►
From YouTube: 2021-08-10 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).
A
B
C
B
B
So
it's
probably
release
week.
I
think
I
just
put
in
a
pr
to
gather
up
to
change
logs
and
in
slack
please
josh
was
asking
if
we
could
get
the
metrics
exporter,
auto,
configure
sdi
place
for
the.
C
B
B
Well,
I
was
just.
I
was
literally
right
before
this
meetings
taking
a
peek
at
j
releaser.
Have
you
seen
this
thing?
It's
a
new
project
inspired
by
goal.
Releaser,
I
don't
know.
Maybe
it
said
jrelisser.org.
B
B
B
B
C
B
Yeah,
it's
hard
to
know
how
much
was
your
yard?
How
much
was
just
yeah
anyway,
I
guess
we'll
watch
it
over
the
next
yeah
week
or
so
yeah
and
then
christian
put
in
this
pr
this
morning
around
version.properties,
which
I
think
looks
fine,
but
as
our
resident
gradle
expert.
C
B
So
I
think
the
main
thing
that
he
was
trying
to
get
at
which
I
asked
the
question
was
like
what
what
is
this,
for,
I
don't
understand,
was
that
if
you're
doing
local
builds
and
and
the
version
gets
rev
like
the
tag
gets
introduced
along
the
way
that
gradle
was
caching
the
version
and
then
that
input
stud
property.
Yes,.
C
B
C
C
I
think
that's
our
intention.
I
mean
it's
sort
of
okay,
but
I've
seen
a
couple
libraries
also
like
they
put
into
every
single
jar
and
that
make
sure
you
know
what
the
version
is
of
each
jar.
You
can
do
some
checks
in
run
time,
also
to
make
sure
you
align
your
dependencies
correctly.
So
I'll
comment
on
that.
C
B
Oh,
the
other
thing
I
wanted
to
ask
about,
and
I
I
posted
this
on
slack,
that
the
new
ot
otlp
over
http
spam
exporter
has
publishing
turned
off
because
jack
wasn't
able
to
get
get
something
to
work
right.
So
if
we
want
to
publish
that
this
release,
we
should
probably
try
on
publishing.
B
B
Would
it
make
sense
for
us
to
kind
of
have
a
process
where
we
publish
an
alf
like
for
new
modules
like
this?
We
mark
it
as
alpha
for
one
release
to
let
people
give
it
a
try.
I
don't
know
just
it
was
it's
just
a
thought,
because
we
might,
we
could
release
it
and
it
could
be
totally
broken
since
no
one's
used
this
thing
in
anger,
yet
anyway
yeah.
C
A
C
B
A
B
B
B
B
C
B
B
Yeah,
oh
yeah.
Well
I
guess
there
was
the
property
option.
Then
there
was
just
like
having
it
be
fail-safe
so
that
it
won't
blow
up.
Although
I
guess
you
need
to
make
sure
that.
B
C
C
C
C
A
B
B
All
right
issue
created
3463
I'll
assign
it
to
the
two
of
us.
Just
so
at
least
it's
there
I'll
I'll
sign
it
to
jack
too.
Maybe
he'll
get
to
it.
Maybe
he'll
be
won't
sleep
tonight
and
then
we'll.
Instead
work
on
gradle,
probably
not.
A
B
Well,
breaking
changes
of
us
if
they
are
like
translatable
like
if
you
can
convert
from
one
from
version
a
to
version
b,
and
you
can
build
the
schema
to
support
that
and
it's
then
it's
legal,
but
if
you
just
have
like,
I
think
the
thing
that
isn't
legal,
if
you
just
completely
drop
an
attribute,
then
there's
no
way
to
or
yeah
drop
drop,
one
of
the
attributes
there's
no
way
to
convert
from
the
from
a
to
b.
In
that
case,
I
think
in
that
case,.
B
Yes,
yeah
definitely
can
be
renamed,
there's
a
weird
case
where,
with
if
you
have,
if
you
like
fold
an
attribute
into
an
existing
one
like.
I
think
this
has
already
been
brought
up
at
one
point
like
if
you
decide
that
we
don't
need
rpc.net
and
you
just
use
net.something
and
that's
the
one
that's
already
exists,
then
you
could
write
a
transform,
but
then
you
might
have
instrumentation.
That
was
writing
both
of
them
and
then
how
would
you
resolve
it?
So
there's
still
some
outstanding
questions.
Yeah.
B
B
B
B
B
C
C
C
B
I
don't
like
having
to
use
reflections
make
up
for
very
terrible
interface
changes
yeah
well
yeah,
especially
since
okay
published
breaking
interface
changes
without
your
immersion
is
a
bummer.
B
B
Anyway,
I'm
reasonably
happy,
as
I
have
come
in
the
pr,
rather
than
trying
to
use
the
tag
methods
if
they're
there
it's
much
simpler,
just
to
use
the
cache
and
just
stick
with
that,
like
no
reason
to
spend
extra
time
trying
to
see
if
there's
a
tag,
I
think
the.
B
Is
that
is
that
cash
I
haven't
looked
at
the
implementation?
Is
that
does
that
come
from
guava,
or
does
that
come
from
some
other
place,
the
one
that
the.
C
Unlimited
keys
version
is
the
weak
cache
map
it
comes
from
bike.
Buddy
sort
of
this
is
the
same
person
that
made
my
buddy
as
his
weak
map
library,
because
anyone
that
does
this
sort
of
thing
needs
to
add
random
data
to
random
objects
if
you're
doing
instrumentation
by
putting
manipulation.
I
guess
that's
right.
Yeah.
B
C
B
C
B
It's
pretty
close,
it's
really
close.
They
just
need
one
more.
I
wish
there
was
one
more
method
on
the
event
listener,
which
is
when
it
got
actually
cued
onto
the
like
dispatcher,
because
if
we
had
that,
then
we
could
totally
do
it
and
it
would
be
a
much
better,
much
better
way
to
approach
it
well,
except
except
for
having
to
launch
the
request
right.
B
B
B
C
C
B
C
C
B
Yeah,
no,
I
mean
it
is
good.
Sometimes
there's
like
the
super
crazy,
the
super
crazy
samplers
that
I
don't
even
understand
why
this
is
that
was
a
different
customer,
but
the
crazy
sampler
that
would
only
like
eliminate
at
all
internally
source
root
spans
like
we'd,
never
want
to
see
anything
that
comes
from
that
starts
on
an
internal
process,
but
they
weren't
willing
to
just
disable
the
instrumentation.
B
B
C
C
C
B
You
get
alolita
on
it,
she'll
she'll
get
that
thing.
B
She
seems
to
be,
she
seems
to
be
perfectly
okay,
with
shaking
some
trees
and
getting
people
to
talk
about
stuff.
B
I
think
carlos
has
started
getting
frustrated
and
just
starting
to
merge
things,
even
if
they
don't
technically
have
enough.
If
they
sat
around
and
no
one
had
said
no,
it
has
a
couple
of
approvals.
C
C
B
I
haven't
thought
I
haven't
cared
very
much
about
the
why
it
shouldn't
be
the
case.
I've
mostly
been
worried.
B
B
I
mean
I
have
two
thoughts,
one.
We
can
add
an
additional
method
to
the
sampler
interface
and
give
it
a
default
implementation
that
calls
the
old
one.
But
that
means
every
implementation
of
a
sampler
that
wants
to
use
the
new.
One
also
has
to
implement
the
old,
the
old
method
right.
They
can't
they
can't
just
implement
the
new
one.
B
C
I've
found
that
method
to
like
having
to
write
a
do
sample
method
in
every
single
unit,
test
right
related
to
sampling.
It's
also
pretty
annoying
just
because
there's
so
many
parameters
and
I
just
wanted
to
pass
in
a
trace
id
or
something.
So
if
that
thought
has
come
to
mind,
it's
still
the
same
problem,
but
at
least
it
avoids
the
problem,
the
future,
I
guess
we're
saying
right:
yeah.
B
C
B
B
A
C
A
B
Whatever
yeah,
I
wish,
we
had
bundled
all
that
stuff
up
in
metadata,
like
sampling,
media,
originally.