►
From YouTube: 2022-01-13 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
Now,
from
that
anything
we
want
to
talk.
Probably
there
will
change
this.
You
can
briefly
test
about.
C
So
regardless
this
has
been
an
ongoing
discussion
for
long.
So
what
exactly
is
happening
is
probably
just
if
we
see
our
trace
api
or
we
can
just
go
there.
Why
do
we
in
the
documentation.
C
This
returns,
the
start
span
returns
a
shade
pointer
with
which
is
I
mean
there
have
been
lots
of
us
coming
that.
Why
should
it
be
telling
me
chase
pointer?
Why
not
just
return
a
reference
to
a
span
or
maybe
spam
the
value
on
the
beginning
pointer,
not
not
the
shape
for
performance
reasons
so,
which
makes
sense.
I
think,
is
aware
of
this
discussion.
This.
C
Yeah
yeah,
this
change
does
breaks
api,
so
I
mean
the
reason
why
I
wanted
to
touch
upon
this.
Is
that
they're
having
some
some
again?
Some
questions
raised
yesterday.
C
C
C
That's
that's.
That's
the
reason
why
we
return
it
as
a
shape
pointer
and
that's.
Why
do
we
store
it
in
a
straight
thread?
Local
storage
is
because
we,
because,
if
somebody
creates
another
span,
nested
span
which
and
now
the
initial
span
the
outer
span
should
be
the
parent
of
that.
So
through
the
thread,
local
storage
context
manager,
we
can
already
we
can
get
the
parent
span
and
we
can
use
that.
C
C
What
I
think
is
this.
This
point
is
coming
again
and
again-
and
yesterday
was
somebody
raised
this
okay,
so
the
set
project
is
using
it
I
mean.
Basically
I
mean
it's
it's
it's
it's
the
object,
storage
database,
just
like
hadoop
or
something
so
they
have
been
using
it
and
I
think
they've
found
some
performance
issues
and
then
they,
just
probably
they
just
commented
it
there.
They.
C
I
don't
know
how
they
tested
it.
We
could
ask
in
using
non-shared
pointers,
fans
knocks
pants.
I
don't
know
how
they
did
this
test,
but
they
said
that
they
have
a
dedicated
impact
on
performance.
We
can
check
with
them.
What
exactly
was
the
impact?
Where
exactly
was
impact,
but
I
was
thinking
I
mean
this.
Just
probably.
C
C
So
the
only
operation
which
would
be
allowed
would
be
just
create
a
span
populate
the
span
and
and
end
the
span
or
it
will
be
automatically
ended
if
it
goes
out
of
the
spokes
code,
so
that
will
will
at
least
remove
the
use
of
the
share
pointer.
It
will
just
give
a
unique
pointer
and
if
somebody
wants
to
create
a
span
and
have
an
explicit,
they
have
to
explicitly
specify
the
parent
using
the
option
in
that
case
start
span
with
shared
and
then
in
start
span
options.
C
They
have
to
specify
the
parent
explicitly
so
somewhere
here
they
have
to
specify
the
parent
letter.
This
is
an
optional
field.
If
this
is
not
there,
then
we
will
get
the
parent
from
these
span
context
which
is
stored
in
thread
local
storage.
So
if
this,
if
we
specif,
if
somebody
wants
to
create
a
non-shared
fan,
then
they
have
to
provide
this
exposure
explicitly
or
otherwise
it
would
be
a
good
span.
C
That's
something
I
think
we
can
do
it,
but
probably
we
have
to
look
more
into
that.
I
mean
if
just
if
we
just
want
to
support
so
returning
and
on-share
span
with
some
limitations
that
it
cannot
be
made
active,
it
cannot
be
currently
up
to
span
and
there
are
limited
set
of
functionalities
you
can
use
with
that.
C
C
The
way
the
safety
is
using
is
they
don't
they
don't
set
that
span
as
an
active,
so
they
they
are
using.
They
create,
they
don't
use
the
nested
expanse,
they
just
create
the
span
and
they
populate
it,
and
then
they
end
it
where
they
don't
really.
We
don't
really
need
to
have
a
shared
pointer,
so
probably
some
performance
into
what
we
can
do
with
that.
But
I
think
it
just
needs
some
more
probably
testing
and
investigation,
whether
it
really
works
or
not.
C
B
B
C
C
Okay,
that's
one
of
the
things
dll
changes.
I
think
I
really
don't
see
it
before
that
you
can
add
it.
I
think
this
person
did
reach
out
to
us
again,
but
I
don't
see
any
solution
as
of
now
it
does
not
make
sense.
I
mean
his
approach
is
right.
That's
totally
fine!
If
we,
if
we
have
a
dot
cc
file
which
can
be
used
in
dll
and
it's
included
in
the
header
file
also,
but
it
does
not
make
sense
if
we
install
and
we
have
in
our
standard
header
file
location.
C
C
C
C
Yeah
so
he's
planning
to
add
an
external
library
which
would
be
wrapping
api
in
sdk
and
that
that
may
be
part
of
open
open
elementary.
I
don't
see.
C
C
C
C
B
This
is
one
thing
this
is
being
making
compiler
warnings
like
comments.
Are
you
giving
a
comment?
There's
no
comment
from
me
interesting.
B
Okay,
I
want
one
second,
so
this
thing
makes
a
compiler
warning
for
switch.
C
C
So
there's
a
deadlock,
I
mean
we're
in
a
situation
where
we're
waiting
on
the
response,
and
the
response
has
already
come
from
these
external
servers
and
before
we
really
start
waiting
for
the
response.
The
is
already
there
and
in
that
case,
waiting
for
response
goes
in
an
infrared
room.
It
just
keeps
on
checking
for
the
response,
but
response
has
been
already.
C
It's
a
good
point.
You
don't
see,
we
have
even
discussed
about
this.
C
C
C
And
that's,
I
think
this
one
is
still
working
progress,
I'm
going
to
close
it
because
I
think
I
have
to
rewrite
it.
I
think
it
took
lots
of
time
for
me.
I
did
raise
this
pr.
I
thought
I'm
going
the
right
way,
but
probably
I
think
the
book
I
have
to
rewrite
this
beer
anyway,
it's
just
a
draft,
but
I'm
going
to
raise
another
pr.
After
closing
this
a
bit
complicated,
I
mean
I
thought
probably
it
would
be
simpler
to
really
create
the
aggregation
and
api
the
sdk
changes.
C
Together,
yeah,
I
thought
of
doing
it.
I
mean
it's
very
difficult
because
unless
until
we
have
one
x
one
and
two
and
running,
we
can't
really
add
multiple,
but
but
let
me
let
me
try
to
do
it.
I
think
that's
your
right,
instead
of
instead
of
trying
to
spend
more
time
with
one
person
trying
to
spend
more
time,
if
you
can
really
split
it,
I
think
that
would
be
much
faster.
C
So
let
me
see,
let
me
logically
whatever
changes
I
have.
It
is
logically
come
to
make
it
logically
complete
by
today
tomorrow.
Let
me
let
me
raise
one
pr
on
that
and
then
I
think
probably
we
can
take
it
from
there.
I
can
raise
another
new
new
tasks
for
that
for
the
change
similar
changes
which
we
have
to
do
for
other,
like
I
probably
have
raised
it
for
histogram
and
then
we'll
be
reading
for
other
aggregation
methods.
C
C
C
So
yeah,
I
I'll
try
to
create
one
one
scenario
for
not
end
to
end,
but
one
one
logical
scenario
for
aggregation
for
histogram
that
won't
be
end-to-end.
So
it's
not
that
I
can
create
a
matrix
from
api
and
I
can
I
can
generate
the
complete
histogram,
but
that
would
be
aggregation
class
for
the
histogram,
which
will
we
could.
C
We
can
stated
substitute
afterwards,
but
we
can
create
similar
aggregation
for
other
methods
between
like
sums,
summarizations
plus
value,
so
got
it
right
I'll
create
one
one
aggregation
for
histogram
and
then
we
will
split
the
task
into
multiple
navigation
methods
which,
if
you
want
you
can
take
it
further.
B
Yeah
and
one
thing
for
this
one,
so
there
are
two
tasks
that
need
privilege
on
the
repo
okay,
yes,.
C
B
B
B
C
C
B
C
B
The
reason
for
drop
is,
is
I
change
from
debug
pin
to
release.
B
B
B
C
C
C
I'm
not
sure
I
tried
on
the
docker
image
whether
the
docker
maze
has
got
updated
with
some
other.
I
mean
with
whether
whether
docker
image
is
the
updated
update,
updated,
ubuntu
1804.
No,
I'm
sorry.
C
C
C
C
C
A
Maybe
we
can
implement
it
like
with
a
lock
like
log
to
to
implement.
I
think
before
this
is
a
new
c
plus
class
facial
right.
Let's
read
the
local
storage
that
keyword.
C
Static
magic
or
static
local
variables,
but
it's
very
specific
to
start
it
thread
locally,
yeah,
yeah,
okay,
I'm
not
sure,
because
what
what
they
have
done
is
they
just
added
these
flags?
No,
no
inline
and
no
clone.
So
probably
that
compiler
will
is
doing.
C
C
B
C
C
It's
coming
in
this
ubuntu
184
has
long
term
support
and
that
compiler
buggy
compiler
is
a
default
compiler
with
this
actually.
A
And
for
this
I
thought
this
type
of
bug.
I
think
that
should
be
a
like
a
back
report
from
gcc
site
right.
C
C
Thanks
thanks
we'll
also
see
it.
Probably
if
I
can
reproduce
it
again.
I
did
try
last
week,
but
I
think
we
just
want
to
try
once
more.
If
I
was
doing
something
wrong
and
asking
yes,
I
just
want
to
test
it
once
again,
but
please
please
have
a
look
and
if
anyone
you
anyone
have
a
solution
for
this,
please
feel
to
ping
it
here
and
assign,
and
we
can
assign
to
that.
I
mean
if
there
is
some
solutions.
C
Okay,
anything
else
to
discuss
otherwise.
C
Yeah,
I
saw
nothing
I
did
discuss
about
the
meeting
we
shared
sorry.
I
didn't
I
wanted
to
talk
about
that.
I
have
to
talk
about
this
once
again.
Probably
let
me
discuss
it
again
and
then
probably
let's
see
if
we
can
do
it
next
week,
if
we
can
report
it,
so
definitely
we
can't.
We
can't
change
it,
both
the
meetings
in
the
morning
time
I
mean
in
the
hour
for
morning
time.
C
Let
me
see
if
we
can
do
it
one
hour
earlier
or
probably
couple
of
hours
earlier.
C
C
A
For
this
mouth
release
yeah,
I
think
we
maybe
release
at
the
end
of
the
month
to
pick
up
some
change
from
metrics.
I
think
there
are.
There
should
be
some
more
yes.
C
C
Okay
thanks,
I
think
that's
if
we
don't
have
anything
else,
probably
we
are
going
to
close
the
meeting
thanks
thanks
for
joining.