►
From YouTube: 2021-08-17 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
A
Like
the
offices
are
open,
like
you
can,
if
you
do
like
a
self
attestation
that
you're
not
sick,
you
can
like
you
can
go
in,
but
they've
been
from
my
understanding.
They've
been
pretty
deserted.
B
C
A
Yeah
wait
till
the
winter
when
it's
dark,
it's
still
dark
at
the
7
a.m.
C
A
I
don't
know
if
we
have
too
much
to
talk
about.
Certainly
the
one
five
release
I
I
saw
andrag.
You
were
already
working
on
that
the
update
how's,
the
metrics
feeling
yeah.
B
C
A
Oh
yeah,
some
of
the
you
had
made
some
change
some
improvements
in
one
five
right
that
we'll
get
to
pull
in
now.
B
Yeah,
so
I
think
I
forgot
one
important
one,
which
is
to
improve
the
assertion
messages
to
not
go
from
bottom
to
top
to
go
from
top
to
bottom.
I
have
that
blend
is
one
improvement
but
forgot
to
do
it.
I
think
oh
okay.
A
Yeah
yeah,
because
it
I
mean
I
mean
you
used
to
from
spock
having
to
know
this-
that
there's
a
particular
way
to
read
it,
but
like
I
had
to
learn
a
particular
way
to
read
the
new
yeah
messages.
B
B
A
Are
there
any
open
prs
that,
besides
the
obvious
that
you
want
in
anybody
wants
into
this
week's
release,.
C
Go
ahead,
it's
not
open
yet,
but
I
was
planning
to
remove
our
hug
with
copying
the
config
properties
to
the
system
properties.
B
A
Do
we
want
we.
A
This
was
needed
for
prometheus
for,
like
for
a
pull-based
metrics.
A
B
A
A
Did
honor
I
did
the
github
pages.
Did
you
see
the
discussion
with
jason's
comments
about
the
github
pages
stuff
that
it
was
something
about?
It
was
more
complicated
to
commit
to
a
different
branch
than
the
one
that
it
was
running
from.
B
Or
I
think
our
main
plant
is
protected,
let's
see
yeah
yeah,
so
we
tried
to
do
that
and
for
our
release
workflow
to
opt
in
docs
from
the
workflow
on
the
sticky
repo,
but
it
just
doesn't
work,
there's
no
way
to
get
a
workflow
to
commit.
At
least
there
wasn't
a
few
months
ago.
It's
improved
it's
good,
but
that
would
be
a
okay
problem.
I
think
yeah,
so
I
gh
pages
might
be
able
to
actually
solve
both.
C
B
Okay,
yeah,
so
that's
another
reason
to
be
worried
and,
like
I
thought
we
were
going
to
be
just
publishing
everything.
Github
pages
anyways,
like
the
text
files
or
whatever
this
is,
I
was
sort
of
expecting,
but
anywhere
is
fine,
but
the
main
branch
protection
is
going
to
be
a
problem
for
this
anyways.
A
And
nikita
is
not
here
but
mate.
Maybe
you
know
I
was
trying
to
make
some
changes
to
muzzle,
and
so
it
was
a
little
confusing
how
to
make
changes
and
get
those
changes
to
be
used.
Now.
A
A
Doing
was
what
worked
for
me
was
publishing.
C
Mostly,
I
think
so
so
unless
you're
changing
anything
in
the
gradle
plugins
code,
you
probably
don't
have
to
publish
anything
to
make
them.
I
mean,
if
you're,
if
all
you're
changing,
is
the
muzzle
muzzle
module,
then
it
should
be
picked
up
automatically
by
cradle.
A
Yeah
yeah
right
it
was
yeah,
publish,
yeah,
okay,
yeah.
I
think
that's
why
I
was
a
little
confused
got
confused,
because
I
was
actually
updating
things
in
both
places.
Okay,
yeah
I'll,
send
a
I'll
I'll,
get
it
sorted
and
send
a
pr
tomorrow
to
explain
what
I
was
messing
with.
A
Oh,
maybe
I'll
ask
you
both
here,
because
so
one
of
the
things
that
I
was
dealing
with
was
we.
You
know
what
the
couch
base
instrumentation.
A
A
We
shade
we
shade
it
so
that,
because
we
otherwise
we
would
end
up
with
two
copies
of
that
same
class
name
in
our
agent
class
loader
and
I
ran
into
I
had
to
do
the
same
thing
with
the
azure
sdk,
with
the
new
version
of
the
azure
sdk
that
they're
just
released
is
now.
A
So
the
the
azure
open,
telemetry
sort
of
library
uses
has
a
meta-imp
services
file
that
the
azure
core
tracing
picks
up
and
so,
of
course,
those
end
up
conflicting,
also
because
I'm
putting
them
both
into
the
agent
class
loader,
and
I
need
them
to
point
to
the
different
versions.
The
different
shaded
versions,
which
they
do
point
to
the
differentiated
versions.
But
the
the
one
meta
imp
file
name
is
the
same.
A
Oh
yeah,
so
you
can
see
here
the
resources,
so
I
kind
of
like
name
namespaced
the
resource,
so
I
put
it
under
a
subdirectory
and
then
over
here
instead
of
just
this,
I
did
this
kind
of
thing
to
like
name,
space
it
and
then
in
the
muzzle
nose
and
then
told
muzzle
that
this
means
a
directory
structure
to
read
it.
But
then,
when
injecting
it
into
the
class
loader,
you
only.
C
A
So
that's
the
strategy
that
was
the
best
strategy.
I
could
think
of
any
it's
not
amazing
for
sure,
but
any
thoughts.
B
B
C
C
But
honestly,
I
don't
think
it
would
probably
make
sense
to
have
like
a
separate
directory
for
sba
files
or
resource
files
for
each
instrumentation.
A
On
the
resource,
the
resource
would
be
have
namespace
and
file.
A
C
A
And
does:
does
this
make
sense
like
having
a
namespace
as
a
folder,
or
do
you
prefer
a
full
like
from
to.
A
B
C
So
so,
like
map
the
file
that
you
can
put
pretty
much
anywhere
in
your
channel,
oh
right,
right,
rename
it
to
the
services
in
the
targets.
A
Sold,
I
will
send
the
pr
for
further
comments.
B
Like
it
was
a
funny
pr
because
around
half
of
those
classes,
I
just
copy
pasted
one
of
the
other
classes
and
then
just
find
her
place
like,
for
example,
find
replace
long
to
double,
and
so
that's
how
you
convert
a
long
counter
into
a
double
count,
then
long
counter
longer
encounter.
I
just
have
to
choose
to
change
the
class
name,
because
the
apis
are
exactly
the
same.
B
A
A
B
So
I
didn't
notice
while
doing
this
and
that
it's
going
to
be
very
important
to
revisit
that
attributes
bridge.
I
think
I
think
I
had
a
draft
like
months
ago.
I
was
just
playing
with
it,
but
with
metrics
many
of
the
apis
except
attributes,
and
I
guess
that's
used
pretty
commonly
so
that's
going
to
be
very
slow
so
going
to
have
to
see
if
we
can
finally
use
the
interface
bridge,
oh
cool.
I
need.
B
A
All
right,
well,
maybe
may
the
force
be
with
you
yeah
for
now,
just
1.5.