's
always
been
the
one
thing
about
our
meria
project,
which
I've
always
been
not
clear
on
like
they're.
Also
the
stance
of
accept
any
feature
if
it
makes
user
happy
just
do
it,
but
then
I
always
wonder,
then:
is
this
really
going
to
be
maintainable
if
they
ever
use
every
single
feature,
there's
a
lot
of
features
in
there,
but
it's
just
that
trade-off.
I
think
they're
making
the
trade-off
and
supporting
it
all.
C
A
It
was
interesting
this
sergey
brought
this
up
and
then
we've
had
like
two
people,
just
in
the
last
day,
asked
something
similar
on
gitter
and
so
just
sort
of
reiterated.
What
we
had
discussed
several
months
back,
which
was
that
we
don't
really
want
to
publish
benchmarks
since
it's
so
dependent
on
workload.
And
it's
you
know
you
look
at
other
apm
vendors.
A
Don't
do
that
because
it's
it's
tricky
situation,
but
it
was
brought
up
that
memory
is
potentially
a
something
that
would
be
reasonable
to
publish
memory
like
a
minimum
memory
footprint
that
you
know
take
a
small
as
you
can
take
a
small,
app
small
number
of
classes
so
that
you
know
we're
not
like
instrumenting
a
bunch
of
stuff
but
minimum
to
load
the
agent
all
the
agent
classes-
and
I
know
we've
seen
a
link
to
this
one
that
I
think
you
opened
onorak,
that
you
know
that
the
you
know,
just
as
a
minimum
we
need
x
memory
seems
reasonable
to
publish
that.
C
That's
important
like
one
of
the
reasons
we
were
looking
at,
that
was
using
the
agent
in
lambda
and
so
the
memory
requirements
there
are
much
more
strict
than
so
knowing
the
minimum
and
not
even
the
minimum
like
knowing
the
minimum
for
operation.
Yeah
is
important
for
those
memory
constrained
areas.
C
A
Oh
so
the
exporter
class
loader,
is
only
if
you
use
that
hotel
exporter
jar,
external
exporter,.
D
A
Internally,
for
the
all
artifact,
we
bundle
everything
together
in
the
agent
class
loader.
So
technically
there
shouldn't
be
any
duplicate
classes
anywhere,
but
I
think
that's
not
correct.
C
C
A
C
C
Numbers
just
for
reference,
like
the
x-ray
sdk,
we
do
copy
like
every
release.
We
run
the
jmh
on
an
amazon
instance
like
the
specs
are
fixed
across
each
release.
It's
like
it's
one,
x3
large
or
something
like
that,
and
then
we
copy
it
in.
I
guess
if
the
instance
is
fixed,
it's
still
sort
of
maybe
useful
information,
but
at
the
same
time
I
don't
know
if
I've
ever
found
those
numbers
very
useful,
but
we
do
do.
A
That
I
could,
I
could
see
doing
that
for
the
sdk
for
the
there's
that
spec
issue
for
the
benchmarking,
which
is
more
sdk
targeted
for
the
agent
it's
a
little
harder
just
because
what
do
you
consider
a
real
application
right
as
opposed
to
jmh
benchmarks?
It's
super
clear
and.
C
A
D
E
B
Oh
yeah,
I
I
got
some
popcorn
out
and
I
I
kept
the
movie
short.
I
think
I
didn't
want
to
deal
with
that,
but
they
for
some
reason.
They
really
don't
want
to
publish
anything.
In
fact,
they
don't
even
release
anything
like
step.
One
is
actually
get
them
to
agree
to
release
proto
like
on
a
yeah
there's
changes.
C
C
E
C
B
Yeah
or
just
they
should,
if
they
really
need
to
have
latest
and
greatest
like
set
up
a
little
project,
it's
not
that
hard
to
type
a
little
project
that
will
publish
your
or
even
you,
don't
even
need
to
publish
it.
Just
like
do
what
the
api
does
right
now
and
pull
it
in
and
generate
the
classes
and
go
to
town.
So
do
you
know
recommending.
B
B
Well,
I
guess
you
know
when
you
do
a
release,
you
can
download
the
zip
of
from
from
github
right
of
the
contents,
so
you
could
always
just
build
it
off
of
release
workflow
and
grab
it
and
stick
it
into
your
project
or
something
I
don't
know
do
you
so
it's
aws
folks
who
are
whining
and
complaining.
Do
you
have
you
talked
to
them
about
like
just
setting
up
a
new
project.
C
B
C
Yeah,
so
about
that,
I'm
not
sure
about
the
use
case
of
that
theme.
Specifically,
I
know
about
the
other
team,
that's
using
otlp
in
their
back
end
and
so
that
they're
not
having
any
problems
like
the
version
is
fine,
but
that
was
just
my
point
in
that
people
use
the
proto
even
without
the
sdk.
I
think
so.
Keeping
them
independent
seems
to
make
sense,
but.
B
Let
me
pull
it
up,
yeah,
I'm
just
I'm
just
puzzled
about
like
if
they're
like
using
the
protocol,
and
they
got
this
new
type
into
the
protocol
that
nothing
yeah
supports.
Yeah,
nothing
uses
this.
It's
it's
curious
to
me.
Are
they
considering
the
protocol
as
being
a
completely
generic
telemetry
protocol
that
they
can
then
add
stuff
to
and
use
whatever
they?
I
just
it's
very
confusing
to
me.
I'm
I'm
puzzled.
C
So,
but
this
is
not
that
this
is
something
else
but
yeah.
This
is
a
amazon
project
to
provide
an
elastic
search
based
tracing
thing,
with
otlp
support
and
so
they're,
actually
using
our
meria,
which
I
recommended
to
them
and
they're
going
to
use
the
because
they
I
mean
now
aws
does
support
grpc,
but
they
didn't
want
to
support
http
also.
So
it
was
very
easy
but
yeah
so
they're
using
the
proto
here-
and
this
seems
like
a
use
case,
that
yuri
wasn't
considering.
I
don't
know.
B
A
What
I
was
just
trying
to
look
up,
it's
the
the
spec
spec
approvers.
B
D
A
Oh,
did
it
get
auto
no
22
hours?
Oh.
A
That's
I
mean
yeah
on
on
face
value.
I
am
without
reading
all
the
details.
It
is
surprising,
there'd,
be
so
much
pushback
on
publishing,
released
versions.
B
E
C
B
C
A
C
B
C
B
C
B
C
A
B
Does
what's
the
what's
the
project
that
will
automatically
create
like
a
maven
repository
from
any
github
repo?
B
C
B
B
Well,
I
mean
you
well,
I
think
what
like
what
you
were
saying.
We
could
just
pull
it
down
as
a
part
of
the
build
as
a
dependency,
but
do
actually
do
we
want
to
publish
something.
I
think
this
is
actually
a
good
question
like
if
we
weren't
publishing,
then
we
wouldn't
have
this
problem
of
people
trying
to
depend
on
it
and
thinking
we
should
be.
C
B
C
A
E
A
Until
oh,
okay,
so
this
is
blocked
until
this.
So
who's
gonna
comment
here
and
ask
bogdan,
if
you
has
any
further
questions,
looks.
B
Like
you,
you've
got,
the
you've
got
the
stick.
There
you
I've.
A
Got
the
ball
cool
all
right,
it's
me
I
will.
I
will
send
a
poke.
C
B
As
much
as
before,
I
actually
have
no
idea,
he
hasn't
been
visible
internally
either.
So
I.
B
He
desperately,
I
feel
like
he
needs
like
to
take
a
month
off
along
those
lines
we
have
splunk
has
like
planning
week
next
week,
so
I'm
my
availability
is
going
to
be
a
little
spotty,
so
we're
doing
our
quarterly
planning
which
I've
never.
I
don't
know
how
I've
never
been
in
quarterly
planning
before.
So
I
guess
we'll
find
out.
A
B
A
B
B
No
and
I
don't
recommend
any
other
country-
people
from
other
countries
fly
here
either
it's
a
mess
at
the
moment,
yeah,
it's
all
virtual
and
that's
going
on
next
week
in
the
middle
of
our
planning
week.
So
I
have
no
idea
what's
happening
I'll
be
busy
next
week,
I
think,
is
what
it
boils
down
to.
A
B
C
A
Oh
honor,
I
got
that
topic.
The
pavel
said
that
he
is
relying
on
those
the
java
agent
module
releases
artifacts
because
he
has
built
his
own
like
custom
instrumentation
and
he
reuses
parts
of
that
like
whether
it's
the
tracer.
C
C
C
One
idea
yeah,
I
think
I
am
okay
with
pavel
having
a
bit
of
an
annoying
shading
dilemma,
because
writing
an
agent
is
supposed
to
be
hard
anyways.
So
if
that
can
save
us
on
publishing
some
stuff
to
jfrog,
make
them
happier
also,
it's
still
worth
thinking
of
something
if
there's
a
way,
and
so
it
sounds
like
it
might
just
be
an
issue
of
the
objection
of
the
normal
jar
and
if
you
had
an
unanticipated
version,
maybe
then
you'd
be
able
to.
I
like.
A
Not
putting
not
putting
it
under
the
iana
yeah
like
a
debug
jar
that,
like.
A
And
this
would
the
idea
would
it
would
just
be
temporarily
getting
rid
of
those
artifacts
until
we
have
a
stable,
hopefully
someday,
we
have
a
stable.
C
C
E
A
It
to
the
agenda,
or
for
next
week
or
open
an
issue
to
chat
debug
jar.
B
Well,
honorable,
like
two
minutes
to
look
at
some,
so
I
started:
making
open
telemetry
sdk
extended
default,
open
telemetry,
I
put
in
a
pr
for
it,
but
I
just
realized
that
it
has
a
major,
fairly
major
issue,
and
I
wonder
if
you
could
take
a
look
at
what
I'm,
how
I'm
saying.
B
I'm
trying
to
figure
out
what
actually
to
do
so.
Let
me
share
my
screen,
and
maybe
you
can
you
can
give
me
some
advice
on
what
the
right
answer
is.
How
do
I
is
it
possible?
I
can't
share
more
than
one:
that's
really
annoying
all
right.
Well,
we'll
assume
this
is
good,
so
I'm
here
in
the
build
method
of
the
sdk
and
I'm
trying
to
figure
out
whether,
if
you're
using
this
builder,
whether
we
ever
want
to
use
the
super
yeah,
if
you're
using
the
builder,
should
we
just
skip
spi
all
together
or.
B
C
C
B
B
C
C
B
Yeah,
there's
something
right
along
those
lines.
So
basically
I
can
get
rid
of
these
supers
and
get
rid
of
these
super
methods
and
not
so,
if
you're
using
the
builder
you're,
not
using
spi,
and
if
you
want
sbi,
then
you
just
use
regular
sbi
from
infinites
in
the
api,
and
you
don't
worry
about
the
sdk
and
we
default.
We
default
to
the
same
defaults
that
the
spi
has
built
into
it.
Basically,.
B
We
don't
need
the
house,
I
don't
think
we
spi
out
there
yeah,
okay,
cool.
I
think
that's,
I
think
that's
great
thanks
for
talking
through
it.
I
was
heading
that
way
just
wanted
to
get
another
pair
of
eyes
on
yeah
cool,
awesome,
yeah
I'll
fix
that
up,
maybe
maybe
tomorrow
morning,
but
that's
all
right.
I
don't
think
we're
in
a
rush,
but
I
think
this
will
fix
nikita's
issue
and
we'll
clear
it
we'll
really
make
it
a
lot
clearer.
B
What
the
purpose
of
the
sdk
builder
is
it's
like,
and
I
actually
the
the
really
nice
thing
like
we've.
I
think
we
talked
about
this
is
I
can
really
as
a
further
further
pr
add
a
new
method
here,
which
is
add,
span,
processor,
right,
yep,
and
then
you
can
collect.
We
can
collect
up
the
span,
processors
and
put
them
on
the
put
them
through
the
build,
actually
just
add
them
to
the
tracer
provider.