►
From YouTube: 2021-04-08 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
B
A
A
D
B
E
Oh
fabulous
yeah,
modifying
the
context,
handling
and
just
created
the
pr
just
now
so,
oh
nice,
it
has
taken
like
three
three
or
four
days
of
work:
oh
wow,
so
yeah,
let's
see
if
it
improves
things,
and
I
I
I
I
couldn't
er
fix
your
networking
issue.
The
header
thing
you
because,
as
you
as
I
wrote
down
in
the
in
the
slack
yeah,
it
was
in,
it
was
in
inserting
the
wrong
feathers.
E
E
Should
it
should
it
should
use
the
span
of
the
request,
because
you
you
must
follow
that
span,
not
not
the
parent
of
the
request
so
yeah.
I
have
to
modify
the
instrumentation.
E
I
did
it
like
that,
but
didn't
so
that
I
did
it
like
that.
So
you
could
recover
the
headers
that
you
are
injecting
if
you
lock
before
after
creating
the
span
yeah,
but
it
must
be
done
in
the
opposite
side
and
you
you
must
recover
that
feathers
in
a
different
way,
because
if
not,
we
we
cannot
get
the
the
current
span,
but
yeah
the
the
fix
should
be
easy,
but
I
hope
I
was
expecting
to
finish
this
pr
a
bit
before
so
yeah
it.
It
took
a
bit
more
than
expected.
E
So
let
me
open
the
document
for
today
james
nice,
to
meet
you
sorry
that
I
didn't
say
anything
nice
to
meet.
D
E
It
was
a
really
nice
pr,
the
one
you
provided,
the
two
pr's
you
provided
so
yeah
welcome
to
the
to
the
team.
If
you
want
to
start
contributing
regularly,
you
are
more
than
welcome.
B
Plan
yeah
I've
been
I've,
been
toying
around
with
the
url
session
config
and
and
I'm
not
sure
if,
if
there
right
now,
I
walked
away,
it
was
working,
it
was
working
decently.
Well,
even
with
these,
these
strange
issues
that
you
were
describing,
I
was
getting
data
sent
up
to
the
elastic
collector
and
it
was.
I
was
getting
data
visualized
it
was.
It
was
a
little
wonky,
I'm
toying
around
with
view
controller
instrumentation
to
like
kind
of
get
a
automatic
wrap
of
network
connection.
I'm
not
a
big.
D
B
Of
it,
I'm
not
gonna,
I'm
probably
not
gonna
push
it
up
to
the
shared
repo
because
it
is
pretty
hokey.
B
But
now
I
come
back
in
the
morning
and
it's
dead,
locking
in
like
the
like.
The
trace
parts
are
like.
So
I'm
like
what
what
did
what
happened?
I'm
a
little
worried
that
it
might
have
to
do
with
the
way
that
it
stores
data
persist,
data,
I'm
not
sure.
If
there's
any
data
persistence
in
there
for
the
you
know
trace
handling
stuff,
because
I
did
just
unplug
my
phone
while
it
was
connected.
E
E
E
You
can
see
it
right.
Okay,
so,
let's
see,
let's
follow
the
the
topics
of
the
last
meeting.
If
there
is
something
that
and
review
them
and
continue
with
what
we
have
for
today,
if
you
have
some
some
other
topic,
you
want
to
talk
or
to
add
you
know
the
document
is
public,
so
you
can
edit
yourself.
E
If
you
want-
or
you
can
just
say
about
it
and
we
we
will
add
it
here:
okay,
so
yeah
for
last
first
topic:
yeah,
it
was
the
issue
that
james
talked
about
using
the
own
instrumentation
of
url
session,
not
the
experimentation,
but
the
timing,
information
that
it
can
provide
with
the
delegate.
E
E
But
if
you
are
using
a
non
non-delegate
url
session
configuration
it
will.
I
expect
that
it
won't
check
for
the
method
and
probably
won't
be
called
so
even
if
we
switch
this
method,
it
won't
work.
Always
that's
my
concern,
that's
why
I
expect,
but,
as
I
said,
I
have
not
written
that
code
to
check.
If
that's
true,
it
shouldn't
take
very
long
when
I
can
take
on
it.
So
yeah,
probably.
B
Would
it
be
an
option
to
just
inject
our
own
delegate
if
one
doesn't
exist
or
overwrite
it?
I
have
a
sort
of
a
pattern
that
I
used
for
that
in
the
ns
url
instrumentation
that
I
submitted
or
like
originally
not
as
good
as
yours,
but
it
does
have
a
sort
of
pattern
in
there
for
for
overriding
the
delegate,
which
maybe
we
could
provide
as
an
option
if
we,
if
customers
or
users,
might
want
to
utilize
that.
E
Okay,
it's
not
about
rewriting
the
delegate
because,
but
they
exist,
that
delegate
must
exist.
B
E
Yeah
we
can
yeah,
we
can
check
if
it's
possible,
to
force
a
delegate
into
the
session
if
it
doesn't
have
one
that
that
would
be
the
only
possibility.
I
don't
know
if
that's.
B
Yeah,
the
the
the
way
that
I
had
done,
it
was
over
by
swizzling
the
actual
init
function
and
then
wrapping
any
incoming
or
any
delegate
that
is
passed
with.
You
know
like
a
instrumentation
delegate.
E
B
Yeah,
it
can
be,
it
can
be
a
little
awkward
though
you
can.
You
can
set
it
up
where
the
the
instrumenting
delegate
will
forward
messages,
that
it
doesn't
respond
to
directly
to
the
original
delegate
and
then
pass
through
any
of
the
messages
that
you
know
we're
instrument
or
we're
interested
in
through
our
code
first
and
then
to
the
the
original.
E
B
E
Yeah,
that's
right.
We
could
provide
a
delega,
an
empty
delegate.
Maybe
if
we
want
to
provide
this
to
the
user,
so
they
can
instantiate
a
delegate
that
just
exists.
So
it
gives
that
that
gives
that
for
us,
but
yeah,
that's
the
kind
of
behavior
that
I
think
that
will
break.
If
we
add
forcefully
at
the
delegate.
C
In
my
investigation
at
work
we
actually
added
it.
It
was
optional,
so
there
was
a
second,
we
essentially
publicized,
a
class
which
was
the
delegate
that
the
clients
could
choose
to
use
if
they
wanted
to.
If
they
chose
not
to,
then
they
would
get
like
the
basic
set
of
information
to
what
you've
already
got,
where
they
chose
to
use
the
custom
delegate,
and
they
would
get
a
little
bit
more
extra
details
which
was
beneficial.
E
Yeah,
the
idea
is
trying
not
to
make
the
user
to
change
their
code,
at
least
for
the
first
version,
so
yeah,
that's
yeah,
that's
something
to
to
check
okay
more
more
about
this.
Anyone
at
this
point.
E
Okay,
the
next
is
yeah
the
review
of
the
implementation-
I
I
know
he
finished
it,
and
I
now
I
with
this
last
pr
I
created
today.
I
think
we
are
covering
almost
all
of
them.
E
B
E
That
that's
that's
one
of
the
because
I
know
you
you
approved
one
of
my.
Can
you
see
my
you
can
see
github
right.
B
E
So
I
have
one
here:
this
is
the
one
you
approved.
I
think
you
approved
it
right.
C
E
Yeah,
I
think
that
you
must
be
in
the
in
the
code
owners
file.
You
have
to
be
an
approved
maintainer
or
approval
by
the
project,
and
there
are
some
some
topics
about
the
level
of
your
what
you
must
have
provided
to
the
project
before
being
one
of
those
it's
in
the
open,
telemetry
docs.
E
I
don't
know
what
where's
that,
but
you
must
do
like
10
pr,
something
like
that
and
continuous
work
and
blah
blah
blah
blah
blah.
Okay,
you
can
check
it's
in
the
docs
in
the
project
page
the
level
for
the
people,
and
I
think
that
that
need
to
be
there
and
added
to
the
code
owners
and
must
be
approved
by
other
approvals
or
committee
people
or
something
like
that
so
yeah.
I
think
it
will
need
some,
maybe
more
time
of
working
the
project
before
having
that
right.
E
E
The
code
ordnance
file
so
yeah,
it's
fun.
Okay.
The
next
point
discuss
how
to
get
started
with
yeah.
B
It's
just
yeah
vanad
and
I
were
just
chatting
about
how
he
can
you
know,
get
started
with
the
swift
project
and
you
know
generate
some
data.
Were
you
successful
at
all,
but
none.
A
D
Yeah
yeah
yeah,
I
mean
I
I
I
haven't
you
know
but
yeah
I
got
the
basic
stuff
all
set
up,
so
yeah
I'll
I'll
be
productive.
Soon,.
E
One
of
the
things
that
I
am
changing
that
maybe
it's
a
big
change
for
you.
I
don't
know
it's
about
the
scope
thing
and
setting
active
expanse
and
not
active
expands.
So
one
of
the
changes
I
have
done
currently
it's
it's
something
that
it's!
Oh,
no,
it's
in
the
next
topics.
We
can
talk
about
that
when
we
reach
that
topic.
Okay,
yeah,
sorry,
I
just
so
something
that
will
change
for
the
starting
thing
and
the
active
span
thing.
E
One
of
my
topics
is
about
what
to
do
with
active
sponsoring.
Now
I
have
removed
the
scope
concept
so
when
you
end
a
span
that
has
an
associated
a
context
that
context
is
closed
with
the
new
implementation.
So
there
is
no
any
scope
you
have
to
close,
but
it's
closed
when
the
span
that
created
the
the
active
span
that
created
a
context
is
ended
in
the
okay.
B
That
was,
that
was
something
I
noticed
was.
I
was
checking
to
see
if
there
was
an
active
span,
but
ending
an
active
span
wouldn't
clear
out,
like
the
active
span
back
to
back
to
nil
or
whatever
yeah.
So
that's
good.
E
Yeah
in
depending
on
the
implementation,
it
had
that
scope
as
a
parameter
of
the
span
that
you
had
to.
If
you
created
a
a
spam,
subclass
or
or
your
own
span,
you
had
to
close
it
yourself.
E
Done
automatically
with
not
where
you
have
to
call
it
at
the
end
of
the
span,
you
have
to
clean
it,
but
you
don't
need
in
the
implementation
of
the
span.
You
need
to
call
it
at
the
end
also,
but
with
a
method
instead
of
having
a
variable
that
you
must
close.
E
E
That
was
one
of
the
things.
The
other
is
that
there
is
no
the
scope
concept
in
the
library,
it's
not
in
the
api
some.
So
now
now
the
context
is
hundred
in
the
open
telemetry
instance
itself,
where
you
can
create
with
a
different
context
provider
right
now
there
are
no
more.
So
that's
not
very
much
finished,
but
it's
it's
an
initial
way
of
supporting
more
yeah
and
there
are
some
changes
there
in
the
behavior,
but
the
test
passes
all
now.
E
E
Yeah,
I
think
it's
better
than
it
was
before
and
covers
better
something.
So,
okay,
so
yeah,
let's
go
with
the
topics
today:
yeah
location
for
importers
or
instrumentation
that
james
talked
about
in
he
in
one
of
his
peers,
yeah.
C
Good
to
go
yeah,
so
this
one
was
basically
thinking
that
if
I
added
the
swift,
metrics
shim
layer,
kind
of
like
a
bridge
between
their
framework
and
hours
apple
has
three
different
frameworks
between
the
tracing
metrics
and
logs
one
for
each
little
category
there.
So
I
feel
at
some
point
we're
going
to
probably
want
importers
for
all
of
those
as
well
and
of
course
we
do
have
the
open
tracing
shim
as
well,
because
it's
been
in
for
a
while.
C
E
I
don't
know
if
using
instrumentation
for
everything
being
an
importer,
just
a
kind
of
instrumentation,
so
it
is
it's
instrumenting.
Another
library
of
you
prefer
having
that
as
important.
Yet
just
that's
my
only
my
only
doubt
about
it
having
two
importers
and
instrumentation
and
just
having
one
folder
for.
C
I
would
count
out
class
importers
as
being
other
telemetry
style
packages
and
where
we're
taking
in
their
data,
transforming
it
and
then
exporting
it
again,
whereas
instrumentation
is
actually
generating
new
data
from
other
packages
or
bits
of
information,
whether
that's
a
system
level
with
obviously
url
session,
or
I
can't
think
of
an
example
of
a
third
party
photo,
but
I'm
sure
that
would
be
one.
B
Looking
at
I'm
just
looking
through
a
couple
other
ones
to
see
if
they
have
anything
like
that,
maybe
maybe
we
should
double
check
just
to
see
like
how
other
open
telemetry
projects
are
dealing
with
with
the
idea
of
importers.
C
E
That
that
talked
about
having
some
important
instrumentation
in
the
main
repo
was
was
was
nice
to
have
so
I
think
we
can.
We
can
have
both.
For
me.
It
works.
I
mean
having
importers,
has
a
importer
of
another
kind
of
telemetry
inside
top
and
telemetry
and
instrumentations
are
only
ingestion
of
expansion
or
tax.
C
E
Define
it
definitely
it's
it's
crazy.
Spm
is
just
I
mean
it's
still
not
easy
to
work
with
when
you
have
a
really
a
real
project
independent.
It
has
a
lot
of.
E
So
I
think
that
currently
having
just
one
project
is
better,
at
least
in
the
meantime
and
yeah.
So
let's.
B
Yeah
yeah
I've
got
no
issue.
C
E
Yeah,
the
the
the
apple
tracer
looks
nice,
but
my
concerns
is
that
the
respect
is
going
to
change
very
soon
and
probably
with
things
that
won't
be
compatible
with
our
current
version
of
the
ios
or
mac,
so
that
that
concept
concerns
me
a
bit.
E
We
want
to
support
other
versions,
probably
our
clients
or
our
apps
will
need
to
run
in
os
13
now,
and
if
this
technology
comes
only
for
os
15,
we
shouldn't
use
that
in
the
in
the
framework,
apple
being
apple,
usually
usually
only
release
for
the
newest
version,
but
that's
not
good.
When
you
have
a
technology
that
must
be
used.
That's
my
only
concern
with
that.
I
told
that
in
the
pr
and
sometimes
yeah.
B
I
was
just
going
to
say
that
might
be
a
that
might
be
a
good
reason
why
we
might
want
to
pull
it
out
into
a
separate
pack
package
so
that
we
can
define
those
os
requirements,
so
it
doesn't
affect
the
whole
open,
telemetry
project,
but
I
mean
that's
something
we
can
worry
about
later
too.
D
I
think
I
think
this
is
probably
something
similar
on
the
dot
net
open,
telemetry
project,
because
you
know
microsoft
has
some
instrumentation
on
the
profile
right,
which
you
know
they
want
to
use
some
of
it.
You
know,
I
think
that
is.
We
can
probably
follow
some
of
that
pattern
here.
D
The
way
we
they
use,
the
microsoft
metrics
out
of
the
their
profiler
and
add
it
to
the
open
tracing.
Oh
sorry,
open
telemetry,.
E
Okay,
okay
and
here
I
also
write
down
three
issues
that
oh,
that
carlos
opened
in
in
the
project
that
I
think
that
we
should
decide
here
what
we
do
with
that,
because
it
has
some
implications-
and
I
didn't
wanted
to
take
a
decision
myself
with
with
that.
I
think
that
they
make
sense
to
talk.
One
of
them
was
consider
splitting
out
experimental
packages.
E
So,
as
you
know,
some
of
the
libraries
are
experimental.
He
says
to
use
a
different
version
or
having
a
suffix
for
it.
E
As
an
idea,
I
having
different
versions,
mean
having
different
packages,
so
I
think
that
we
or
at
the
suffix
that
will
also
break
in
the
future,
but
maybe
it's
what
we
want
just
to
know
that
you
are
moving
to
a
non
experimental
thing
or
alpha.
E
But
what's
your
opinion?
I
I
think,
if
we
do
something,
we
have
three
options:
moving
some
packages
of
out
of
the
project
in
another
different
project,
adding
a
suffix
or
just
do
nothing
and
say
to
carlos
that
we
just
document
it
in
the
rhythmic
and
people
should
know
about
it.
C
One
example
would
be
firebase,
they
have
experimental
packages
as
part
of
their
definition
and
they
do
the
suffix
approach.
They
just
add
a
beta
to
the
end
of
them,
and
then
they
also
document
them.
I
feel
like
that's
a
fairly
sensible
move.
I
think
versions
causes
too
many
issues
just
from
a
client
integration
perspective
convoluted
what
we
talked
about
a
minute
ago
as
well.
B
Yeah
would
that
be
like
the
in
the
swift
package
manage
or
in
the
the
package
file?
Would
we
name
the
the
portions
that
are
experimental,
dash,
experimental
in
there
or
okay,
and
then
the
versions
alternatively,
would
be?
Would
there
be
like
a
separate
version
branch
with
those
experimental
files
in
the
in
the
github
or
in
the
repo?
Is
that
the
idea,
those
would
be
the
different,
the
the
different
choices.
C
Okay,
okay
or
all
three
of
them
kind
of
break
down.
In
my
opinion,.
B
E
Yeah,
I
think
that
it,
it
will
be
easier
to
just
add
a
suffix
that.
E
Having
different
branches,
because
we
will
have
to
keep
synchronization
there
if
we
want.
E
Agreed
yeah
so
yeah,
so
let's
rename
it
we
put
it
alpha
experimental.
What
do
you
think
it's
better.
C
B
Yeah,
maybe,
depending
on
yeah,
depending
on
what
it
is,
you
can,
because.
B
E
Oh
yeah,
so
so
you
prefer
alpha.
You
prefer
experimental.
B
I
think
for
things
like
the
the
metric,
just
because
I'm
more
familiar
with
that
as
it's
like
totally
out
of
spec-
maybe
it
should
just
be-
it
should
just
be
like
experimental,
whereas
things
that
are
like
inspect
but
maybe
not
complete
are
alpha,
I'm
not
sure.
E
Okay,
next
one.
E
E
I
think
we
don't
want
to
have
own
packages,
as
we
have
said.
Maybe
this
was
something
that
that
it's
an
extension
to
the
span,
the
one
that
he
pointed
out.
So
probably
we
can
remove
this
code
because
I
don't
know
anyone
using
are
you
using
it?
E
E
So
probably
so
I
don't
know
if
there
are
some
other
code
like
this,
but
I
I
think
that
we
say
we
could
say
in
the
document
that
semantic
conventions
is
not
totally
defined
as
it
is
now,
but
I
think
that
we
are
following
the
spec
with
the
pr
from
from
jen
james,
so
just
say
that
and
delete
that
code
that
he
points
out
thing
could
make
sense
for
you
yeah,
or
do
you
think
that
he's
expecting
some
other
changes?
I
I
don't
know.
C
B
I
yeah
I'm
not
sh.
I
wish
that
he
had
linked
to
like
some.
Maybe
maybe
we
can
ask
him
for
some
clarification
on
what
he's
referring
to
as
semantic
conventions
like
a
link
to
a
to
a
spec
doc
or
something
because
I'm
wondering
if
this
is
just
like
kind
of
like
helper
methods.
That
he's
has
issue
with
that
aren't
in
spec
or
if
or
if
he's
actually
talking
about
semantic
attributes
as
well.
B
Yeah
yeah
and-
and
maybe
you
could
ask
if
there's
like
a
a
reference
spec,
that
we
could
that
we
could
use.
C
This
more
on
the
topic
of
semantic
combination,
but
it
kind
of
I
had
an
idea
when
I
did
my
pull
request
that
automated
the
the
pulling
in
of
that
date.
So
I
wanted
to
kind
of
see
what
you
felt
for
about
it
as
part
of
the
the
specifications
in
the
ammo
files.
We
get
the
return
type
out
of
it
at
the
moment.
I
just
put
that
as
a
comment
to
kind
of
say
that
you
should
use
a
string
here,
you
could
use
an
interior
cetera.
C
C
So
I
feel
like
the
course
I
would
essentially
be
like
semanticconventions.int
dot,
http
port,
for
example.
The
reason
that
that's
important
is
because
then
elsewhere
in
the
project,
we
actually
have
helper
functions
on
the
span,
which
is
to
set
attributes,
and
then
it's
the
enum
that
you
can
use
and
then
the
value
to
it.
C
E
Okay,
the
next
issue
here
this
one.
E
He
says
that
as
an
optimization
rather
than
correctness,
be
because
they
exist
as
primitives
in
other
languages.
E
E
I
don't
know
what
what's
your
opinion
about
it?
I
think
they
need
that
for
java
and
python,
because
those
are
languages
that,
with
their
dynamic
nature
and
dynamic
memory,
it's
probably
better
to
have
that
in
a
common
type.
But
I
don't
think
for
swift
that
really
makes
sense.
B
Yeah,
I'm
not
I'm
not
sure
what
we're
optimizing
for
as
what
carlos
calls
out.
So
I
don't.
I
don't
think
we
need
to
remove
them.
I
I
kind
of
I
kind
of
like
being
able
to
interact
with
them
like
it.
You
know
print
them
out
as
strings,
but
then
they're
kind
of
living
as
integers
under
the
hood.
E
B
E
Great
excellent,
this
one
active
span
active
spam.
This
is
a
quite
nice
thing.
I
have
talked
about
it
a
bit
at
the
beginning,
but
currently,
when
you
create
an
span,
the
span.
E
Is
not
an
active
span,
so
the
the
thing
is
that
if
you
create
a
span,
it
will
take
the
current
span
as
its
parent.
If
you
don't
provide
one,
so
you
don't
need
to
know
inside
a
method
if
there
was
a
parent
span
created
that
just
call
that
method
in
that
class,
because
if
there
is
one
and
it's
active,
it
will
assign
it
as
a
parent
automatically.
E
So
you
don't
need
to
pass
the
parent
spam
context
yourself,
because
the
active
spam
works
for
you
and
that's
something
that.
B
E
The
thing
is
that
you
have
to
make
it
yourself
active.
I
really
think
that
we
could
make
a
active
by
default.
So
when
you
start
an
span,
it
automatically
runs
that
code
or
you
can
start
as
root
span,
so
it
doesn't
run
that
code
or
just
add,
or
just
or
also
you
can
in
the
builder
in
the
spam
builder,
configure
it
not
to
make
it
active
but
make
it
active
by
default.
E
So
whenever
you
create
a
span,
it
will
all
the
childs
all
the
spans
created
after
that.
In
the
same
context,
if
that
one
has
not
ended,
will
be
child
expanse
of
that
span
automatically.
B
E
B
E
That
when
that
span
is
created,
and
so
everything
that
happens
after
that
will
will
be
related
to
that
active
response,
so
you,
the
idea,
is
handling
the
context
at
the
same
time
that
the
span
is
created
or
ended
by
default.
E
C
E
I
mean
you,
you
can
create
already
a
root
span
or
you
can
create
a
span
with
with
a
given
parent,
so
you
can
set
the
parent
automatically
for
a
span.
C
E
Yeah
right
because
you
want
a
root
span,
it
does,
but
sometimes
you
don't
know
I
mean
sometimes
your
code,
you
don't
know
if
you
have
a
parent
spam
if
your
method
is
called
from
another
code
that
has
another
span.
So
you
want
that
to
be
your
parent.
B
So
I
think
I
think
final,
just
posted
something
from
the
spec
that
might
just
nip
this
all
in
the
bud.
It
says
in
languages
with
implicit
context.
Propagation,
I
think
that's
that
covers
us
right
spans
create
our
span.
Creation
must
not
set
the
newly
created
span
as
the
active
span
in
the
current
context
by
people,
but
this
functionality
may
be
offered
in
addition,
as
a
separate
operation.
B
So
maybe
we
could
add
a
like
set
active
span
flag
on
the
on
the
on
the
span
builder.
E
Okay,
but
we
must
do
it
actively
right.
It
says
that
okay,
I
didn't
check
the
spec,
so
the
spec
says
that
you
must
not
make
it
the
by
default.
The
active
span.
E
E
E
E
Yeah,
okay,
yeah,
I
think
it
it
will
improve.
Also
this
at
least
you
handle
the
state
before
creating
the
span.
You
don't
have
to
start
it
and
after
that,
make
it
the
active
yeah.
I
think
it
it
will
improve.
Also
cool,
okay,.
E
So
that's.
E
What
sorry
there
is
one
more
sorry,
the
mobile
specific
attributes.
C
You're
not
not
strictly
for
the
swift
repo,
but
as
a
mobile
language.
It
does
kind
of
come
down
to
impact
us
a
bit.
There's
someone
I
brought.
I
know
you've
looked
at
this
one
already,
I'm
gonna
put
it
up
from
the
planet,
a
few
comments
on
it,
but
it
might
be
good
to
quickly
chat,
make
sure
I
haven't
missed
anything
else.
C
Essentially,
it's
adding
a
couple
more
semantic
attributes
that
are
specific
to
mobile
and
something
that
would
make
good
use
of,
especially
in
our
automatic
instrumentation,
and
to
make
it
a
bit
easier
for
mobile
clients
to
get
onboarded
the
main
ones.
I've
added
is
os
type,
that's
not
name,
sorry.
C
So
that's
basically
ios
watch
os
ipad
os
that
kind
of
stuff
you
can
get
quite
easily
I've
added
os
version
as
well
so
14.1
and
that
one's
machine
readable
because
it
says
there
is
one
already,
but
that's
not
machine
readable.
So.
B
B
Oh
yeah,
I
was
because
that
is
like
a
reese,
a
re,
a
respecting
of
that
attribute
right
because
because
they
already
have
os
version
defined
as
non-machine
readable.
B
C
And
I've
also
added-
and
this
one
has
changed
a
few
times
with
all
the
people's
comments,
but
model
identifier
and
model
name
and
then
also
the
model
identifiers.
C
Identifier
for
you
protecting
unique
users.
C
B
Yeah
the
the
way
I
like
to
use
it
is
grouping
events
or,
like
you,
know,
a
session
as
well
that
that
helps
as
well
to
see
like
what
yeah
so
but
yeah.
I
usually
use
like
vendor
id
for
that,
but
that
can
be
a
little
spicy
these
days
right.
C
One
of
the
comments
did
say
about
the
gdpr
from
the
day
of
directing
that,
so
I
have
that
I,
it
is
something
that
does
apply
to
more
attributes
than
just
this
new
one.
It
just
seems
to
be
the
first
time
it's
been
picked
up
on,
for
example,
like
the
host
type
in
all
that
sort
of
stuff,
depending
on
the
use
case.
It's
definitely
personal
information,
so
I
feel
like
that's
going
to
need
to
have
like
a
bigger
conversation
with
people
who
are
like
actively
involved
in
the
specification
but
yeah.
B
C
E
Yeah,
I
think,
definitely
is
useful
information.
Then
we
need
that
for
devices
yeah
we
I
I
have
this
kind
of
tasks
in
my
own
implementation,
for
I
I
can
imagine
all
of
us
have
something
like
that.
So
yeah
having
something
common
will
be
will
be
great.
I
don't
know
if
we
can
add
some
classes
or
something
I
don't
know.
Yeah.
C
B
B
Yeah,
I
don't
have
anything
else.
E
Yeah,
you
see
you
have
this
approved
the
pr.
You
told
that
you
approved
it,
but
I
cannot
merge.
E
I
don't
know
sometimes
he
approves
fast,
but
sometimes
he
he
takes
more
time.
I.
B
Don't
know
well,
I
wonder
if
he
saw
the
check
mark
that
I
put
on
there
and
was
like
I
don't
need
to
deal
with
this.
Maybe.
B
E
E
I
will
wait
until
he
if
he
approves
other
peers.
I
will.
I
will
ping
him
in
this
one
or
something
like
that.
I
don't
know
okay,
so
I
think
that's
all
and
we
are
almost
on
time
any
other
stuff,
maybe
check
in.
If
we
have
some
other
issue
here
that
we
haven't
tackled
today
we
have
the
spanish
scope
that
is
with
1pr.
Now
we
have
talked
about
this
one.
E
E
We
have
talked
about
this
one
also-
and
these
are
two
are
with
pr,
so
I
think
we
are
more
or
less
done
with
the
recent
issues.
So,
let's
see
if
we
can
move
to
a
having
a
1.0
beta
version
soon
and
if
they
we
can
get
some
public
noise
from
the
open,
telemetry
community
or
something
like
that
and
get
more
people
using
this.
E
So
I
think
it
it
needs
usage
for
for
for
issues
to
to
appear
and
nowhere
where
things
are
not
working
as
expected.
C
Yeah
we're
planning
on
throwing
into
one
of
our
sample
apps,
hopefully
the
next
few
weeks,
well
sooner
or
later,
but
yeah
I'll.
Let
you
know
how
that
goes
over
here:
issues:
okay,
great
yeah,.
E
I
am
using
this
for
a
project
data
doc.
Currently
and
it's
I
have
been
working
with
this
coding
sort
of
production
code
for
a
long
time
and
tracer
things
are
more
or
less
working
working.
Well,
the
other
areas
are
less
tested,
but.