►
From YouTube: 2021-02-24 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).
B
B
C
Yeah,
thank
you
for
the
the
nettie
work.
Those
concurrency
failures
are
very
unnerving.
D
A
A
A
A
B
A
Yeah
I
just
wanted
to
because
my
issue
didn't
didn't
bring
any
attention,
so
I
just
wanted
to
ask
you
in
the
face:
what
do
you
think
about
we
smoking
tests
on
windows?
Do
we
need
them
or
not?
I
mean
we
have
them
in
the
splunk
already.
So
the
question
is
just
stop
trimming
them
to
open
telemetry
repo.
B
Let's
be
careful,
I
know
that
github
actions,
windows
runners-
are
way
slower,
but
if
it's
only
smoke
test,
maybe
it
doesn't
matter
because
the
rest
of
the
build
is
slow.
A
Justin
gay
yep,
almost
twice
as
I
slow
linux
tests,
yep.
B
Maybe
it
doesn't
matter
just
because
they're
the
rest
of
the
build
is
also
slow,
but
if
it's
causing
annoyance
because
the
build
gets
too
slow,
then
there
is
cost
I
mean
we
could
only
do
it
on
lightning
build.
It
doesn't
have
to
be
on
every
build.
Like
there's,
of
course,
many
options
I
think,
having
some
tests
in
some
way
does
sound
like
a
good
idea,
though,.
A
C
So
do
they
use
windows
containers?
How
do
they
yep.
B
A
A
B
C
D
C
I
see
yeah
because
we
can
also
potentially
split
out
parallelize
this
some
of
the
smoke.
A
B
A
C
Yeah
we
have.
We
have
some
really
good
tests
now
between
the
instrumentation
ones,
running
with
the
java
agent
and
all
the
smoke
tests
and
all
the
different
app
servers.
C
C
B
C
B
C
B
B
Stuff,
or
is
just
it's,
it
uses
electricity
like
it's
sort
of
similar
to
zipkin
jager
in
some
sense,
like
it
uses
elasticsearch
as
a
back
end
and
ingests
from
the
open
filament
collector
using
otlp.
C
B
B
C
C
I
would
imagine
the
makita
the
the
splunk
stuff
given
how
all
in
on
otlp
they
are
that
that
will
be
otlp
native
if
it's
not
already
what
exactly
you're
your
back
end
ingestion.
A
A
B
C
Oh
the
stability,
the
comment
there
in
the
meeting
was
about,
given
that
it's
stable
or
once
it's
stable,
having
duplication,
might
not
be
bad.
Yeah.
A
Duplicate
that
and
keep
it
in
sync
and
and
be
a
target
of
bug,
reports
that
we
should
that
we
have
to
delegate
back
to
outer
configuration
module.
B
C
B
B
C
B
You
can
talk
about
next
item,
the
caching
so
yeah.
The
only
reason
I
even
went
into
this
thing
about
maybe
removing
usage
of
guava
was
wanting
to
separate
out
my
instrumentation
api
or
library
instrumentation,
and
so
one
thing
I
found
was
that
we
use
some
things
that
can
only
be
used
in
the
agent
like
caching,.
B
C
A
Yeah
pulling
jar
file
and
trading
it
it's
cleaner
from
licensing
perspective,
then
our
license
generator
will
take
care
of
that
yeah
as
well
automatically
yeah.
C
Just
yeah,
I'm
not
sure
I
support
that
suggestion.
Okay,
yeah
me,
neither
so
I
guess
none
of
us
did
but
yeah,
I
guess,
but
we
can
shade
it
in.
I
thought
we
were
gonna
have
to
copy
the
source
code
into
our
repo.
C
I
guess
looking
at
the
like,
I
would
say
if
it's
you
know
a
meg,
you
know
that's
significant,
something
that
we
would
care
about.
If
it's
100k,
I
would
say,
use
that.
A
B
B
C
C
Yeah,
my
guess
is
that
our
I
mean
our
caching
is
not
maybe
so
critical
that,
let's
start
with
the
small
one,
if
we're
we're
shading
it
in
and
not
exposing
it,
we
should
be
able
to
swap
it
out
in
the
future.
If
we
want.
B
B
C
B
B
B
Right
now,
so
maybe
right
now
yeah
so
right
now,
because
we
have
guava
in
the
agent
class
loader,
we
needed
the
bounded
cache
interface
to
use
it
from
instrumentation,
but
if
we
shade
it
and
then
it'll
be
in
the
bootstrap
class
loader
and
that
problem
goes
away.
So
this
interface
is
not
as
important.
C
C
B
C
C
B
C
Hard,
so
we
would,
but
we
would,
if
we
don't
put
it
in
an
interface,
hide
it.
Then
we
would
hide
it
in
an
internal
package
and
say
that.
B
C
B
I
mean,
of
course,
after
maintaining
our
groovy
tests
for
months.
I
really
hate
them,
but
that's
a
separate
topic.
What
I
really
want
is
to
make
sure
our
java
testing
solutions
like,
for
example,
the
testing
library.
B
We
have
the
sdk
every
world
coverage
so
to
say,
like
they're
used
in
instrumentation,
so
we
know
that
they
actually
work
for
real
use
cases,
not
just
unit
tests,
and
so,
if
it's
okay,
to
have
at
least
some
of
the
instrumentation
tests
in
java
to
make
sure
we're
using
both
versions
and
someone
can
write
their
tests
in
java
and
be
happy.
Someone
can
write
their
tests
in
groovy
and
not
be
happy,
but
I
think
they
think
they're
happy.
A
And
they
are
based
on
how,
on
our
artifacts,
the
new.
C
D
B
A
I
mean
as
long
as
our
that
usual
tasks
mean
http,
client,
test
or
http
server.
That's
how
all
all
the
same!
But
if
you
have
some
standalone-ish
or
separate
instrumentation
for
whatnot,
which
which
is
not
like
our
usual
tests
and
it
has
its
own
specific
tests
they
can
be
in
java,
we
will
not
prohibit
that.
A
C
Move
to
java
just
to
see
what
you
know
see
how
they
look
kind
of
evaluate
if
it's
something,
because
I
know
this
you
know-
does
keep
coming
up
and
I
do
think
the
general
new
contributor
experience
would
be
better
if
it
was
not
ruby
and
spock.
C
B
C
B
Really,
when
I'm
running
the
debugger,
when
I
need
to
use
the
debugger,
that's
where
it
really
just
kills
me
anyways.
I
don't
want
to
push
for
like
migrating
all
of
our
tests
to
java,
even
though
I
would
like
it,
I'm
more
right
now
just
curious
about
if
it's
okay,
to
have
some
instrumentation
in
java
and
some
in
groovy
from
the
step
stance
that
it's
not
good
to
have
inconsistencies
in
run
repo,
but
at
the
same
time
it
doesn't.
Let
us
show
that
both
versions,
work
well
in
real
world
use
cases.
C
I
also
think
that
you
know
it
gives
us
a
chance
to
that's
the
only
way,
we're
going
to
evaluate
genre
tests,
and
I
do
think
that
we
should
stay
open
to
that
option
of
migrating
at
some
point.
If
you
know
if,
but
we
need
to
take
that
flow
and
see
what
it
looks
like
and
see
how
I
feel.
B
B
B
Like
one
example
is
like,
when
I
compare
I
by
I,
I
could
find
that
our
testing
didn't
have
the
something
like
has
entry,
satisfying
like
where
you
provide
a
function
to
match
an
attribute
value
like
ours
does
and
this
sort
of
thing
I
would
never
notice,
while
writing
unit
tests
for
the
class.
I
have
to
write
real
instrumentation
tests.
C
Cool
so
under
rug,
are
you
still
wanting
to
release
one
zero
of
the
java
agent
like
this
weekend
after
the
1-0
of
the
sdk.
C
I
don't
see
any
reason,
not
why
not
on
nikita.
Can
we
lose
me
is
that
peter
frozen.
A
A
C
A
A
A
Let's
see
that,
so,
let's
try
for
sdk
to
release
new
version
and
see
how
how
we
can
handle
that
yeah,
but
maybe
it's
good
for
sales
engineer
to
be
able
to
claim.
Yes,
that's
1.0
stable,
please
try
it.
Please
use
it!
Yeah.
C
Yeah,
I
can
do
this
and
then
I
don't
really.
I
I
think
let's
do
this
first
and
this
will
kind
of
tell.
I
think
this
is
more
important
than
the
documenting
those
config
options,
because
we
first
have
to
decide
what
we're
declaring
is
stable,
even
yep.
If
it's
not
stable,
I
don't
it's
not
really
that
important.
If
we
document
it
or
not,.
A
C
All
right
anything
any
anything
else
that
any
pr's
that
we
need
to
chat
about.
C
J
center:
good
luck,
good.
A
A
Because
there
are
versions
on
maven
central
that
we
want
to
muzzle
check,
but
for
that
dependency
is
broken
so
metadata
it
does
doesn't
list
them.
So
when
our.
C
A
That
yeah-
that's
that's
this
there's
exactly
some
jakarta
faces,
so
muscle
give
me
a
range
from
these
to
this
and
megan
central.
Look,
no
versions
here.
A
C
B
A
B
A
C
B
C
A
B
A
B
B
A
C
Yeah
is
there
some
way
with
gradal
that
in
our
build
we
could
monkey
patch
the
maven
metadata
xml.
A
B
A
A
B
A
No,
they
are
red
mostly
due
to
various
latest
episodes,
so
these
are
like.
No,
let
me
let
me
fix
myself.
Yesterday
we
had
a
lot
of
build,
fail
failures
due
to
j
center
timeouts.
I
see
those
rb
brb
and
the
wheels
we
have
nightly.
Failure
due
to
the
latest
latest
depth
failures.
We
should
separate
and
this
and
then
the
latest.
B
C
C
Yeah
I
mean
the
muzzle
muscle,
I'm
trying
to
think
of
what
do
we
lose
specifically
here
by
not
testing
because
we've
already
tested,
I
mean
we
know
that
the
muzzle
was
passing
on
all
these
versions
as
new
versions
come
out
for
now,
oh,
but
we
don't
have
two
plus
anymore,
so
it's
not
gonna
pick
up.
C
Yep
all
right,
I
I'm
fine
with
I
mean
it's
jsf,
I
don't
care
too
much
about
jff,
I'm.
I
would
be
fine,
commenting
it
out
and
creating
an
issue
to
revisit
yeah.