►
From YouTube: 2021-03-18 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
A
B
A
A
A
A
A
So
and
those
libraries
which
use
that
fail
so
from
one
point
we
do
have
already
support
in
our
test
infrastructure.
To
say
this,
like
this
instrumentation
should
be
tested
only
on
java
versions
until
up
until
this.
A
But
if
we
use
that
for
16,
that
means
that
we
we
we
don't
test
on
16,
like
half
of
our
library,
so
significant
portion
of
our
integrations
and
our
current
policy
that
we
test
on
java,
8
java
11.
A
That
I
don't
know
yet
I
I
haven't
been
living
up.
I
just
saw
several
several
failures,
which
are
clear
due
due
to
libraries,
and
so
I
step
step
back
so
so
if
we
just
limit
instrumentations
now
testing
up
to
java
15.
That
means
that
we
actually
should
test
both
on
15
and
16.,
because
the
because,
if
we,
if
you
don't
test
on
16
like
quarter
of
our
instrumentations,
that's
bad,
I
assume
we
still
have
to
test
with
latest
version
and
that's
strange
as
well.
C
A
A
D
Good,
so
it
seems
like
what
we
need
to
do
is
we
need
to
disa,
or
I
don't
know
if
we
have
the
mechanism
sounded
like.
Maybe
we
do
to
just
say
you
know,
test
up
to
jdk
version.
15,
for
this
set
of
libraries
find
all
of
those
libraries.
I
mean
it's
going
to
be
a
pain,
but
you
just
have
to
grind
it
out
and
get
that
done,
and
we
don't
mean
obviously
the
agent's
not
going
to
work
with
those
libraries
on
java
16.
So
there's
no
reason
to
test
it
right
like
it's.
D
A
But
then
there
are
two
questions
for
for
me.
First,
should
we
still
test
on
50m,
which
is
like
the
last
version,
the
the
majority
or
all
of
our
info
current
instrumentations
work
on?
And
the
second
question
is
what,
if
somebody,
some
library
in
two
weeks
or
in
a
month
start
supporting
16.
D
D
A
Well,
unless
you
buy
vendor
support,
which
I
don't
know
either
any
vendor
who
is
going
to
continue,
support,
provide
paid
support
for
15,
they
may
probably
not,
though
probably
not,
but
well,
if
you
have
money.
A
D
C
I
think
it's
as
needed,
like
I
think
it's
going
to
be
as
it
comes
up
as
people
want
it
as
people
see
it
as
we're
doing
regular
updates
to
our
supported
library,
instrumentation
versions,
like
we
kind
of
already
have
this
problem
right
like
anywhere
that
we're
using
that
constraint
for
java
versions.
Already
we
have
this
problem.
A
E
Yes,
sorry,
my
phone
died
last
night,
so
my
alarm
didn't
go
off
this
morning
and
happy
at
daylight.
Hopefully
everybody,
the
u.s
friends,
are
doing
better
with
daylight
savings
this
week
than
I
am.
E
I
was
gonna
say
for
like
say,
for
example,
it's
spark.
That's
spark
java
that
doesn't
run
on
java
16..
If
there
are
like
java
16
spark
users.
E
I
assume
that
they
would
just
use
that
flag
for
now,
so
when
one
may
be
so,
should
we
like,
if
we
use
the
flag
for
testing
as
opposed
to
constraining,
at
least
for
now,.
A
A
A
A
E
Yeah,
I
think
this
is
going
to
be
the
kind
of
the
key
piece
potentially
is.
If
we
can,
if
there's
a
way
for
us
to
keep
testing,
you
know
the
old
versions
of
the
library
on
normal
on
other
java
versions,
but
latest
do
we
need
or
do
we
need
two
now
do
we
need
a
latest
depth
test
15
and
a
real
latest.
E
Not
well
not
15
right
because
we
would
basically
drop
15,
but
maybe
a
latest
test.
A
A
A
Don't
get
supported
from
bender
or
you
don't
get
security
patches
into
it.
So
that's
not
exactly
the
same
that
it
doesn't
exist
so
anyway.
Yes,
so
I
I
don't
have
a
perfect
solution
and
I'm
even
not
sure
what
is
my
preferred
solution,
but
I
think
that
we
soonish
have
to
come
up
with
something.
A
E
A
Okay,
I
I
certainly
can't
set
up
17
ci
job,
which
will
be
read
for
foreseeable
future.
Should
I
just
for
the
time
being
start
constraining
that
mac
java
version
15
and
add
16
into
our
ci.
A
If
nobody
wants
to
comment
something,
then
I
will
go
to
the
next
one
and
the
next
one
is
just
fyi
that
the
hotel,
spec
slash
maintainers
is
going
to
start
like
old,
open
telemetry
initiative
initiative
to
essentially
to
answer
the
question:
how
do
we
get
the
whole
internet,
contributing
library
instrumentations
to
hotel
and
not
us
doing
all
the
work
which
you
can
join
the
discussion
on
open,
telemetry
slack
in
that
channel,
which
is
in
in
notes?
A
A
E
E
Proposes
the
kind
of
the
new
name
so
potentially
hotel
java
extensions,
which
could
be
also
we
could
allow
that
to
be
a
directory
and
we
could
consume
all
the
jar
files
in
there.
E
So
that
people
could
drop
in
you
know
those
third-party
java
agent
instrumentations,
in
addition
to
their.
A
E
So
if
it's
on
the
system
class
path,
the
problem
is
that
most
of
these
things
even
spring
boot
and
certainly
web
apps-
that's
gonna
end
up.
In
a
you
know,
a
nested
class
loader
somewhere.
A
E
Something
to
think
about
for
sure,
because
we
could
you
know
we
have
access
to
it.
At
that
point
we
can,
I
mean
it's
tricky
we
gotta.
Basically,
we
want
to
consume
those
the
the
resources,
the
class
files
redirect
those
into
the
the
agent
class
loader
right
to
me
match
sort
of
the
existing,
because
we
need
to
be
able
to
call
it
from
there
kind
of
like
almost
copy
and
move
the
class
bytes
from
the
user's
class
loader
over
into
the
agent
class.
A
A
B
E
Is
in
the
agent
class
loader
and
it
needs
to
be
able
to
call
those
matchers
in
the
user
class
loader,
so
we
need
some
kind
of
bridge.
E
B
E
E
A
A
E
Does
this
sorry
just
does
this
kind
of
capture
what
you
were
same
yep,
I
think
so.
Okay,
cool,
sorry,
nikita.
A
That
probably
will
be
a
good
first
first
step
or
second,
because
we
need
stable
api
anyway,
but
then
we
can't
think
about
those
user
class
paths,
because
that
certainly
will
be
very,
very
convenient
for
for
end
users.
E
So
that
there
was
a
that
ties
a
little
bit
to,
I
would
definitely
you
know
this
idea
of
integrating
or
customer
applications
being
able
to
configure
the
java
agent.
Also,
like
we've
gotten
a
couple
of
requests
for
that.
A
A
But
and
let's
not
forget
about
one
more
aspect
of
these
open
telemetry
world
domination-
is
that
we
do
want
that
libraries
instrument
themselves
with
open
telemetry
api.
So
that's
certainly
the
ideal
north
star.
E
E
Yes,
we
need
some
pinky
in
the
brain
background.