►
From YouTube: 2021-11-16 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
Having
fun
with
cashing.
B
C
B
C
B
B
C
A
And
it's
isn't
it
at
least
frequently
released.
C
B
A
So
the
week
the
week
cash
it
sounded
like
I
saw
in
the
chat
that
I
had
or
not
quite
either
gotten
that
or
written
it
correctly.
A
A
A
A
A
And
that's
a
good
point,
well
that
what
anarag
said
that
I
mean
this
is
this
can
be
later?
I
mean
sorry
this
one
can
be
later.
This
one
would
be
fine
in
the
agent
for,
for
example,
one.
B
B
B
A
C
C
B
B
D
B
D
D
D
Yeah
and
yeah
now
that
I
think
of
it,
it
probably
would
be
better
if
we
don't
hold
strong
references
to
classes
so
yeah,
yes,.
B
So
yeah,
what
probably
I
I
am
saying
is:
do
we
actually
have
to
provide
cash
implementation
for
andra
or
we
or
can
we
have
bounded
cash,
separate,
artifacts,
separate
from
api,
and
so.
B
B
B
B
B
B
B
C
C
That
being
said
like,
for
example,
I
think
I
can
I
mean
I
don't
remember
all
the
cases
I
know
high
strix
uses
it.
The
database
sql
stuff
uses
it.
I
think,
there's
probably
at
least
four
places
that
use
the
bounded
cache
and
it
just
seems
like
instrumentation,
has
a
common
case
for
caching.
C
C
B
C
B
B
A
A
D
A
D
No,
I
that's,
if
I
remember
correctly,
it's
cache
of
like
header
name,
to
attribute
key.
So
it's
a
infinite
number
of
these,
since
they
have
to
be
configured
explicitly
in
the
property.
So
I
think
I
just
used
cache
because
it
seemed
to
be
like
a
natural
thing
to
use,
but
it
could
as
well
be
a
concurrent
link
like
a
current
hashmap.
D
A
C
B
C
B
B
A
B
D
C
D
A
D
D
A
Pull
it
are
you
saying
pull
the
that
into
the
jdbc?
Oh
no!
This
is
an
instrumentation
api
yeah.
It's.
D
Used
by
all
sql-like
instrumentation,
so
cassandra
cloud-based,
it's
not
just
sql,
but
we
could
get
rid
of
bounded
cache
as
an
ati
and
just
use
the
concurrent
link
hashmap
directly
in
this
class
and
make
it
internal
dependency
or
something
it's
very
small
and
there's
exactly
one
usage
of
cash.
A
C
A
B
That
just
needs
reviewing
anyway,
so.
B
C
C
B
I
mean
if
that's
inside
the
agent
anyway.
C
Okay,
like
to
venture
at
the
same
time,
we
know
this
is
never
going
to
be
updated.
So
that's
also
the
reason-
and
it
says,
version
1.4.2,
so
I
would
actually
be
okay
with
not
shading.
A
Yeah,
I
just
worry
about
downstream
upstream.
C
A
C
A
B
A
B
C
I
realized
that's
probably
safe
to
I
don't
know.
If
anyone
saw
my
comments,
I
was
worried
that
we
would
accidentally
leak
the
logs
sdk
into
people
that
are
using
the
instrumentation
and
suggesting
some
more
grams.
I
realize
that
if
we
just
compile
on
the
sdk
for
now,
while
that's
getting
cleaned
up,
that
should
still
be
safe.
C
C
C
C
A
And
I
can
probably
unblock
this
one
materials,
because
I
had
reviewed
it
all,
except
that
one
problem
that
looked
like
didn't
work
out
anyway.
D
I,
if
I
remember
correctly,
that's
what
nikita
has
encountered
when
he
extracted
the
attribute
keys
thingy
into
a
separate
library
in
eddie.
So
if
you
have
two
instrumentation
modules
and
one
of
them
passes
in
the
other
phase,
it's
pretty
much
guaranteed,
as
you
would
sorry
fail.
The
muzzle
check
in
build
time
yeah.
It's
just
this
pattern.
D
A
D
A
D
C
D
But
if
we
include
it
as
a
implementation
dependency,
then
muzzle
checks
will
basically
fail,
because
we
we
have
two
reactor
native
models,
modules,
reactivity,
0.9
and
1.0,
and
they
have
different
of
their
matches
different
muzzle
references,
and
we
expect
one
zero
to
fail
on
three
one:
zero
versions,
but
neddy
nettie
for
one
instrumentation
will
always
always
pass.
D
There
is
one
instrumentation
module
that
passes
the
muzzle
check
on
the
reactor
in
the
t0.1,
0.9
and
yeah,
and
that
makes
the
whole
thing
break.
A
C
A
Jungle
of
artifacts:
what
is
what
are
we
looking
for
muzzle
transitive.
A
A
Nikita
thank
you.
I
saw
you
were
trying
to
help
this
user.
B
B
Different
opinions
on
that,
so
essentially,
what
he's?
What
he
wants
is
to
make
a
couple
of
our
packaging
package.
Private
classes
public,
because
then
some
right
back
either
functionality
or
tests
will
become
easier,
and
I
was
trying
to
battle
that,
but
he
he
seems
to
insist
and
I'm
not
sure
that
I
am
that
much
of
right
back
expert
to
properly
object.
To
that.
Like
my
only
objection
is,
can
we
not
make
more
classes
public
and
that's
pretty
weak
objection.
B
D
Look
at
this
pr
for
a
long
time,
but
at
the
beginning
I
remember
that
the
the
this
user-
they
only
wanted
not
actual
classes,
but
just
you
know
the
dot
class
values
for
those.
So
maybe
it
would
be
enough
to
just
implement
some
kind
of
get
what
is
called
get
handler
class
and
get
interceptor
executor
class
in
the
main
tracing.
B
That's
that's
one
way.
Yes,
I
believe
there
is.
B
Some
code
in
his
pull
request,
which
would
be,
would
be
easier
with
just
like
static
blah
blah
blah
dot
class,
as
opposed
to
I
have
to
grab
that
instance
and
then
scroll
something
on
that.
But
if
somebody
has
time
please
review
that,
I
I
mean
in
a
little
bit
like
struggle
internal,
I
don't
know.
C
B
Acting
great
well,
they
want
they
didn't
want
to
use
getters
yeah
after
my
push
that
now
kinda
use
it.
But
if
you
scroll
a
little
bit
scroll
there.
C
B
B
C
C
A
A
C
A
A
B
A
A
B
C
B
C
A
C
A
More
complicat,
if
it
was
yeah
if
it
was
more
complicated,
then
that
would
weigh
on
me
but.