►
From YouTube: 2020-07-10 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).
B
C
E
C
E
Okay,
we'll
see
well
done
I
guess
that
he's
probably
busy
this
week.
Okay,
shall
we
start
or.
F
E
A
E
A
F
F
B
F
E
E
B
E
B
Me
the
metrics:
well,
we
don't
have
specification
on
the
metrics
yet
so
nothing
to
do
yet
until
like
all
of
Josh's
metrics
stuff
still
has
still
is
either
a
no
tip
or
NPR
in
the
specs.
So,
okay,
well,
that
gets
done
and
views
haven't
even
made
it
through
otep.
Yet
so
it's
gonna
be
some
time
before
the
metrics
work
is
there,
but
it
is
coming.
Ok,.
E
Ok
good
to
know
because
I
started
to
catch
up,
I
already
started
here
at
like
set
to
ask
for
more
time
for
these,
so
I
can
actually
help
you,
ok,
but
it's
good
to
know.
So.
In
that
case,
I
can
probably
take
day.
This
part
myself
I
kept
idea
how
it
will
be
implemented,
but
I
can
take
you
stuff
and
we
can
iterate
to
yeah.
B
E
B
Yeah
and
I
don't
want
I,
don't
necessarily
want
to
debate
the
issue.
I
just
want
to
figure
out
how
what
we
can
do
to
get
ourselves
orbs
a
decision
like
what
do
we
need
to
do
in
order
to
make
a
decision
about
what
we
want
to
do
here?
Nope
mm-hmm,
that
that's
my
mate,
that's
the
main
goal,
my
I
don't
have
a
super
super
strong
opinion
on
this
or
I
would
be
aggressively
pursuing
it,
but
I
don't
feel
like
I
feel,
like
we
kind
of
heard,
an
impasse.
G
E
So
I
really
like
the
idea
of
how
we
can
independent
an
artifact
like
extracting
your
busy
context
from
GRDC
and
I,
know
that,
but
that
has
to
be
like
this
decided
on
set
in
stone.
If
that
doesn't
happen,
we
should
probably
create
a
linear
ourselves.
I
think
there's
no
perfect
solution
here
and
actually
I,
don't
know
what
happened
honestly.
I
remember
that
poke
thing
was
supposed
to
contact
the
generous
guys
and
asked
for
their
blessing
or
and
or
feedback
I.
E
C
About
Springs
cold
drink
and
months
ago,
he
did
investigate
tomorrow
a
doctor
and
he
showed
us
there's
something
like
snapshot
caught
about
reactor
context.
To
me
too
well
known
and
I.
Think
John
was
there,
but
after
he
dies,
it
was
absent
for
months.
I
think
there's
no
movement
in
this
direction,
so
I
think
we
are
at
the
same
place
where
he
was
months
ago.
E
C
F
C
E
E
B
B
H
H
B
H
The
fact
that,
in
order
for
propagation
to
work,
you
have
to
trust
the
color
that
they
have
the
right
context,
but
in
the
meantime,
if
there
is
a
leak
like
what
you
said,
you
have
a
problem.
If
you
don't
trust
the
color,
no
propagation
works,
I
mean
it's
just
not
this
example,
but
in
any
example,
if
you
don't
trust
that
the
color
put
the
right
thing
into
the
into
the
context
into
the
tread
water
context,
then
no
propagation
words.
H
H
H
How
is
that
gonna
work
even
in
even
if
it's
your
context,
even
if
it's
your
context,
how
do
you
know
that
the
color
is
not
actually
somebody
that
you
know,
for
example,
if
somebody
erupts
JRTC
into
let's
say
HTTP
framework
or
something
like
that,
and
they
create
a
span
on
that
side
and
they
wrapped
the
grdc
or
any
propagator
like
or
even
vice
versa.
They
wrap
every
HTTP
call
into
a
JRPG
call.
H
The
gel
PC
will
already
installed
something
into
the
context,
and
when
you
get
the
plugins
of
your
HTTP
framework,
which
was
actually
wrapped
by
the
GRDC,
you
have
something
in
the
context
that
you
have
to
respect
so
so
it
is
a
huge
problem
and
trust
me
I
had
to
deal
with
these
for
five
years
in
Google.
As
a
wonder
of
that
context,
it
was
not
easy.
B
H
Gathering
one
requirement
I'm
missing:
do
we
need
to
be
able
to
swap
the
implementation
order
and
which
part
should
be
configurable?
Remember
the
whole
screen
touches
that
once
I
have
that
information
I
can
go
to
2g
RPC
and
propose
something
to
them
and
we
can
make
a
decision,
but
that's
the
missing
piece
that
I
yes,
indeed,
I
mean
I,
have
five
and
I
but
Carlos.
If
you
let.
H
We
need
to
understand
very
well
what
is
the
part
that
needs
to
be
configurable,
and
what
should
we
allow
people
to
do
dependency
injection
on
or
something
maybe
it
is,
and
what
was
that
part,
because
that
that
was
the
whole
discussion
that
we
had
three
month
two
months
three
months
ago
about
G?
Is
the
configuration
well,
what
should
we
have?
The
ability
to
use
dependency
injection
for.
H
E
H
Means
interaction
with
the
Redrock
yeah,
maybe
maybe
it's
more
than
that.
Maybe
it's
the
data.
Let's,
let's
clearly
understand
what
exactly
needs
to
be
to
be
configurable.
Okay,
make
sense,
and
after
that,
as
I
said,
we
can
I
can
introduce
to
the
dirty
seaports.
I
can
send
an
email
but
I
need
to
I.
Have
this
answer,
because
I
need
to
go
with
some
proposal
to
them,
not
just
saying
something.
B
B
D
D
Start
with
when
we
create
a
server
span,
do
we
want
in
this
moment
of
time
say
we
have
a
clean
state?
We
are
the
start
of
the
local
Finley
o
as
baculum
claimed.
We
shouldn't
do
that.
So
that's
that
that's
the
question,
but
it
it's
not
pressing
for
me
right
now.
My
problem
with
all
those
tests
is
soft
anyway,.
H
To
answer
your
question,
what
I
would
do
is
the
following:
I
will
think
this
way-
and
this
is
this-
was
my
mindset
in
Google.
If,
if
a
trap,
if
a
thread
picks
up
a
work,
a
new
work,
that's
when
a
new
context.
It
is
that's
the
only
moment
when
you
have
a
clean
context
when
the
thread
picks
up
a
word.
H
D
I
understand
that,
for
like
a
general
attribute,
but
specific
problem
is
that
when
I
create
spam
and
I
want
to
extract
remote
spam
context
from
the
HTTP,
for
example
like
in
time
and
if
I
fail.
If
there
is
nothing
income
so
should
but
I
create
a
I,
create
server
spam
and
by
specification
the
only
the
only
parent
of
the
server
spam
can
be
a
removed
spam.
So
I
cannot
have
the
parent
of
the
server
responded.
I
have
just
created
to
be
arbitrary
span
from
the
context.
B
D
Well,
the
way
it's
currently
implemented
in
in
out
instrumentation
repo-
it's
in
fact
it's
extremely
easily.
We
just
we
extract
spam
context
and
we
want
to
extract
spam
context
and
put
that
as
a
parent
of
the
newly
created
path.
So
I
can
just
use
an
empty
route
context
as
a
destination
for
that
spam
context
to
get
that
set
as
parent
and
be
done
with
that,
Nikita.
H
Remember
I
pointed
you
that,
yes,
it's
nice,
that
we
returned
context
because
it's
a
standard
object,
but
we'll
have
this
much
problem,
and
this
is
exactly
I
feel
exactly
what
Nikita
is
hitting,
because
he
had
something
in
the
context.
He
gets
some
values.
He
does
not
know
what
values
he
gets.
So
it's
very
hard
to
to
to
merge
these
two
results.
So
I
think
I
think
there
is
a
merging
problem
between
what
the
what
was
before
versus
what
the
propagator
returns.
H
Why?
Because
propagator
should
work
for
any
value
that
you
have
in
the
context,
not
only
for
spans.
We
want
to
propagate
the
correlation
context
as
well.
We
want
to
propagate
other
parts
of
the
context
with
that
and
would
have
been
a
bit
uglier
the
interface
to
every
propagator.
To
return
the
whole
thing
we
can
discuss,
Nikita
I
think
one
option
could
be
to
return
key
and
object
from
the
context,
and
then
you
can
do
the
manual
managing
of
the
things.
E
All
the
different
continents,
by
the
way,
that's
what
I
was
not
sure
about
that.
Okay,
now
I
know
you
are
oh
I
with
you
Carl,
but
not
with
boredom
Carlos.
H
How's
the
pain
in
August
think
about
a
problem
think
about
the
problem
that
we
just
mentioned
with
merging
and
the
fact
that,
because
we
turn
context,
we
have
something
previously,
we
return
the
context
from
the
propagator.
How
do
we
merge
them?
Remember
exactly
this
I
was
asking
you:
how
do
we
know
that
this
guy
has
something
in
this
context
that
we
need
to
merge
with
the
previous
one
in
sure.
E
H
C
C
I,
don't
think
I
follow
killer.
If
did
please
I
mean,
for
example,
you
have
like
a
lot
of
issues,
you
don't
know
how
should
it
be
nurse
yes,
but
we
can
like
implement
the
difference
that
you
destroy
or
give
a
user
sunlight
custom
implementation
on
user
to
implement
the
different
strategies
for
merging
it
and
depend
on
choose
strategy.
It
will
be
different.
H
H
E
B
B
Yes,
it
took
about
two
and
a
half
hours
to
go
through
all
of
there
mornings
and
get
them
cleaned
up
thanks
for
being
there,
yeah
I
was
wondering,
if
I
kind
of,
had
to
make
some
executive
decisions
on
the
settings,
because
there
are
at
least
a
couple
that
I
don't
think
applied
like
the
obsolete
JDK
or
whatever
it
was
one
that
was.
It
was
a
funny
one.
It's
like.
B
E
D
E
E
B
E
H
B
B
Fully
qualified
names,
if
you
could
import
them
and
not
allowing
importing
of
inner
classes
and
actually
I,
think
that
I,
although
I,
don't
agree
with
the
not
allowing
importing
of
inner
classes
JavaScript
the
Javadoc
generator
gets
very
confused
with
if
you,
if
you
do
import
with
them.
So
it's
probably
a
good
idea,
because
Java
the
Java
doc
doesn't
work
very
well
with
that.
So
yeah.
E
B
E
B
E
H
It's
only
at
the
beginning
that
you
can
set
the
resource.
There
is
another
item
to
set
it
in
the
cloud
with
whatever,
but
that's
maybe
but
again
not
I
would
not
worry
to
not
just
simply
not
accept
it
like
just
ignore
that
I,
don't
think
I,
don't
think
there
will
be
the
case
where
there
is
something
said,
and
somebody
will
over
life.
Let's
put
this
really
because.
A
B
And
bought
yeah,
we
don't
air
out
right
now,
I
think
bargain
is
saying
we
should
ignore
them,
which
is
probably
true
and
I.
Think
the
question
is:
do
we
need
fancy
logic
like
we
have
for
span
attributes
when
we
delete
them
and
I
agree?
I,
don't
I
can't
think
of
a
use
case
where
you'd
want
to
remove
a
resource
that
attribute
once
it's
set
yeah.
A
B
So
it
would
be
one
of
these
things
where,
if
you
had
like
an
auto
detector,
you
had
to
Autry's
or
solder
detectors
like
look
at
that
auto
detector
spec
and
they
both
were
trying
to
propagate
the
same
key,
and
one
of
them
knew
more
than
the
other
one
or
something
and
decided
that's
not
valid
and
deleted
it
I.
Don't.