►
From YouTube: 2021-07-09 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
B
A
B
Ugly,
well
I
mean
I'm
I'm
messing
around
with
that
pr
right
now
trying
I'm
creating
functions
for
any
any
question
mark.
You
know
the
best
thing
ever
like,
let's
put
an
extension
function
on
any
possible
thing,.
B
Oh,
that
two
string,
you
might
be
able
to
do
a
dot.
It's
less
cool,
though.
B
B
B
I
don't
think
I
definitely
don't
need
to
merge
the
semantic
convention
update.
I
think
that's
just
that's
just
there
for
the
next
release.
So
don't
bother
to
worry
about
that
one.
So
we
can
work
out
the
formatting
like
that
formatting
issue.
A
B
Yeah,
that's
like
that
was
the
big
question.
I
mean
there's
no
support
for
this.
Unless
you
do
it
explicitly
in
the
sdk
right
like
it's
not
going
to
be
in
the
agent
there's
not
going
to
be
any
support
for
doing
this
in
the
agent
or
anything
like
that
right,
yeah,
there's!
I
guess
it's
not
the
agent,
it's
the
api
yeah.
I
actually
that
fields
thing
is
a.
I
think
it's
I
don't
think
it's
spec.
B
C
B
B
Yeah
clearing
out
well
the
javadoc.
Basically
just
parrots-
was
in
the
spec,
but
there's
also
this
thing
about
being
able
to
optimize
for
a
single
pass
through
collect
through
a
collection
or
an
iterable.
A
B
A
B
A
B
C
A
B
B
A
B
Just
I
always
run
the
build
before
I
push
anything
anyway.
So
that's
just
you
know
I
just
got
in
the
habit
hey.
I
wanted
to
talk
about
pr
3369,
real
quick,
which
I
believe
both
of
us
were
completely
too
chicken
to
respond
to,
and
I
finally
decided
to
dive
in.
B
A
A
B
Still
haven't
I
I
added
an
additional
comment
this
morning
about
it
so
yeah,
I'm
just
not
sure
what
to
do
like.
I
would
like
him
to
get
it
like
get
it
through
I'll
get
it
working
respond
to
all
the
comments
we'll
get.
A
I
mean
definitely
an
emergent
clean
up.
The
I
don't
know.
I
guess
it's
private.
Maybe
that
is
still
an
option,
but
I
think
it's
reasonable
for
the
computer
to
finish
that
up,
yeah
yeah,
I
mean
if
they
disappear,
then
we
have
to
though
so
that's
one
thing.
It's
true.
They
haven't
disappeared
yet,
but
that's
always
an
option.
If
someone
did
right,
yeah.
C
B
I
have
no
idea
what
that
one's
about,
but
yeah
I
haven't
had
a
chance
to
look
at
it
at
all.
B
A
A
B
A
A
B
By
the
way,
okay,
but
I
did
want
to
talk
about
so
in
the
meeting
this
morning,
nikita
and
trask-
and
I
did
talk
about
the
fact
that,
right
now,
when
we
cut
a
release,
we
end
up
with
all
of
our
documentation,
not
with
the
proper.
B
On
the
proper
tag,
which
is
a
little
bit
of
an
annoying
experience,
so
nikita
had
proposed
a
slight
a
new
workflow
that
feels
like
it's
a
little
bit
manual.
I
don't
know
how
much
of
it
we
can
automate
nope,
but
I
wondered
whether
you
wanted
to
try,
because
it's
a
lot
simpler
in
our
case,
and
we
don't
have
to
worry
about
the
examples,
because
we
haven't
changed
any
of
the
apis
that
the
examples
are
using.
B
So
the
examples
is
less
much
of
a
less
of
an
issue
when
you're
dealing
with
things
that
are
stable,
but
I
wondered
if
you
were
interested
in
trying
out
that
his
proposed
workflow
for
this
release,
just
to
see,
if
there's
any
stuff,
that
we
didn't
account
for
for
things
that
just
don't
end
up
working,
I
mean
the
only
part,
that's
a
little.
That
is
a
little
we
have.
I.
B
B
A
B
B
Is
there
something
about
the
main
branch?
Well,
we
have
to
get
the
tag
onto
the
main
branch
right,
because
otherwise
the
version
version
versioning
won't
work
right.
That
tag
has
to
live
on
the
main
branch
history,
because
we're
going
to
make
that
release
branch
right,
but
but
but
for
what's
it
called
the
that
netflix
module
nebula
for
the
nebula
plug-in
that
tag
has
to
be
on
somewhere
on
the
main
bench
history
for
it
to
pick
it
up.
B
B
A
Right
yeah,
I
don't,
but
there's
nothing
to
tackle
in
the
main
breaks
with
this
proposal.
That's
the
problem,
like
you
have
to
tag
the
release
branch.
I
think.
B
Right,
you
tag
the
release
bench,
but
then
you
have
to
but
then
that
the
last
commit
on
that
release.
Branch
has
to
be
cherry,
picked
onto
the
main
branch
or
you
won't
pick
up
the
tag
and
if
you
do
cherry
pick
that
onto
the
main
branch,
since
that
shaw
has
the
tag
on
it,
then
it'll
be
picked
up
in
the
history
and
nebula
will
work.
B
That's
the
only
part
that
I'm
not
sure
about
that
seems
squirrely
to
me,
but
I'm
pretty
sure.
C
A
A
B
B
Work
doesn't
work
like
it
literally
just
blows
up
like
it.
Won't
it
won't
even
build
period
if
you
try
to
put
android
and
the
android
plug-in
and
the
nebula
plug-in
in
the
same
project
so
yeah,
I
looked
through
their
code
to
try
to
figure
out
how
it
all
worked,
and
I
think
it
had
to
be
in
that
history
before
you
pick
it
up.
Okay,.
B
I
mean
oh,
this
is
something
I
do
want
to
get
merged.
There's
a.
There
is
an
outstanding
pr
to
update
the
change
log
which
never
got
reviews
from
like
two
weeks
ago,
and
I
updated
it
today
so,
okay
that
one
we
want
to
get
merged
and
then
updated.
B
A
B
A
B
A
B
A
A
So
that
there's
one
in
this
pr,
but
so
do
you
remember
the
one?
I
guess
it's
a
method
called
attributes
which
switches
to
an
application
right.
A
B
You
wanna
put
a
pr
in
real
quick,
like
I'm
gonna,
be
around
for
at
least
another
three
hours,
so
they
won't
be
working
actively.
But
if
you
wanna
put
a
pr
to
clean
that
up,
I'm
happy
to
to
approve
it
before
we
release.
A
A
A
B
Tests
with
suppressed
warnings
on
them
yeah.
Well,
so
if
you
disable
error
prone
locally,
then
you
don't
get
the
deprecation.
It
won't
fail
to
build
for
deprecation
anymore
you'll,
just
get
warnings.
Oh
really.
I
was
messing
around
with
today
when
I
was
deprecating
the
aws
id
generator,
because
I
was
like
oh.
A
B
Rid
of
the
errors,
and
it
would
let
me
let
me
do
a
build
and
it
showed
a
warning.
I
don't
know
all
that
stuff
is.
A
B
That's
true,
I
mean
they're,
really
real
annoyances.
If
you
want
to
keep
the
test
there
and
your
test,
like
the
real,
the
real
pain
was,
the
jmh
test
was
the
jmh.
Where
I
had
to
then
I
couldn't.
I
had
the
import,
and
so
I
had
this.
You
know
the
annoying
issue:
yeah
yeah,
where
you
have
to
fully
qualify
yeah
yeah,
that's
the
worst.
A
B
B
B
B
B
Yeah,
it's
a
lot
of
a
lot
of
question
mark
question
marks
in
a
row,
but
that's.
B
A
B
A
Yeah
is
it
that
hard
to
read
like
I,
I
don't
find
it.
I
don't
find
it
yeah
right,
like
my
life,
is
reading
java
code,
so
it's
like
once
this
is
java
coded
my
brain
handles
this
much
easier
than
any,
even
cockpit
or
anything.
I
think.
B
So
I,
when
we
get
and
then
when
we
get
whatever
the
multi-line
multi-line
strings
in
java,
whenever
that
came
in
15,
whatever
they
put
the
14
whatever
they
put
that
in
so
in
17
we'll
be
able,
we
could
write
a
dsl
in
multi-line
string
language
and
have
it
all
I'm
sure
that
would
not
be
a
problem
at
all.
Right,
just
put
a
bunch,
a
bunch
of
triple
coated
strings
with
a
whole
bunch
of
tables
in
it.
B
B
Was
there
anything
else
interesting
today?
What
did
we
talk
about?
We
talked
more
about
the
metrics
prototype
and
the
exemplars
and
prometheus
issue.
A
B
B
B
What's
closure
all
the
way
down?
Yeah?
Oh,
I
think
nikita
wants
to
start
a
closure.
Sig
all
right
closure
closure,
sdk
for
open
telemetry.
B
A
B
A
way
to
learn
closure,
I
mean
I've.
I've
been
really
itching
to
have
a
project
to
pick
that
up
and
learn
how
to
do
it,
but
really
have
you
seen
closure
before?
Oh
yeah,
absolutely
I've
attended
many
talks.
I've
read
I've
watched,
rich
hickey,
talk
about
it
tons
of
times
and
the
thing
that
I
like
about
it.
So
I'm
not.
I
wish
I
could
have
closure
without
the
lisp.
B
I
wish
I
could
have
closure,
but
with
curly
brackets
syntax,
but
I
love
the
fact
that
everything
is
immutable
by
like
out
of
the
box
like
you
get
no
mutable
data
structures,
and
that
is
wonderful,
and
that
is,
and
so
this
that
more
strictly
functional
style
is
what
I
would
that's
why
I
would
like
to
like
to
use
it.
A
Yeah
I
in
one
previous
company,
we
had
an
acquisition
where
the
code
base
came
in.
It
was
like
this
huge
closure
codebase,
and
so
I
just
was
curious
and
I
looked
at
one
file.
I
think
I
was
like.
Oh
my
gosh.
I
have
no
idea
what's
going
on
here.
Let
me
close
this
file
and
just
go
away
forever.
So
you
closed
the
file
and
then
we.
B
B
B
Anyway,
I
don't
think
there
was
too
much
more
there's
a
lot
of
metrics
a
lot
of
metric
stock.
B
Yeah
I
I've
kind
of
gotten
lost
like
for
a
while.
I
thought
I
understood
what
josh
mcdonald
was
proposing,
and
now
I
don't
understand
anymore
at
all,
so
we
spent
the
metrics
today
arguing
about
or
discussing
views
and
pipelines,
and
so
riley's
proposal
is
that
so
right
now
the
current
java
sdk,
you
basically
get
one
pipeline,
which
means
you
have
one
exporter
you
have
one
set
of
aggregators.
B
Every
instrument
is
in
that
same
pipeline
and
riley's
proposal
is
that
you
can,
when
you
construct
your
your
meter
provider
implementation,
that
you
can
set
up
multiple
pipelines,
each
of
which
gets
its
own
exporter,
its
own
configuration
essentially
like,
basically
you
and
and
it's
the
own
set
of
views,
and
then
you
by
defining
your
view,
selectors
you
can
make
some
instruments
only
go
to
one
pipeline
or
another
pipeline.
A
A
A
A
B
I
have
a
I've,
been
developing
a
theory
about
this,
oh
by
the
way,
the
merging
my
I
think
I
just
got
another
failure
of
yeah
of
the
stupid
rate
limiter
test.
My
theory
is
that
there
are
actually
very
few
people
who
have
enough
expertise
to
feel
confident
to
chime
in
on
things
when
we
were
just
talking
about
the
basics
of
tracing,
I
think
there
were
a
lot
of
people
who
at
least
the
harder
stuff
now
we're
getting
into
the
harder
things
and
people
don't.
A
C
A
B
Strategies
have
no
idea
about.
Google
must
also
sample
like
crazy,
like
they
must.
So
I'm
surprised
that
google
folks
haven't
chimed
in
much
yeah.
I
don't
know.
B
A
Yeah,
oh
at
least
x-ray,
when
I
have
my
internal
one.
So
it's
still
okay,
but
I
think
what
I
mostly
and
I
think,
even
with
john
paul's
work
on
the
otp
for
remote
sampling,
and
I
always
imagine
like
the
configuration-
should
be
very
similar
to
environment
variables
like
that
composes
samplers
also,
and
so,
if
one
of
those
remote
samplers
wants
to
use
right
domain,
then
if
we
just
don't,
have
it
anywhere
in
respect,
they're,
just
sort.
B
Oh,
I'm
so
annoyed
by
this
this
rate,
the
throttling
logger,
why
it's
it's
failing
like
crazy
now?
B
Is
it
always
on
my
quest
or
is
it
any
us?
I
thought
I
saw
ubuntu
11
today.
A
A
Yeah
I
mean
obviously
that's
gonna
have
some
factors,
so
that
should
motivate
us
to
fix
the
flakiness
yeah,
but
if
we
ever
find
something,
that's
unfixable,
but
if
we
see
a
pattern
like
it
only
happens
on
macos
or
something,
then
of
course
there's
also
workarounds
like
we
don't
even
have
to
run
all
these
jvms
on
macbooks.
I
think
I
noticed
that
after
the
fact
deciding
is
there
a
point
to.
B
This-
and
you
know
the
one
we
probably
should
never
run-
is
mac
os
java
11,
because
that's
probably
what
almost
everyone
is
developing,
but
it's
locally,
so
it
always
it'll
allow
like
that
always
passes
locally.
So
why
do
we
need
to
run
it
on
ci,
like
everyone
just
uses
that
as
their
local
implementation
right.
A
A
Yeah
so
which
is
okay,
I
mean,
I
don't
think
we're
expecting
the
17
build
to
fail
by
itself,
it's
sort
of
a
nice
to
verify
build
we
have,
but
if
someone
was
to
see
a
17
failure,
unfortunately,
that
flag,
wouldn't
by
itself
just
work
like
the
other
ones,
do
right,
which
is
not
going
to
be
a
problem
in
practice.
I
think
something
that
could
happen.
Yeah
yeah,
no
fair
enough.
A
A
A
C
B
A
B
A
I
mean
monotonic
clock
itself
has
an
implementation.
Detail
is
a
good
one,
I
think,
and
then
the
clock
interface,
I
probably
would
have
split
it
into
clock
and
ticker
like
guava
causation
or
something
the
monotonic
thing
and
like
having
two
interfaces
would
have
been
clear.
That's
probably
what
I
would
have
done.
Yeah.
A
A
B
A
A
A
C
A
B
Oh
man,
I
always
make
the
mistake
of
trying
to
look
into
these
jdk
classes
and
understand
what
the
crap
is
going
on,
and
then
they
use
all
these
weird
internal
stuff,
vm.get
nano
time
adjustment
on
the
vm,
which
is,
of
course,
a
native
method.
So
who
knows.
A
B
B
C
B
B
You
just
want
system.net,
basically,
no,
because
I
create
events
during
it
that
I
need
time
stamps
on.
So
I
need
wall
time
stamps
for
the
events
that
I
create
during
the
creation
and
your
recording.
A
B
B
If
you
would
prefer,
if
that
would
help
you
it's
not
that
big
a
deal
I
mean
I
can,
I
mean
I've
always
assumed
that
if
it
got
moved
out
of
internal
and
hidden,
I
could
I
can
re-implement
the
eight
lines
of
code
that
it
takes.
Yeah,
yeah,
okay
or
I
can
just
go
to
and
go
and
use
I
can
so
what
I'm
actually
doing
most
of
the
time
is.
I
just
assume
that
no
one's
going
to
adjust
the
clock
on
their
phone
while
they're
starting
up
an
app
like,
I
literally,
don't
think,
that's
possible.
B
So
they
have
to
be
something
really
weird
for
like
system.current
time
millies
to
actually
get
off
while
they're
starting
up
the
app
like.
Maybe
they
start
up
the
app
in
that
like
100
milliseconds,
where
it
takes
to
start
up
like
something
happens
with
daylight
savings
time,
and
the
clock
turns
like
it's
going
to
be
such
a
narrow,
weird
case
that
I
haven't
normally
been
worrying
about
it.
So
I
might
just
go
yeah
you're,
like
yeah,
so
so
one
one
session
on
one
phone
thinks
it
took
an
hour
to
start
up
the
app
like.
A
Sort
of
cool
how
rum
is
has
all
these
case.
These
unique
cases
like
google
analytics
has
always
had
that
same
feature
right,
like
it
queues
up
all
the
events
until
the
javascript's
loaded,
and
then
it
sends
them
right
and
so
like
technically.
If
I
sdk
cared
about
this
case
a
lot,
we
could
somehow
implement
that.
I'm
sure
I
don't
know
if
we
would
but
like
being
able
to
queue
up
events
like
having
a
pre-loaded
api.
That's
just
queuing
until
the
api
actually
gets
initialized
yeah.
A
B
Little
event
class
hold
on
to
the
time
and
then
it's
literally
just
a
time
stamp
and
a
name,
it's
all
sticking
in
there,
but
I
do
want
to
be
able
to
have
it.
I
want
to
see
the
timings
like
if,
for
some
reason
the
zipkin
exporter
starts
taking
you
know,
300
milliseconds,
to
initialize.
That
sure
would
be
nice
to
see.
B
A
B
Cool
well,
if
you
run
into
problems
with
the
release,
I
can
always
do
it
tomorrow.
Hopefully
this
one
will
go.