►
From YouTube: 2020-09-24 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
B
I'm
not
sure
I
I
I'm
I
follow
like,
like
I
created
my
own
event
for
this.
I
don't
use
the
google
calendar
once
so.
Oh
man.
E
You
guys
have
the
google
group,
are
you
in
the
google
group
like
listserv
or
are
you
on
like
the
open,
telemetry
calendar.
E
E
Maybe
probably
there's
just
like
a
link
for
a
group
and
then
you
can
join
it
with
your
google
account
and
then
it
will
like
yeah.
B
Yeah,
I'm
not
sure,
I'm
not
I'm
not
the
one
who
has
having
problems
with
this.
So
like
amos
nathaniel.
What
what
link
did
you
guys?
Click
to
get
to
you
can
send.
F
F
B
B
Yo
that
one
is
oh
okay,
I
see.
So
if
you,
if
you
take
a
look
at
the
description,
one
is
actually
wrong.
The.
Where
is
the
actual
zoom?
B
If
you
see
the
disc,
if
you
see
it,
so
we
probably
should
change
the
description
right.
Oh,
I
see
yeah.
D
Yeah
is,
I
think,
oh
wait
was
was
owie
the
person
that
created
this
one
or
I
don't
know
if
o.a
has
the
ability
to
change
it.
B
B
F
B
All
right,
gotta
close
all
these
launching
zoom
windows,
nice,
if
you
guys,
could
fill
out
the
attendees
thing.
That'd
be
great,
so
we
know
who's
responsible
for
past
stuff.
D
Yeah
sure
I
can
I'll
I'll
take
the
notes
here.
Just
give
me
a
second
cool.
B
All
right
do
we
have
any
new
people
in
here
that
haven't
joined
a
python
sig,
yet.
B
Hi
yeah,
you
want
to
just
do
a
little
intro.
G
Not
sure
so
my
name
is:
will
I'm
a
software
engineer
at
aws
on
the
x-ray
team,
much
like
bontik
and
nathan,
I'm
just
joining
to
contribute
to
the
python
sdk.
Add
some
extra
compatibility
and
generally
just
contribute
as
much
as
I
can
so
I'm
excited
to
get
started,
learning
about
it
and
diving
into
the
repo.
B
Awesome:
okay,
cool
cool
nice,
great
I'm,
the
only
intern
nice
looks
like
b-h-a-u-tick.
C
Hey
yeah,
it's
me
I
am
bad.
C
I
like,
as
well
mentioned,
like
started
coming
up
to
the
python
sync,
so
we
hope
to
contribute
to
the
community
and
I
work
with
aws
extra
team
as
a
full-time
software
engineer,.
B
B
Awesome
nice,
do
you
guys
know
if
a
lolita
would
be
joining
the
python
6?
B
No,
I
could
say
no.
I
see
I
don't
know
I
I
talked
with
her
like
a
couple
days
ago
and
I
said
it
would
be
good
for
her
to
join
and
see
what
was
going
on.
But
I
guess
we
could
talk
about
that
offline,
anyways
cool,
so
basically
the
formats
of
these
cigs.
We
go
over
like
any
big
topics
that
people
want
to
talk
about.
B
You
know:
release
candidates
like
milestones,
we've
got
a
hit
and
then
we,
after
that,
we
just
go
through
the
you
know
open
pr's
and
issues
that
we
have
so
are
there
any
like?
I
guess,
for
the
people
who
have
been
here
before
any
open
issues
that
people
want
to
talk
about
like
big
topics
that
are
unrelated
to
prs,
or
you
know
open
issues,
I
guess
for
the
new
folks.
Is
everyone
able
to
find
like
stuff
to
do?
Are
they
like?
Is
anyone
lost
on
like
where
to
find
things?
B
This
is
like
kind
of
the
time
to
like
talk
about.
You
know,
pain,
points
that
you
guys
might
have
so.
B
B
Cool
okay,
so
if
you're
talking
about
specific
pr's
or
specific
issues
you
can
just,
and
if
you
want
us
to,
if
you
wanna
us
to
have
more,
you
know
eyes
on
it,
you
can
put
it.
You
can
link
the
actual
issue
or
pr
under
the
in
the
actual
google
doc.
Okay,.
F
B
That
yeah
and
then
we'll
we'll
get
to
it
during
the
meeting
so
sounds
good
cool.
So
right
now
alex
is
adding
this
ga
spec
compliance
matrix.
B
This
is
pretty
much
what
we're
using
to
see
that
you
know
python
is
ready
for
the
for
the
ga
release,
which
should
be
at
the
beginning
of
november
of
this
year,
alex
if
you
can
open
that
up
the
the
compliance
matrix,
real,
quick
yep.
So
this
is
this
is
generally
like.
You
know
a
good
summary
of
like
all
the
features
that
we
need
for
python
and
alex,
and
I
have
been
like
actively
updating
these.
B
So
what
whatever
you
see,
there
should
be
the
most
updated
stuff
pretty
much
like
for
us.
This
is
what
we're
gonna
be
focusing
on
until,
like
probably
the
end
of
october,
for
rga
related
stuff
so
like,
and
a
lot
of
them
are
pretty
straightforward.
It's
it's
simply.
B
Just
our
our
implementations
don't
comply
with
the
specs
so
and
since
the
specs
are
pretty
well
defined,
they're
usually
pretty
easy
changes,
so
we
created
a
bunch
of
issues
for
these
and
marked
them
as
required
for
ga
in
if
you
filter
for
those
and
a
bunch
of
them,
are
you
know,
labeled
as
good
first
issues
as
well?
So
you
know
if
you
guys
ever
want
to
like
you
know
like
if
you
don't
know
what
to
do
like,
we
don't
have
to
explicitly
give
you
stuff
like.
B
This
is
the
great
way
to
get
started,
and
it
will
definitely
help.
You
know
push
us
towards
this
first
milestone
so
yeah
we
can
go
over
the
specific
ones
later
after
when
we
talk
about
the
issues
so
cool.
D
Yeah
it's!
It
might
be
worth
just
mentioning
that
there's
a
there's,
a
required
for
ga
label
that
we've
tagged
all
those
issues
with.
So
just
there's
a
lot
of
stuff
to
do
here,
but
there
it
it.
They
should
all
be
linked
to
details
in
the
specs
so
that
you
get
a
pretty
good
idea
of
what
to
kind
of
how
to
how
to
tackle
the
work.
B
Yeah,
if
you
ever
want
to
like
work
on
them,
just
like
comment
on
the
issue
and
like
one
of
us
will
assign
it
to
you.
I
don't
think
you
guys.
Many
of
you
have
the
ability
to
assign
stuff
yourself
yet
so
yeah
cool,
so
I've
added
some
like
prs
and
stuff,
mostly
because,
like
these
are
pretty
old
and
we're
trying
to
like
kind
of
clean
up
the
old
pr's
that
we
have.
B
Hopefully
the
people
who
wrote
them
are
in
this
call
right
now.
But
if
you
go
to
the
issue,
the
first
pull
request
a85.
B
This
would
be
right.
This
is
diego's
very
old
support
for
config
files,
pr
it's
still
like
in
draft
mode,
and
I
think
there
is
something
holding
this
up,
but
I
don't
recall
what
it
was
so
diego.
Are
you
still
continuing
to
plan
on
working
this
working
on
this.
H
Well,
I
don't
have
any
reason
not
to
accept
for
the
fact
that
I've
been
working
on
stuff
that
is
required
for
ga
yeah.
So
this
this
one
I
I
think
it
would
be
a
good
thing
to
have
it
someday,
but
right.
C
B
To
work
on
that,
how
do
you
feel
about
like
maybe
we
can
just
close
this
one
and
like
if
you
start
working
on
it
again,
we
can
open
it
back
up
all.
B
Trying
to
clean
up
some
stuff,
so
cool
alex.
I
guess,
are
we
good
for
just
closing
this
then
and
then
just
reopen
it
when
we
start
working
on
it
again,.
B
B
D
B
It's
aaron
in
this
call
yeah,
hey
so
aaron.
It
might
be
a
long
shot,
but
do
you
still
have
like
contact
with
conor
or
anything
or
like
know
the
progress
of
this
exemplars
prototype.
E
No,
I'm
not
talking
to
conor,
but
I
mean
I
think
I
think
I
said
last
week
too.
It's
it
is
merged
into
the
spec,
it's
not
in
the
sdk
spec
yet.
But
it's
totally
like
you
know
something
we
are
supposed
to
support
but
yeah.
I
think
there's
spec
work,
but
basically
blocking
this.
B
E
Which
yeah-
and
I
think
I
think,
yeah
you
guys
all
know
that
it's
probably
google
is,
is
sort
of.
Like
the
I
don't
know
the
key
stakeholder
yeah,
the
one
who's
driving
this
yeah
yeah.
I
guess
we're
pushing
it
so
I
mean
I
can.
I
can
definitely
see
if
somebody
is
willing
to
work
on.
It
could
be
me
or
from
the
other
folks
that
are
actually
joining
our
team.
So
I'll
get
back
to
you
on
that.
B
Sure
should
we
just
wait
on
that,
like
do
you
recommend
just
like
opening
up
a
new
pr
for
this,
and
or
should
we
just
like
branch
off
this
one?
What
do
you
think.
D
Yeah
and
I
think
the
like
the
spec
uses
this
as
the
one
of
the
prototype
examples
right
like
I
don't
think,
there's
many
more
of
them,
so
I
mean,
even
if
we
mark
it
as
closed
it
doesn't
mean
that
the
pr
is
going
to
disappear.
So
yeah
like
that
that'll
still
be
there,
but
it'll
be
good
to
know.
If
there's
going
to
be
progress
on
it.
Otherwise,
every
time
I
look
at
the
bottom,
I
always
see
it
and
I
think
man
we
should
really
do
something
with
this
today,
yeah.
D
Okay,
all
right!
This
is
the
other
one.
B
Yep,
okay,
so
this
issue
is
kind
of
blocking
the
rga
release,
because
we
need
the
one
of
the
requirements.
Is
that,
like
our
api
surface
for
tracing
is
concurrent
and
these
tests
help
like
verify
that?
So
basically,
this
this
issue
is,
for
you
know,
adding
tests
to
you
know,
test
out.
Concurrency
related
use
cases
that
relate
to
the
batch
spam
processor.
B
B
B
So
if
anyone
has
any
ideas
or
like
experience
and
how
to
do
this,
you
know
it'd
be
greatly
appreciated.
If
someone
were
to
like
help
out
on
this
or
if
anyone
has
any
ideas
so.
B
This
open
pr
actually
is,
doesn't
address
the
actual
problem
and
the
person
who's
working
on
this
is
no
longer
working
on
the
project.
So,
as
for
this
pr
like,
I
think
we
can
close
it,
but
the
original
issue
is
still
still
exists,
so.
B
All
right,
this
is
diego's
global
error.
Hello
thing
I
still
haven't
taken
a
look
at
it,
but
I
guess
yeah,
I'm
just
you
could
just
like
assign
it
to
me
add
me
is
a
approver.
We
also
need,
like
some
other
people,
to
look
at
it
as
well
so
yeah,
but
I
guess
one
more
time
for
the
people
who
haven't
seen
this
before
d.a
diego
you
wanna.
Just
briefly
talk
about
it.
H
H
A
It's
all
right
as
a
small
question:
diego
does
this:
does
this
eat
the
original
error,
or
does
it
still
origin
letter
still
raised,
or
does
it.
A
H
Know
the
the
errors
are
not
raised
anymore,
they're.
A
A
H
Yes,
if
something
goes
wrong
during
the
handling
of
an
exception,
then
the
exception
gets
locked
but
not
raised.
I
mean
an
additional
message
shows
up
saying
something:
went
wrong
with
and
handling
this
exception.
Yeah.
B
Cool
nice
all
right,
so
next
I
think
we
have.
B
Nathaniel's
pr
for
the
id
generator,
this
one
is
pretty
straightforward,
but
you
could
talk
about
it
nathaniel.
If
you
want
to
bring
up
any
issues,
you
have.
F
Thanks
yeah,
it's
pretty
small.
The
idea
that
we
have
going
on
here
is
that
for
some
back-ends,
but
more
specifically
for
the
aws
back-end,
we
need
the
trace
to
be
of
a
specific
format.
It
can't
just
be
random.
Digits
like
it
is
right
now
we
need
it
to
involve
seconds,
otherwise,
our
back-end
service
is
just
going
to
throw
it
away.
So
this
first
pr
just
follows
the
model
of
the
previous
java
and
js
open
telemetry
sdks,
which
have
the
notion
of
an
ids
generator
for
the
tracer
provider.
F
But
yeah
so
much
to
it.
I
just
made
it
as
a
draft
right
now
because
the
tests
are
failing.
B
Cool
yeah
so
for
your
reasoning
behind
like
the
putting
it
not
putting
in
the
same
full
file,
you
can
just
like
write
your
reasoning
in
the
issue
and
then
we
can
address
that
there,
okay,
yeah,
that's.
E
A
good
idea,
two
quick
questions
is
there:
is
this
in
the
spec
as
well,
or
is
this
just
something
that
we're
copying
from
java.
F
I
don't
think
it's,
I
haven't
checked
the
spec.
To
be
honest,
I
don't
know
if
it's
in
there,
but
I
can
check
that
and
reference
it.
If
it
isn't,
I
reference
reference
it
if
it
isn't.
If
it's
not,
then
maybe
I
can
add
it
there
too.
E
Yeah,
it
might
be
worth
brazilian
issue
there.
F
E
Cool
yeah-
and
I
think
my
other
question
was
there's:
what
is
it?
There's
a
trace
in
tracing
there's
a
sampler
that
it
works
by,
like
it
kind
of
it,
samples
which
one
is
it
called?
E
E
Yeah,
it
kind
of
it
kind
of
depends
on
the
numbers
being
random,
or
at
least
like
a
portion
of
them
being
random.
If
I
understand
how
it
works
correctly,.
E
B
Like
the
trace
id
and
like
sees
like
the
first
bits
or
something
like
that,
and
then
like
compares,
if
it's
less
than
like
a
threshold,
it's
like
a
basically
zero
to
one
random
distribution
based
off
of
the
trace
id
so
yeah.
That
is
a
good
point.
E
Okay,
so
yeah,
you
might
want
to
just
check
that.
However,
you
guys
generate
ids
for
the
aws
thing.
It
doesn't
break
the
assumptions
of
that
because
I
think
basically
the
idea
is
that
all
if
like
a
parent
spend
is,
is
sampled,
then
all
child
spends
will
be
assembled
as
well
and
it
works
based
on
the
the
child
spends
having
like
a
smaller
smaller
value
than
the
parent
spin
id.
I
think
I'm
not
super
across
that,
but
yeah,
that's
all
I
was
gonna
say.
F
B
Awesome,
okay,
cool!
I
have
a
pr
that
I
would
like
to
get
some
eyes
on,
so
basically
we're
we're
trying
to
get
metric
collection
as
part
of
the
instrumentations
that
we
have.
B
Prototype
for,
like
the
requests
library,
we
want
to
get
certain
metrics,
such
as
like
the
duration
of
the
requests,
as
well
as
like
the
count,
and
basically
what
I'm
doing
is
just
you
know,
adding
some
api
methods
to
the
base
instrumenter
to
be
able
to
enable
metric
automatic
metric
collection
and
users
would
have
the
up.
The
you
know
have
the
option
to
export
it
if
they
want
to,
but
this
is
purely
just
like
the
collection
portion
of
it.
B
Yeah
just
need
a
couple.
More
just
need
some
more
eyes
on
this.
B
It's
kind
of
large,
so
yeah,
if
you
have
any
questions,
just
feel
free
to
comment
on
the
the
pr
or
just
ping
me
on
gator.
B
I'm
trying
to
push
for
this
because
I
don't
know
microsoft
really
needs
this.
So
yeah
be
great.
If
we
get
some
reviews
on
this.
D
Yeah,
I
forget
someone
on
the
call
could
pick.
This
up
looks
like
aaron's
already
approved.
E
The
drink
yeah
and
I
only
have
like
I
mean
I-
I
talked
everything
over
layton
in
the
discussion
there,
the
only
the
only
thing.
I
really
think
that
from
like
a
design
perspective,
we
need
other
opinions
on
is
this
makes
the
request?
Instrumentation
depend
on
the
sdk,
so
I
don't
know
if
that's
an
issue
or
if
you
have
any,
if
you
guys
have
any
thoughts.
B
I
think
like
me
personally,
I
didn't
want
to
take
the
dependency
but
like
after
your
suggestion
of
like
purely
just
recording
the
the
metrics
like.
If
we,
if
we
change
our
metrics
api
to
like
you
know,
not
have
to
do
that,
funky
thing
that
we're
doing
where
we
have
to
specify
the
actual
class
in
order
to
construct
it.
We
actually
don't
need
to
take
a
dependency
on
the
sdk
anymore.
I
think
so.
E
Yeah,
except
for
the
are
you
saying,
if
you
remove
the
push
controller
thing
too,
and
just
yes.
B
I
I
already
removed
it,
I
already
removed
it
actually
yeah.
So,
like
I
agree
with
your
point
like
originally
like,
I
only
wanted
to
do
the
collection
part,
and
so
I
think
that's
actually
a
good
advocate.
It's
like
instrumentations,
like
it
shouldn't.
B
E
Cool,
I
mean
yeah.
I
think
we
already
have
an
issue
open
for
that
just
for
for
like
changing
the
the
methods
for
creating
the
instrument,
so
it
won't
depend
on
that.
So
yeah.
B
Cool
all
right,
we're
halfway
through.
Does
anybody
have
else?
Have
any
other
pr's,
I'm
taking
a
look
at
some
of
them?
That
I'll
like
have
some
open
discussion
like.
A
Yeah
I
have,
I
have
one
one:
zero.
Three,
six,
it's
title
is
automatic
exposure
provider
set
up
right.
A
So
we're
waiting
for
the
spec
to
clarify
how
this
should
work
and
a
couple
of
weeks
ago
the
proposal
was
merged
and
the
proposal
says
hotel
underscore
exporter
can
take
a
comma
separated
list
of
exporter
names
and
and
that's
how
the
sdk
or
or
an
implementation
should
automatically
configure
a
trace
or
a
metrics
explorer.
A
So
this
pr
does
not
implement,
does
not
support
matrix.
Yet
it
starts
with
trace
exposure,
but
it's
a
bit
more
complicated
than
it
sounds
because
our
spam
processors
and
exporters-
sorry,
our
exporters,
don't
have
a
unified
api.
Some
take
a
service
name
as
in
in
an
argument,
some
don't
so
so.
There's
some
conditions
depending
on
which
exporter
the
user,
selects
then
how
to
instantiate
the
processor
or
or
the
exporter
itself.
A
For
example,
datadog
exporter
uses
a
special
processor,
so
there's
some
some
checks
around
that,
but
other
than
that
this
confirms
to
the
spec.
So
if
someone
could
take
a
look
also,
while
I
was
working
on
salary
instrumentation,
I
realized
this
is
actually
required,
because
without
this,
if
someone
uses
a
telemetry
instrument
command
to
start
a
salary
worker,
it
it
crashes
or
doesn't
work.
If
I
remember
correctly,
because
the
tracer
implementation
is
not
set
up,
so
salary
tries
to
use
the
api
classes
and
that
composes
some
errors.
B
D
Is
part
of
ga
yeah,
I
I
don't
know
if
the
rest
of
it
is
so
I
I
do
wonder
if
yeah
I
guess
we
can,
we
can
kind
of
review
how
how
this
works,
but
would
you
be
open
a
way
to
kind
of
splitting
off
the
hotel
exporter
like
the
parts
that
are
part
of
the
the
spec
out
of
this
pr
and
maybe
make
it
easier
to
review
and
get
those
get
those
merged.
A
Sure
yeah,
so
I
try
to
make
it
so
because
we
are
in
addition
to
exporter,
we
also
dynamically
instantiate,
spam,
processor
and
and
other
things.
So
maybe
I
can
actually
just
remove
support
for
all
those
things
and
just
for
now
hardcode
using
a
single,
a
specific
like
batch
processor
everywhere
and
just
implement
the
exposure
thing
for
now,
and
we
can
look
at
the
other
things
later.
D
I
guess
I
have
a.
I
have
a
pr
also
that
I
just
want
to
bring
to
attention.
It's
still
work
in
progress,
and
I
still
have
to
get
to
test
passing,
but
so
parent
used
to
be
passed
in
as
either
a
span
or
a
span
context,
and
there
was
a
change
that
was
merged
earlier
this
week
to
the
spec
that
changes
that
all
together
to
only
be
propagated
as
a
context
instead.
D
So
this
is
changing
our
apis
for
starting
the
spans.
So
I
I
did
get
started
on
this
work,
so
this
is
more
of
a
fyi.
This
is
coming
down
the
pipe.
It's
not
it's
not
ready
to
be
reviewed
or
anything
yet,
because
the
tests
are
still
failing,
but
yeah.
I
don't
know
if
anybody
has
any
questions
on
this.
B
Be
merged
for
a
while,
so
it's
good
that
you.
D
Proactive
with
it
so
there's
also
now
talks
of
if,
if
the
spans
or
the
span
context
will
disappear
so
we'll
just
we'll
just
go
with
the
plan
that
they're
still
around
for
now,
but
yeah.
So
that's
that's
just
something
to
be
aware
of
the
other.
Pr
that
I
opened
that's
a
work
in
progress
is
the
support
for
otlp
v05.
D
I
I
open
the
pr
just
to
see
what
the
delta
would
be
as
far
as
what's
what's
broken,
there's
a
bunch
of
stuff,
that's
broken
as
it
turns
out,
because
the
metric
spec
changed
pretty
significantly
in
the
latest
version
of
otlp.
So
I
think
diego
is
going
to
be
taking
this
this
pr
on
I'm
already
working
on
it.
So
yeah,
I'm
already
assigned
right.
H
C
B
Nice
all
right
does
anybody
else
have
any
other
pr's
that
they
want
to
talk
about.
B
I'm
seeing
one
here
with
the
might
want
to
talk
about
is
a
most
in
the
call
right
now.
B
Yeah,
so
the
one
that
is,
the
you
know
ad
required
yeah.
B
I
So
I
think
we
decided
on
using
a
private
class
for
spam
instead
and
then
just
using
new
type,
two
for
for
type
as,
like
the
public
api.
B
I
I
think
it
was
aaron
who
suggested
that
if
I'm
not
mistaken,
yeah.
B
Nice
yeah,
so
I
feel
like
this
is
a
pretty
good
way
to
do
it.
It
requires
like
much
less
code.
I
feel
and
also
like.
B
I
think
the
the
general
thing
is
that
like
are
we
going
to
be
doing
this
for
all
classes,
I'm
leaning
towards
yes,
but
like
if
that's
the
case
like
yeah,
we
should
have
like
the
least
amount
of,
like
you
know,
invasive
code
as
possible,
so.
E
I
did
have
one
other
question
on
this,
and
that's
like
so
is
this
only
for
sdk
spans
or
like
all
these
things
in
the
sdk
like
what
about
all
the
like
trace,
trace,
api
subs
or
the
default
default,
tracer
and
stuff
like
that.
B
I
think
it's
only
for
like
the
the
sdk
classes,
I
I
think
it's
purely
for
like
we
don't
want
users
to
call
the
constructors
and
I'm
not
sure
if
there's
much
value
for
them,
working
with
like
the
the
default
default
objects
so
like,
I
don't
think
we
care
if
they
instantiate
the
default
spans
like
they're,
using
it
up
to
their
discretion.
E
Yeah
that
makes
sense.
I
also
kind
of
thought
the
same
thing
about
the
sdk
spin
like
if
they
instantiate
one
it's
like.
No,
it
doesn't
really
do
anything
right,
so
I
mean
I
guess
I
was
kind
of
like
it's
in
the
spec,
but
I
feel
I
honestly
feel
like
maybe
when
it
was
written,
it
was
just
kind
of
like
must
not
allow
you
know
what
I
mean
like
I
don't
know.
I
don't
know
like
how
important
it
really
is,
at
least
at
least
to
me.
E
B
Yeah,
I
think
in
general,
it's
like
there's
a
lot
of
things
that
the
users
can
do,
and
I
know
we
just
have
to
be
consistent
on
like
what
we
force
them
not
to
do
and
like
how
much
discretion
we
we
give
them.
B
I
think
this
is
a
safe
like
not
like
the
pros
are
are
valid
enough
to.
You
know
justify
the
work
that
we
need
to
do
for
this,
but
I
wouldn't
expect
us
to
do
this
for
like
every
single
class
that
is
possible,
especially
the
ones
where,
like
we,
don't,
really
expect
the
users
to
interact
with
so
yeah,
okay,
okay,
cool
right,
okay!
So
that's
that
for
that
all
right,
anyone
else
have
any
other
prs.
B
They
want
to
bring
up
all
right
so
for
all
the
open
pr's
that,
like
are
not
brought
up
today,
we're
just
going
to
review
them,
as
is
so
it's
just
meaning
like
this
there's
not
any
impeding
discussions
that
might
block
it
so
cool
all
right.
Now
we
can
move
up
to
on
to
open
issues.
We
have
20
minutes
left.
B
A
A
But
if
you
follow
the
public
dogs
and
set
it
up
like
that
and
it
doesn't
work
turns
out,
the
resource
class
has
a
factory
method
create,
and
this
has
been
implemented
inside
the
create
method,
but
not
in
it.
So
what
the
I
think
we
got
sidetracked
here
about
whether
we
want
in
it
whether
we
want
create
or
not
or
if
create,
should
cache
resources
or
not.
I
think
that's
probably
slightly
different
discussion.
A
This
is
more
about
either
a
resource
and
it
should
do
the
same
thing
and
detect
those
environment
variables
and
use
them,
or
we
should
update
the
documentation
to
say
that
you
should
use
create.
Instead
of
calling
the
clause
directly.
E
A
E
Yeah
I
updated
the
or
added
some
sdk
resources.
I
added
some
documentation
in
the
resource
it's
like
in
the
api
documentation,
so
it
should
be,
it
should
say,
to
create
with
dot
create
but
yeah.
I
agree
that
examples
need
to
be
updated.
Then
I
also
I
wanted
to
just
offer,
like
I'm
kind
of
in
favor,
of
getting
rid
of
the
factory
method,
because
to
me
the
the
caching
doesn't
make
much
sense.
A
A
I
think
they
might,
they
might
be
referring
to
resource
detection.
Maybe
maybe
detection
can
be
slow.
Maybe
it
needs
to.
Let's
say
if
it's
running
in
kubernetes
it
might
call
some
communities
api
to
figure
out
what
the
cluster
like
metadata
about
the
cluster
or
something
maybe
I'm
just
guessing.
E
B
Yeah
just
reading
this,
it
makes
it
seem
like
like
if,
if,
if
someone
were
to
try
to
create
a
resource
with
like
an
attribute
map
that
has
already
been
created,
I
think
that's
that
might
be
useful.
You
know
yeah,
like.
E
Yeah
I
mean
I
left
this
comment
here.
I
think
the
other
side
of
that
is
like
if
you
look
at
this
code
right
when
they
call
resource.create
other
tribs,
then
that's
going
to
be
cached
forever
and
then,
when
they
call
dot,
merge
once
that
merged
one's
gonna
be
cached
forever
and
then
like
every
single
time
that
you
do
that.
B
Actually,
like
it
really
depends
on
what
they
mean
by
caching,
so
I
guess:
should
we
just
raise
this
in
the
specs
to
get
a
clarification
first
before
making
a
decision
on
whether
or
not
we
want
to
remove
it
or
not.
B
B
All
right,
so,
when
I'm
looking
at
the
issues
and
I'm
sorting
by
the
required
for
ga,
there
is
a
bunch
that
we
need
to
get
done.
The
good
thing
is
a
lot
of
them
are
assigned
to
people
already.
That's
that's
pretty
good
is.
Is
anyone
here
like
looking
for
stuff
to
do
or
anything?
If
not,
I
don't
know
alex,
do
you
think
there's
much
value
in
going
through
them
individually
or
what.
D
E
B
Yeah
not
much
room
for
discussion.
Most
of
them
are
pretty
straightforward.
We
just
need
people
to
pick
them
up.
As
you
know,
people
have
time
not
much
left
for
discussion
really,
unless,
if
someone
has
some
issues
that
they
want
to
talk
about
that,
they
don't
really
know
what
to
do
with.
I
Addre,
like
some
questions
to
clarify
which
kind
of
wanted
with
this
issue:
it's
the
create
coding
samples
for
span
outline
and
context
behavior.
So
is
the
idea
and
I'm
guessing.
This
is
going
to
be
like
part
of
the
runnable
examples.
So
it's
the
idea
that
we
just
like
have
a
bunch
of
these
situations
and
we
want
to
capture
the
context
capture
every
context
in
each
of
the
situations
and
then
kind
of
log,
the
the
baggage
that
for
each
context,.
B
Yep,
that
would
be
great,
like
just
make
sure
yeah,
like
the
just
make
sure
the
we
want
to
just
verify
that
what
is
expected
is
actually
happening
and
a
great
way
to
do.
That
is
just
logging
exactly
what
you
said
like
what
exists
in
the
current
context.
B
What
the
current
spans
context
is
what
the
what
the
the
parent
of
that
spans
context
is.
This
issue
was
brought
up
like
in
the
maintainers
meeting,
where
they
they
defined
a
bunch
of
behavior
with
the
related
to
context,
and
in
order
to
like,
I
guess,
bring
up
these,
this
feature
set
it'd,
be
good
to
just
have
examples
for
them,
so
you're
trying
to
think
of
like
in
the
in
the
eyes
of
like
the
customer
like
if
I
was
a
customer
and
be
like.
Oh,
I
want
to
light
up
this
scenario.
B
B
End
user,
rather
than
us.
D
Cool,
unless
anybody
has
any
other
issue,
we
can
call
us
early
and
get
everybody
back.
Nine
minutes.
B
B
Wasting
people's
times
right,
yeah
all
right
all
right!
Well,
nobody
else
has
any
other
topics.
Then
we'll
see
you
guys
next
week
sounds
good
all
right.