►
From YouTube: 2023-03-21 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
C
B
C
B
All
right,
just
mostly
me,
Arielle
and
Rob,
just
kind
of
like
shooting
the
or
whatever
it's
fine
trying
to
you
know,
Ariel's
been
pretty
active,
so
we're
trying
to
just
kind
of
review
his
stuff,
good,
otherwise,
relatively
quiet.
How
are
you
not
too.
C
C
B
Didn't
take
you
out
to
but
the
CEOs
his
Ranch
in
Bozeman
on
blanket
on
his
name
amp
it
up
Frank,
salutement,
servicenow
founder
excuse
me.
B
D
D
A
B
C
Shall
we
go
ahead
and
do
a
specific
recap,
I'm
sure
excited
yeah,
Others
May
show
up,
others
may
not,
but
we
will
see
so
yeah
I
guess
this
is
the
Tito
specsig
recap
spring
break
Edition.
C
Most
people
are
on
spring
break
apparently,
but
yep
right
first
issue
was
not
like
super.
It's
not
very
controversial.
It's
just
a
spec
update,
clarifying
parent
contacts
passed
in
and
tracing
SDK
and
I.
Think
it's
just
specifying
that
there
should
be
an
Israel
property
on
it
to
figure
out.
If
the
spam
contacts
came
from
somewhere
else
or
not,
I
think.
F
C
This
is
definitely
old
news,
but
it's
controversial
so
it
keeps
coming
up
adding
spam
links
after
Statin
creation.
People
really
want
it.
There
is
an
issue
with
a
good
summary
of
where
things
stand
and
ultimately,
I
think
that
the
front
loaner
is
a
span
ad
link
operation.
C
So
adding
an
ad
link
method
to
span,
you
can
call
it
whatever
and
it
would
add
a
link,
there's
kind
of
the
second
one
that
people
are
not
super
enthusiastic
about,
where
you
would
have
a
tracer
record
link
and
it
would
kind
of
stuff
all
of
this
stuff
into
an
event
and
I.
Don't
think
that
yeah,
that
seems
like
generally
accepted
as
being
a
hack,
because
you
wouldn't
want
to
just
for
those
opposing
adding
this
directly.
C
This
fan
and
I
think
the
the
biggest
problem,
or
the
biggest
like
dissenters
here
are,
is
the
go
seg
and
mainly
just
around
like
the
fact
that
adding
a
method
to
an
interface
and
go
it's
like
you
can't
provide
like
a
default
implementation,
so
it
causes
breakage,
I,
guess
at
least
with
the
way
the
API
is
designed
and
I.
Don't
know.
C
There's
this
big
kind
of
a
circular
discussion
that
happens
every
time
this
comes
up
where,
like
we
need
to,
we
need
to
have
ways
to
add
methods
over
time
and
not
break
things,
but
in
go.
It
seems
to
be
you
know,
part
part
of
the
language
and
I
guess
C
plus
plus,
is
in
the
same
boat,
but
ultimately
there
needs
to
be
a
solution
and
I
think
towards
the
end.
The
go
Sig
was
like
all
right.
C
We
every
time
we
discuss
these
things,
we're
going
to
throw
out
all
these
other
options
that
aren't
adding
a
method,
but
in
the
end,
if
you
tell
us
that
we
must
add
a
a
new
method,
we
will
so
I
think
that's
where
things
probably
are
are
headed.
C
So
I
guess
TVD.
This
will
probably
be
something
that
we
end
up,
adding
to
our
span.
B
A
C
I
think
maybe
on
and
would
be
the
latest
I
would
see
you
being
able
to
do
this,
but
like
after
end,
I
feel
like
the
stand
becomes
immutable
and
I
would
be
surprised
after
span
end,
but
I
think
the
one
beef
that
most
people
will
have
with
this
is
around
sampling,
but
I
feel
like
the
sampling.
Sig
has
done
something
to
has
come
up
with
some
scheme.
That
makes
us
not
completely
terrible,
but
I,
don't
know
what
that
is
off
here.
F
C
C
I,
don't
know
enough
to
summarize
this
just
yet,
but
it
exists.
There
was
a
bit
of
a
discussion
about
schemas
and
schemas
in
schema,
URL
and
I.
Think.
C
I,
don't
know
what
the
schema
URL
is
for
anymore
after
this
discussion,
because
my
impression
was
that
you
would
be
able
to
use
this
thing
to
identify
the
schema
of
your
incoming
Telemetry,
and
if
you
wanted
to
do
like
some
sort
of
conversion
to
like
downgrade
it
or
whatever
upgrade
it.
You
could
potentially
do
this,
but
in
practice
it
seems
like
there
are
lots
of
issues
with
this,
and
mainly
there
are
two
things
that
were
brought
up
were
that
there
is
in
The
Collector.
C
It's
like
you
can
run
all
these
processors
to
totally
mutilate
the
data
on
on
the
way
through
and
then
you're
going
to
send
this
out
in
otlp,
and
if
you
there's
nothing
stopping
you
from
just
using
the
the
schema
URL.
That
was
on
a
data
to
begin
with,
and
but
the
data
might
not
actually
match.
You
know
the
schema
on
the
wrapper,
so
I
think
that
needs
to
be
figured
out
before
anybody
can
say
what
what
if
it
guarantees
the
schema.
C
C
There
ends
up
being
like
another,
pretty
big
problem
with
resource
detectors
and
people
who
are
using
a
a
number
of
resource
detectors
in
that
they
may
use
different
schema
URLs.
Each
detector
may
use
different
schema
URLs
and
when
you
merge
them,
I
think
you
can
only
merge
attributes
if
it's
giving
URL
matches.
C
C
C
I
think
the
Khan
is
that
all
of
these
dudes
tend
to
like
cross
over
with
the
larger
spec
work.
So
there
is
kind
of
arguments
that
having
them
in
in
the
stack
is,
is
a
better
place,
but
at
at
any
rate
there
is
at
least
a
proposal
to
separate
those
out.
C
And
then
yes,
this
one
is
probably
the
biggest
one
headline
news.
Perhaps
there
has
been
this
discussion
about
adopting
ECS
or
some
kind
of
merger
of
ECS,
an
open,
Telemetry
semantic
conventions.
C
I,
don't
know
that
the
I
don't
know
everything
is
like
officially
official,
but
it
seems
pretty
very
likely
that
this
is
happening
and
that
ECS
is
going
to
or
elastic
is
going
to
donate
ECS
to
otel
and
they're
going
to
go
through
some
period
of
seeing
what
oh
tell
us
to
find
seeing
what
ECS
has
to
find
and
trying
to
just
kind
of
have
a
unification
process
there.
C
People
think
about
this,
but
I
could
see
this
being
like
a
pretty
big
benefit
to
the
semantic
conventions
work
in
general
because
it
seems
like
there
is.
You
know
at
least
there's
at
least
like
15
months
of
scheduled
bikes
shedding
on
the
calendar
already
for
semantic
conventions
and
that
stuff
usually
usually
runs
longer
than
you
like.
So
this
could
potentially
be
a
shortcut
to
actually
getting
something
out
the
door
yeah
I
will
I'll
keep
everybody
updated.
C
If,
if
this
is
not
really
happening,
I
will
also
take
back
everything
that
I
just
said
in
a
future.
Specsig.
C
There's
this
long
discussion
about
x-ray
propagation
for
AWS,
client
calls
and
I
don't
know
really
as
as
evident
by
the
long
bullet
list
here.
But
yeah
there
seems
to
be
AWS
seems
to
be
an
example
of
something
that
does
not
really
fit
into
the
overall
kind
of
propagation
framework
and
guidelines
for
otel
like.
C
Ultimately,
the
idea
is
that
you
know
there's
a
propagation
API
and
you
register
a
propagator
or,
like
a
you,
know,
a
a
composite
propagator
with
with
that
system,
and
then
you
just
get
your
propagator
from
the
propagation
API
and
do
your
work,
but
for
for
AWS
it
doesn't
work
that
way
it
needs
to
have
x-ray
or
client
calls,
because
that's
the
only
way
that
AWS
will
actually
Stitch
things
up
inside
AWS.
They
don't
honor
any
other
formats.
C
They
kind
of
like
sidestep
the
propagation
API
and
use
like
x-ray
propagators
directly,
which
is
kind
of
like
I,
don't
know
like
it
feels
wrong,
I
guess
in
in
the
implementation
phase,
but
I
guess.
The
answer
is
that
everybody
is
doing
it
and
And
when
everybody
is
doing
it.
Things
work
in
in
AWS,
but
when
only
some
people
are
doing,
it
does
not
work.
So
this
is
kind
of
like
unofficially
like
what
is
already
happening
in
a
lot
of
the
AWS
client.
Instrumentation
is
definitely
happening
in
the
JavaScript
side
of
things.
C
I
think
the
Ruby
version
that
we
have
is
kind
of,
like
a
is
very,
very
heavily
based
on
the
JavaScript
implementation,
so
I
think
it
is
also
already
doing
this,
and
really
this
spec
PR
was
just
a
way
to
kind
of
formalize
that
this
is
what
you
should
do
so
that
any
new
stuff
that
crops
off
kind
of
follows
the
patterns
already
out
there.
But
it
did
lead
to
quite
a
bit
of
a
discussion
about
about
how
to
actually
handle
this
from
from
a
spec
perspective.
A
C
Okay,
all
right:
well,
we
will
see
what
we
could
do
with
it.
Then
we
will.
We
will
not
let
the
Tito
spring
break
Edition
go
go
over.
So
in
the
final
minute
there
is
a
PR
to
add
Cosmos,
DB,
attributes
scientific
conventions
and
the
end.
C
We
shall
now
move
on
to
our
agenda.
If.
A
I
find
such
a
thing.
D
D
So
we've
got
a
user
of
open,
Telemetry
and
Amazon
doing
x-ray
things,
and
they
have
an
issue
saying
the
ID
generator.
The
X-ray
ID
generator
should
be
split
out
from
the
X-ray
propagator,
which
I
I
think
maybe
I'd
cause
the
Ruckus
by
asking.
Why?
Because
I
don't
know
what
problem
that
solves,
I
think
the
thread
in
the
slack
had
concluded
that
it's
roughly
parity
with
the
way
that
other
sigs
have
done
it,
because
JS
and
I
think
maybe
python
have
have
split
them
up
into
very
small
libraries.
D
A
D
Resource
detectors,
I,
imagine,
are
the
things
making
API
calls,
maybe
that
need
the
X-ray
ID
I
I'm,
trying
to
time
box.
Maybe
our
discussion
about,
because
from
the
thread
from
the
thread
we
were
like
well,
Amazon
should
be
writing
this
and
then
they're
like
well.
If
we
wait
for
Amazon
to
write
this,
we
we
won't
be
able
to
do
what
we
need
to
do
so.
We'd
write
it
ourselves,
which
is
sad
so
to
try
to
frame
this
conversation.
What
can
we
do
with
the
code?
D
That's
in
Ruby
contribute
today
to
make
it
less
bad
I
don't
want
to
I,
don't
want
to
call
it
better
because
better
is
like
pilot.
Let's
pylon
features
like
let's
make
it
that's
bad,
what
the
implementation
that's
in
there,
what
pains
are?
Is
it
causing
and
there's
I
think
confusion
that
the
ID
generator
has
been
tucked
in
with
the
propagator
and
the
discoverability
of
the
ID
generator
itself?
D
I
think
it
would
be
weird
to
use
the
open,
telemetry
ID
generator
library
to
generate
x-ray
IDs
when
you
make
SDK
calls
like
I.
Don't
know
why
that's
not
in
the
Amazon
SDK.
D
Anyway,
so
we've
got
an
open
issue
to
split
out
the
ID
generator
and
and
a
PR
to
do
so.
My
hesitance
to
accept
the
pr
is
that
it's,
a
new
gem
and
and
Gem
more
gems
is
somewhere
between
a
linear
and
an
exponential
increase
in
maintenance
work.
D
D
Does
anybody
on
the
call?
No,
what
maybe
problem
would
be
solved
by
a
standalone,
ID
generator
yeah,
because
I
think
it's
only
discoverability
that
an
ID
generator
lives
inside
the
propagator,
and
it
confuses
some
people
that
why
do
I
have
to
get
the
propagator
when
all
I
want
to
do
is
generate
an
ID,
because
that's
where
we
put
the
logic
because
within
tracing
that's
the
only
place
that
cares.
C
A
D
And
I'm,
like
What's
the
foreign,
my
understanding
is
for
open,
telemetries
purposes
of
being
able
to
generate
x-ray
traces.
The
propagator
is
the
only
thing
that
cares
about
being
able
to
generate
an
ID
is.
B
B
To
shove,
start
emitting,
you
know,
headers
and
and
stuff
you
just
are
doing
the
math
of
I
think
that's.
The
implication
here
is
like.
After
the
fact
in
the
Lambda
layer
you
get
like
a
trace,
that's
been
generated
by
x-ray
and
they're.
Saying
like.
Can
you
make
this?
Look
like
an
Hotel
formatted
Trace
ID
was
I'm,
not
I.
Think
is
what,
and
so
in
that
case
you
don't
need
to
propagate.
You
just
need
to
do
the
translation,
math
right,
no
right
making
it
up.
Maybe.
A
D
My
understanding
is
that
whatever
is
in
front
of
a
Lambda
invocation
would
generate
the
initial
Trace
ID,
like
you,
got
a
some
sort
of
application
load,
balancer
sitting
in
front
of
your
lambdas,
something
comes
in,
is
web
request
coming
into
your
Lambda
that
load
balancer,
you,
you
would
say
I
want
X-ray
and
it
would
start
an
x-ray
trace
and
it
would
generate
the
trace,
ID
and
then
I.
Don't
know
what
the
Lambda
service
itself
does
for
tracing.
What
goes
on
inside
the
Lambda.
You
might
yeah.
B
F
B
You
could
install
an
otol
SDK,
you
couldn't
start
an
extra
SDK
and
there
are
certain
bits
within
your
Lambda
that
that
would
instrument
and
then
I
think
there
might
be
some
spans
that
are
like
you.
Don't
even
really
control
them,
I'm,
not
sure,
like
they
just
kind
of
get
generated
in
x-ray
format,
and
then
it
used
to
be
that
you
would
log
all
that
stuff
will
get
logged
to
cloud
Cloud.
B
B
And
so
yeah
I
don't
know
much
about
that.
One.
A
B
Tlbr
I
don't
know,
but
I
do
think
if
the
goal
is
I
just
they
just
want
to
do.
There's
some
place
where
they
don't
care
about
adding
headers
onto
requests
and
all
they
care
about
is
format.
You
know
a
bunch
of
digits
in
x-ray
format,
gotta
make
it
fit
into
otel
format,
or
vice
versa.
B
D
B
D
I
I
don't
think
that
open
Telemetry
Ruby
maintain
contrib
maintainers
are
in
the
business
of
having
an
x-ray
Trace
ID
generator,
that's
a
utility
that
the
world
can
use
to
generate
whenever
you
need
to
generate
a
trace.
An
x-ray
Trace
ID,
like
our
my
understanding,
is
that
the
only
reason
we
care
is
so
that
we
can
generate
Trace
IDs
in
an
x-ray
compliant
format,
and
the
propagator
is
the
only
thing
that
cares
about
that.
In
case
you
don't
have
something
ahead
of
your
Ruby
process.
E
D
An
incoming
x-ray
ID
with
an
incoming
x-ray,
Trace
header
and
you
want
to
be
x-ray
tracing
with
open
telemetry.
The
propagator
needs
to
know
how
to
generate
the
for
the
trace
ID.
It's
because
it's
creating
the
root
span,
so
it
needs
to
create
the
trace,
ID
and
then
propagate
that
out.
So
it's
it's
a
propagation
concern,
not
a
Ruby
World
utility
on
how
to
x-ray.
E
B
B
D
B
B
I,
don't
I
don't
have
enough
space
I,
don't
want
to
bike
shed
more
because
I
don't
have
the
right.
I
haven't
done
the
work
except
to
I'm,
proposing
as
I'm
saying
there
is
a
situation
where
this
would
happen.
I
don't
know
if
that
situation
exists
in
the
real
world
so
until
I
can
validate
that
it
can
like
I.
D
B
D
C
Know,
yeah
I
I
think
the
ID
generator
was
added
in
the
first
place,
because
Amazon
demanded
it
for
exactly
this
purpose
for
being
able
to
generate
face
IDs
in
the
X-ray
format,
and
it
is
slightly
adjacent
to
propagation
in
that,
when
you
are
creating
yeah.
C
D
Okay,
that
that
wording
has
has
me
recognize
yet
another
use
case
where
your
Lambda
is
not
propagating
to
something
else.
You
don't
want
to
propagate
x-ray,
so
that
other
call
like
calls
that
you
make
out
of
the
Lambda,
also
have
an
x-ray
trash
headers
so
that
whatever
happens
elsewhere,
participates
in
the
trace.
It's
just
I'm,
generating
with
Hotel
spans
that
are
going
to
get
sent
to
the
X-ray
service.
So,
therefore,
the
trace
IDs
on
these
fans
need
to
be
x-ray.
Flavored.
D
Well
and
that's
and
that's
what
I
mean
that
I
I
now
can
see
that
case.
I'm
not
propagating
I
just
need
these
spans
that
I've
made
within
Lambda
to
be
acceptable
to
when
I
ship
them
to
x-ray
x-ray
you'll
accept
them,
because
it's
a
valid
Trace
ID.
C
Okay
and
then
kind
of
the
second
thing
I
see
behind
all
this
is
just
like
more
than
anything,
I
feel
like
this.
This
issue
was
opened
due
to
some
level
of
confusion.
C
Isn't
you
know
the
why
something
this
live
in
the
exact
same
place
and
Ruby
as
it
does
in
JavaScript,
but
even
I?
Think
as
the
author
kind
of
points
out,
what's
going
on
in
other
cigs,
it
becomes
yeah
I
mean
there
there's.
Definitely
an
incongruence
between
JS
and
Java,
Java
and
python
seem
to
kind
of
agree
on
this
and
if
anything
else,
it
seems
like
Java
and
python.
C
Just
have
this
other
AWS
AWS
anything
package
where
they're
just
like
glomming
anything
AWS
in
there
other
than
the
propagator
and
essentially
moving
the
problem,
I
feel
so
I.
Don't
because
I
think
if
you're
really
worried
about
package
size-
and
you
only
want
the
ID
generator
but
you're
forced
to
take
the
AWS
everything
package,
you're
kind.
D
So
I
could
see
a
world
in
which
we
have
every
little
bit
of
like
Amazon
service.
We
could
ship
the
Amazon
console
and
say
Here's
the
resource
detectors
for
this
service
and
here's
the
research
detectors
for
this
service,
and
then
we
can
have
one
gem
that
depends
on
all
of
them
and,
oh
my
God.
That's
a
lot
of
work.
B
D
D
E
D
Shipping,
an
ID
generator
is
fine
like
like
it's.
We
we
make,
the
propagator
depend
on
the
ID
generator,
there's
a
standalone,
ID
generator,
and
then
we
have
to
document
for
people.
You
have
to
compose
these
yourself
and
you
have
to
know
if
you,
which
fits
you
want
to
use
enjoying
like
I,
could
just
trust
that
a
standalone,
ID
generator
Library
solve
some
problem.
I,
don't
understand.
The
problem
is.
E
C
After
having
this
entire
discussion
and
just
kind
of
peeking
at
this
PR,
it
does
look
like
it
at
least
brings
us
in
line
with
what
Java
and
python
are
doing,
and
AWS
will
forever
have
a
home
to
glom
whatever
they
would
like.
C
All
in
all,
like
I,
think
this
is
probably
acceptable
and.
F
C
Will,
at
least
if
nothing
else,
it
will
prevent
confusion
from
users
that
are
working
in
multiple
languages.
As
you
know,
the
package
that
contains
the
ID
generator
will
kind
of
be
you
know
analogous
between
between
most
of
the
things,
but
I
do
I,
definitely
aligned
with
the
spirit
of
asking
the
hard
questions
of
like
do.
We
need
this,
and
why
do
we
need
this
before
blindly
going
forward.
A
C
F
C
That
makes
sense,
I
think
actually
getting
an
answer
from
from
the
folks
proposing.
This
would
would.
C
D
C
Detectors
which
I
think,
if
we
did
have
them
I,
would
expect
them
to
go
into
this
newly
created
package.
The
extensions
package.
B
B
I
could
see
it
being
contentious
all
of
a
sudden
and
they
want
to
have
some
home
where
they
can
just
shove
arbitrary
things
into.
D
Which
Francis
brought
up
in
the
slack
red
like
to
compare
with
honeycomb,
came
and
said:
I
got
some
vendor
specific
stuff.
Should
I
put
it
in
Kindred
and
in
the
conversation
went
well.
This
stuff
is
sort
of
generic
implementation
could
be
beneficial
for
anybody.
This
one's
very
vendor
specific.
Maybe
you
own
that
which
is
a
reasonable
stance.
B
C
I
think
AWS
yeah,
I
I
definitely
think
there
is
a
foreign
is
blurry.
There's
a
gray
area.
There
I
think
the
one
thing
that
AWS
gets
to
use
to
their
advantage
is
that
they're,
a
cloud
provider
and
a
lot
of
people
use
their
junk,
and
this
somehow
helps
all
that
work
together.
A
little
bit.
C
D
A
A
D
C
Yeah
I
was
just
getting
through,
it
looks
like
Francis
has
been
having
some
level
of
conversation,
but
Jordan
would
be
good
to
get
a
few
more
eyes
on
it.
B
The
discussion,
oh
god,
I-
think
the
discussion
Rob
and
I
had
last
week
was
looks
good,
let's
see
how
it
dog
foods-
and
this
is
an
off
by
default-
config
option
that
people
can
opt
into
and
it
appears
to
be
an
improvement.
Generally
speaking,
may
have
some
changes
to
as
our
area
was
saying,
it
will
not
it
it.
It
has.
Basically
it's
changing
the
measurement
that
we're
doing
slightly,
so
we
should
be
thoughtful
about
if
we
were
to
like.
B
Let's
say
this
is
incredibly
more
performant
and
is
better
and
allows
all
this
stuff
like
even
then,
we'd
still
have
to
have
put
some
care
into
like
how
we
were
to
roll
us
out
into
the
default
where
we
left
it
was
I
think
the
idea
was
like
we
want
to
kind
of
get
this
merged
at
least
I
do
with
no.
You
know
someone's
not
even
really
working
on
tracing
so
so
that
you
know,
Ariel
has
Kentucky
effect
essentially,
which
is
maybe
it
should
go
the
other
way?
B
Maybe
he
should
be
dog
fooding
on
a
fork
or
something,
but
because
of
the
way
it's
being
merged
which
is
like
off
by
default.
In
the
combat
option,
it
just
felt
like
a
low
impact
way
to
move
the
ball
forward
here
and
I.
Don't
have
the
Practical
day-to-day
expertise
with
you
know
this
particular
portion
of
rack
to
really
be
able
to
give
more
feedback
than
that.
D
Think
it's
buying
a
feature
flag.
So
let's
get
it
out
there
so
that
people
can
run
it
in
prod
and
actually
see
how
it
dances.
C
Disabled
people
would
enable
this
disable
the
other
rack
stuff
so
kind
of
like
a
a
double
opt-in
process.
I.
D
C
The
key
point
being
that
that
it
is
opt-in
and
then
yeah,
then
GitHub
and
maybe
Shopify
or
whoever
is
is
bold,
could
kick
the
tires
on
this
thing
and
see
what
they
think.
D
Do
we
I
we
could
for
people
to
dog
food?
It
needs
to
be
out
there,
it's
easier
to
dog
food
with
a
gem
instead
of
a
branch.
Do
we
have
a
pattern
of
releasing
any
of
our
gems
core
or
contrib,
with
like
pre-release
addendums
to
the
version
number,
and
that
would
be
a
way
if
we're
like
nervous
about
this
thing,
if
we
prove
it
RC
or
pre-release,
on
the
rack
instrumentation,
that
has
rack
events
in
it.
So
now
anybody
can
choose.
That
would
be
the
double
opt-in.
C
C
Think
that
that
would
be,
that
would
be.
C
And
if
keeping
it
simple
goes
wrong,
then
we
can
kind
of
evaluate-
and
you
know
add
complexity
later-
to
make
up
their
mistakes
but
yeah
and
then
longer
term
like
if
this
is
kind
of
panning
out
and
working
for
people,
we
would
probably
eventually
just
say,
hey
we're
going
to
deprecate
the
other
stuff.
This
is
going
to
become
the
default.
You
can
opt
back
to
the
other
stuff.
If
you
want
to
for
a
limited
time
and
then
we
will
transition
fully
to
rec
events
once
everybody
is
dissatisfied.
C
C
I
have
a
good
deal
of
confidence,
because
I'm
pretty
sure
Arielle
is
going
to
run
this
at
GitHub
and
fix
anything
that
he
finds
with
it
there,
and
that
is
usually
a
pretty
good
trial
for,
for
some
software.
A
B
Yeah,
it's
really
just
a
matter
of
getting
approvals
on
it,
which
I
think
we're
all
a
little
hesitant
it's
a
little
chicken
and
eggy.
But
maybe
this
weekend
I'll
have
time
to
review
it.
I
don't
know,
maybe
maybe
we
might
get
asleep.
A
D
B
Okay,
yeah
I'll
I'll
make
an
effort
to
try
to
look
at
this
this
week,
but
I
just
I've
just
I've
almost
zero
time,
so
I
can't
make
any
promises,
but
I
do
think
we
should
get
some
approvals
that
are
real.
Approval
is
not
just
like
rubber
staffs
before
merging
yeah.
C
All
right,
yeah,
so
I
think
all
of
us
are
fairly
non-committal
here,
but.
B
C
B
Excellent
I
know
that
good
all
right
take
it
easy
all
good,
seeing
you
all
yeah.