►
From YouTube: 2020-07-07 .NET SIG
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
A
A
A
A
A
B
B
C
C
C
D
B
C
Yeah
I
think
we
can
start
didn't,
have
a
lot
of
agenda
today.
So
I
just
put
my
own
agenda.
There
was
nothing
else
added,
so
I
just
quickly
share.
It's
mostly
just
a
status
update,
so
I
updated
the
beta
tracker,
which
Mike
started
a
bit
ago.
It
should
be
like
reflective
of
what
we
have
done
and
what
we
still
have
missing
before
we
can
call
ourselves
beta
so
we'll
be
working
on
like
these
issues
this
week
and
next
week
and
I
tagged
like
Peter
us
like
the
timeliness
next
Friday.
C
So
hopefully
we
should
have
everything
ready
by
then
and
will
released
when
you
get,
and
then
we
like
take
the
next
set
of
work
item
for
making
us
GA
I
would
be
sharing
like
a
overall
structure
of
the
dogs
much
sooner
than
that,
probably
by
tomorrow
or
Friday.
It
won't
contain
the
full
dog
but
it'll.
At
least
it
should
contain
the
overall
structure
on
how
the
document
will
be
organized.
So
I
can
share
comments
early
enough.
C
That's
pretty
much!
What
I
wanted
to
talk
about
beta
tracker,
but
from
that
we
have
some
nice
updates
in
the
repo
which
we
moved
from
Devore
assured
orbs
to
github
actions
and
code
coverage
is
also
enable
now.
So
we
get
a
nice
comment
showing
the
code
coverage
I'm
still
to
figure
out
why
it
is
sometimes
showing
incorrect
values,
maybe
I'm
just
interpreting
it
wrong,
but
at
least
we
have
it
enabled
now.
So
it's
not
enforced.
C
So
whether
code
coverage
is
increased
or
decreased,
we
don't
it
doesn't
prevent
the
PR
from
being
merged,
but
we
make
it
a
requirement
sometime
later,
probably
after
beta
and
feature
wise.
We
have
like
one
PR
think
it
was
started
like
a
couple
of
months
back
and
it's
for
auto
instrumentation
for
gr,
PC
client,
so
I'll
I'm
just
merged
it
like
yesterday,
and
he
will
be
working
on
the
gr,
PC
server
side
instrumentation
as
well.
C
So
that
should
give
us
like
nice,
set
of
auto
collectors
or
instrumentation
adapters,
so
we'll
be
covering
HTTP,
client,
sequel,
client,
including
the
seeker
for.net
framework
and
GSPC
client
and
then,
of
course,
asp.net
and
a
speed
of
narco.
So
that
should
be
like
good
set
of
instrumentations
in
the
main
triple
have
one
cube
topic
to
us.
There
is
an
intern
in
Microsoft
who
wants
to
do
a
contribute
to
sampling,
and
it's
not
a
full
internship.
Internship
is
almost
done
like
just
a
couple
of
weeks.
C
Left
so
wants
to
make
some
productive
use
of
the
time,
and
the
proposal
is
to
implement
something
called
adaptive
smuggling.
So
the
name
is
from
our
monitor,
slash
application
insights.
So
essentially
the
idea
is
element.
We
currently
have
this
probability
at
the
start
of
time,
just
sits
on
top
of
it
politically
monitors
the
constantly
monitors
the
flow
of
overall
traffic
and
adjust
something
right
instead
of
for
a
user
perspective,
instead
of
specifying
the
probability
they
just
say.
C
This
is
a
volume
of
telemetry
I'm
willing
to
accept,
and
if
I'm
below
that
then
example,
everything
like
put
the
probability
200,
but
if
I'm
exceeding
that
I
just
assembling
to
a
lower
probability.
So
that's
essentially
the
idea,
and
since
there
is
not
enough
time
to
push
this
all
the
way
through
the
specs
I
thought
it's
best
to
just
do
it
in
the
our
country,
repo,
where
we
can
publish
a
separate,
no
get
anything.
It's
mostly
like
a
POC.
C
If
things
are
good,
then
one
of
us
can
take
it
to
the
spec
repo,
because
there
are
like
lot
of
discussions
around
something
in
one
of
main
specs
repo,
so
we
probably
able
to
find
some
place
for
adapt
recently.
So
if
there
are
no
objections,
I
will
ask
to
start
this
on
the
country,
repo,
not
the
main
metric.
E
C
I'm
not
sure
so
it's
going
to
be
implemented.
Don't
really
know
whether
there
is
a
notion
of
like
extensions
to
existing
some
little
product.
There
is
something
called
composite,
sampler
like
and
big,
which
can
embed
like
once
unloading
inside
and
but
I
haven't
really
followed.
The
spec,
so
I
am
short
on
details
on
this
one.
Like
do
you
have
a
specific
suggestion
on
how
it
should
be
done?
C
E
C
See
like
yeah
I'm,
not
sure
how
much
time
is
left
like
there.
Indeed,
she
pieced
actual
weeks
and
almost
eight
weeks
are
done.
So
it's
like
less
than
three
weeks
so,
depending
on
the
time
we
see
we
can
cover
like
poor
generally
rather
than
directly
volt
ammonia,
so
yeah,
that's
it
okay,
so
follow
has
some
topics
to
discuss.
Yeah
Paulo.
Can
you
walk
us
through
that
yeah.
D
F
Our
conversation
we're
at
the
priorities,
so
one
of
them
is
to
have
Auto
instrumentation
available
by
the
time
that
open
telemetry,
which
general
availability
and
I'm
reading
and
but
anyone
feel
free
if
they
disagree
or
think
that
is
not
a
priority,
should
not
be
guiding
the
direction
that
we
are
take.
I.
Think
now
is
the
moment
you
just
caused
them.
F
The
second
thing
that
comes-
and
these
are
not
in
order-
are
kind
of
just
the
things
that
I
think
from
our
conversation
stand
out
as
the
things
that
we
we
agree,
that
the
implementation
for
this
should
be
pluggable
and
pluggable,
in
the
sense
that
anyone
that
wants
to
kind
of
use
these
out
of
limitation
should
be
able
to
kind
of
relatively
easy
change.
This
part,
of
course,
this
is
not
a
design
so
I'm
not
discussing
if
these
blog
abilities,
which
at
vir
build
or
via
configuration.
F
This
is
something
that
we
discuss
later,
but
basically
the
tracer
that
is
the
equivalent
to
the
open,
telemetry
SDK.
That's
the
part
of
the
code
that
the
application
writes
is
the
part
that
we
want
to
Auto
cementation
to
make
kind
of
pluggable
the
context
propagation,
because
there
are
some
proprietary
formats
there
are
TT
see.
There
is
b3
and
perhaps
other
formats
that
people
need
to
support
in
context
propagation
and
the
data
exporter.
F
Implementations
of
integration
should
move
to
the
new
active
source,
because
this
is,
we
think,
that's
going
to
make
much
more
usable
and
easy
to
plug
the
other
components.
So,
instead
of
using
the
specific
type
yours
or
something
to
instrument
the
libraries,
we
would
like
to
use
the
active
source
directly
and,
of
course,
the
goal
is
to
have
a
lot
of
integrations
out
of
the
box
for
the
most
popular
frameworks
and
libraries
they're
out
there.
F
One
thing
that
was
a
non
priority
at
this
moment
was
supporting
the
CLR
instrumentation
engine
because,
while
that
is
desirable
because
of
the
limitation
of
a
single
profiler,
this
didn't
rank
high,
at
least
in
the
parties
that
talked
in
this
discussion
about
out
of
cementation
as
a
high
priority.
It's
desirable
in
the
future,
but
it's
not
something
that
is
considered
a
priority.
Having
the
auto
store
meditation
for
dotnet
is
a
higher
priority
than
that
so
I,
because
I
want
to
put
these
in
a
dork
and
kind
of
outline
a
road
map
having
these
as
guidance.
F
G
So
we
are,
you
know
we
are
trying
to
strike
the
right
balance
between
these
kind
of
things
that
we
internally
sort
of
considers
features
versus
reliability
and
performance
work,
and,
while
all
of
these
things
that
polo
outlined
here
make
sense
to
us
some
of
our
reliability
performance
work
is
more
important
to
us
than
some
of
these
features.
And
then
the
same
applies
to
CLR.
G
Cementation
engine
I
think
the
the
as
general
important
goals
first
rank
Microsoft
might
have
other
priorities.
I
know
Microsoft
is
very
important
thing
is,
is
a
support,
ability,
they're
happy
and
you
have
fewer
features
portability.
Some
smaller
companies
might
be
more
old
and
saying
that
let's
have
more
teachers
and
euro
stability
as
it
comes
at
the
end
of
the
day.
G
All
we
need
to
do
is
we
need
to
figure
out
a
roadmap
where
what
people
would
like
to
contribute
and
in
the
foreseeable
future
that
warrants
everybody's
investment
into
having,
because
if
we
start
having
a
drunk
base,
I
think
in
the
long
term,
every
everybody
benefits,
but
it
also
costs
for
us
to
synchronize.
You
have
joint
PRS
and
all
this
kind
of
stuff
right.
We
don't
have
to
work
on
this
road
map
versus
some
microsoft
road
map.
G
All
just
do
you
have
a
joint
understanding
that
we
all
benefit
enough
from
the
feature
that
other
people
are
contributing
so
that
we
participate
in
the
mutual
review
and
collaboration
process
so
that
the
benefit
of
set
the
cost?
And
then,
if
suddenly,
Microsoft,
says
well,
you
know.
Actually,
the
the
instrumentation
engine
is
too
important
for
us.
We
will
work
on
it
and
will
contribute
it
back
to
the
joint
codebase,
because
we
are
getting
activity
source
for
free
from
slack,
for
example,
rate
well
great.
Then
we
have
way
forward.
G
If
everybody
says
no
I
have
zero
interest
on
all
of
the
stuff.
I
need
I,
don't
have
one
I,
don't
want
to
spend
cycles
from
Kota
viewing
and
integrating
these
changes,
because
I
just
want
to
seal
our
engine
is
nothing
else
then
yeah.
Well,
then,
we
don't
have
enough.
So
I
think
that
this
would
be
a
figural.
F
Sounds
very,
very
reasonable
for
me,
and
and
also
hearing
from
the
perspective
about
the
Microsoft
and
the
instrumentation
engine
I
think
to
have
a
kind
of
make
these
things
concrete.
Are
you
I,
and
also
we
discussed
more
than
what
is
here
in
the
meeting
I'm
gonna,
take
her
out
from
the
dark
in
the
medium
prepare
a
dark
there
based
on
what
we
just
talked
it
kind
of
proposes.
F
Roadmap
and
I
will
try
to
cover
this.
These
concerns
I
thinking
they
in
the
roadmap
itself
and
all
I.
Think,
for
instance,
have
the
data
dog
as
I
start
and
then
putting
the
roadmap.
They
stage
that
we
are
gonna,
move
to
use
the
instrumentation
engine
kind
of
mix
concrete
for
as
a
plan
choo-choo
have
that
no
I
can.
G
Improvements
around
reliability,
so
we
are
going
to
reduce
the
set
of
dependencies,
especially
third
parties,
maybe
also
within
the
framework,
and
then
there
is
also
work
on
actual
features
like
bunch
of
stuff
overlapping
with
with
process
so
I
think
because
they
see
more
mostly
teachers
in
polos
roadmap.
I
think
we
can
exaggerate
this
design
discussions
joint
designs
and
then
have
different
companies
focus
on
implementing
actual
implementation
of
different
features,
because,
like
things
like
activity,
source
I
think
we
would
like,
even
if
there
is
ZERO
collaboration.
F
So
I
would
like
to
discussing
in
these
next
step.
So
then
one
it's
clear
is
to
produce
the
document
and
put
the
road.
The
road
map
and
I
think
we
can
take
the
discussion
for
the
parties
interested
on
that
we
will
keep
up
to
date
here
they
see
in
the
meeting,
but
we
taking
the
discussion
further
on
that
document.
It.
G
G
Personally,
I
don't
really
care
too
much
about
general
availability
tab,
because
the
end
of
the
day
we
already
have
different
production
for
us
general
available.
You
guys
the
same
so
whether
we
call
we
will
have
certainly
a
joint
three
for
under
the
auto
television
umbrella,
but
whether
this
repo
is
a
has
a
mark
of
general
availability,
whether
this
is
just
a
long
term
bitter
and
we
keep
integrating
from
there
into
our
own
copies
and
release
through
that.
To
me,
this
is
a
secondary
decision.
G
F
G
I
expect
will
happen
with
companies.
Is
that,
in
terms
of
hello
non-technical
management,
they
will
want
a
good
track
record
before
they
are
happy
to
make
the
steps.
Even
if
we
are
on
the
same
page
of
contributions
going
directly
into
the
joint
think,
I
am
pretty
sure.
There's
no
company
is
the
high-level
manager
will
say
great.
Let's
keep
our
own
copy
for
now
and
after
we
have
an
established
track
record
of
this
being
successful,
then
we
can
actually
start
thinking
about
killing
it
up.
G
They
will
ever
essentially
want
to
hedge
their
bets,
so
I
agree
with
you,
but
I
don't
see
that
removing
a
the
own
copies
of
each
company
like
let's,
let's
sink
it
in,
let's
be
successful
with
the
joint
copy.
Why
we
have
the
backup
of
our
own
stuff
so
and
because
of
that,
essentially
to
me
whether
we
call
it
GA
or
not,
is
sort
of
irrelevant
as
long
as
it's
getting
better
every
sprint,
yeah.
F
D
G
So,
but
that
doesn't
make
my
question
go
away.
We
can
have
a
GA
less.
That
is.
This
is
what
we
need
to
do
for
GA
and
that's
that
makes
sense
and
I
like
like
it
should
at
least
well.
Maybe
we
tweaked
in
some
detail
whatnot,
but
I
think
that
the
roadmap
will
not
only
contain
let
things
that
are
required
for
GA,
because
what
data
doc
would
say
is
we
have
things
that
a
high-level
priority
that
will
actually
benefit
everybody
in
the
community,
but
might
not
be
on
the
GA
list
like
we're,
not
we're
like.
G
Essentially,
we
are
not
happy
to
stop
working
on
our
performance
improvements
in
order
to
achieve
this
GA
as
soon
as
possible,
but
everybody
was
still
benefit
from
the
performance
from
juju
Damien.
So
because
of
that,
the
roadmap
will
end
up
not
only
contain
things
that
are
requiring
fuji
a
but
also
other
things,
and
so
because
of
that,
my
question
still
remains.
When
you
target
the
roadmap,
what
sort
of
timeline
should
we
cannot
go,
because
we
can't
go
into
great
detail
about
everything?
What
is
your
planning
horizon?
I.
F
I
I
think
for
for
the
timeline
we
need
to
anyway.
We
need
to
start
soon
to
make
this
concrete
to
see
if
we
are
gonna
get
the
collaboration
that
we
want,
that
we
desire,
let's
say,
but
I
I
would
like
to
kind
of
have,
because
now
I
think
we
prioritize
in
the
last
two
months
kind
of
trying
to
get
a
view
of
active
source
and
kind
of
reach.
F
I'd
say
that
I
would
like
to
start
on
these
as
soon
as
possible.
So
I'm
gonna
finishing
right
this
dark.
We
should
not
be
a
huge
dork,
I
I,
can't
it's
not
gonna
be
engineering.
Dark
is
gonna,
be
kind
of
a
roadmap
with
high-level
proposals,
and
if
we
have
the
agreement,
then
I
would
love
to
kind
of
have
data
dark
data
dog
doing
the
first
commit
to
under
the
open
telemetry
umbrella.
For
this,
and
then
kind
of
in
light
of
the
roadmap
is
start
to
do
the
change
proposed
it
as
the
roadmap.
You
know,
okay,.
G
So
the
when
we
talk
about
the
roadmap
when
I
see
the
timeline
I
mean
how
far
ahead
you
want
to
plan
this.
Not
when
you
want
to
do
this
this
we
can
do.
Essentially
we
can.
We
can
target
you,
have
this
document
ready
by
next
week
and
we
can
even
meet
at
the
end
of
this
week
to
get
a
quick
sink
in
a
small
circle,
Josh
vo
the
same
page,
so
that
that's
fine,
but
how
far
ahead
you
want
to
plan.
G
G
F
G
F
G
G
D
Yeah
I'm,
just
trying
to
figure
out
from
the
Microsoft
side,
I
mean
I'm,
certainly
happy
to
well
to
a
limited
degree,
to
sort
of
continue
to
try
and
be
you
know,
consultant
or
just
offer.
You
know
technical
advice.
That's
hopefully
helpful
in
terms
of
you
know
who's
building
this
thing
from
Microsoft.
D
Their
team
seems
like
the
most
impacted
by
it,
but
it's
also
not
clear
that
they
have
that
they
have
the
manpower
to
to
do
anything
about
it.
In
the
short
term,
I
can
I'm
actually
going
to
be
talking
with
them
just
later
today,
so
I'll
at
least
mention
this
and
give
them
the
opportunity
either
to
sort
of
get
more
engaged
right
now,
if
they,
if
they
want
to
or
or
hopefully
at
least,
to
kind
of
relay
a
message
that
says
you
know
maybe
they're
interested
in
the
future,
but
but
right
now
that
they.
G
G
D
G
This
is
something
that
Microsoft
wants
it's
a
very
short
term,
then
Microsoft
will
need
to
do
it
ourselves.
So
if
this
is
a
if
Microsoft
is
like,
we
cannot
even
start
dogfooding
it
for
some
test
customers
without
instrumentation
adjoined,
and
this
is
a
contribution
that
Microsoft
ninja
to
themselves.
However,
in
terms
of
how
I
see
this
evolving
like
the
core
of
the
trace,
there
was
one
the
TAS
action
instrumentation
the
way,
I
see
this
evolving
in
the
long
term.
I
do
not
see
that
we
will
avoid
switching
to
instrumenting
called
targets
in
the
long
term.
G
D
D
That's
that's
totally
fine,
I
I
mean
I
I,
yeah,
I
I
would
imagine
if
I
would
imagine
if
Microsoft
isn't
quickly
contributing
it
probably
also
means
Microsoft,
isn't
quickly
going
to
try
and
use.
What's
getting
reduced,
that'd
be
my
guess,
but
again,
I
I
feel
like
I'm
I
should
I
should
defer
somewhat
to.
A
D
To
let
mikail
or
I
I
guess
the
other
thing
that
I'm
also
unclear
about
within
Microsoft
is
that
I
I've
heard
some
things
that
there
was
some
reorganization
going
on
within
the
you
know,
within
the
azure
Monitor
team,
and
that
but
I
don't
know
what
reorganization
it
is
and
I
don't
necessarily
know
what
impact
it
would
have
in
term.
You
know
some
things
might
rise
and
some
things
might
lower
in
terms
of
priority
and
also
in
terms
of
like
free.
D
You
know
free,
not
people,
to
work
on
specific
tasks,
so
I'm
still
crossing
my
fingers
that
I'll
learn
more
soon,
but
so
far
I
don't
really
know
anything
more
anything
beyond
what
I
knew
last
week
anyway.
Yes,
I
will
bring
it
up
when
I
talk
to
Macallan
Alex
later
today,
and
hopefully
at
least
get
a
little
bit
of
response
back
from
them,
how
they
want
to
engage
or
not
engage
in
the
in
the
immediate
future.
C
C
There
are
like
few
questions:
it's
not
easy
to
like
tag
people,
so
nobody
got
notified
when
there
was
specific
question.
So
let
me
take
that
as
an
action
item
and
say
because
I
know
that
no
ways
here,
we
can
probably
ask
some
of
the
questions,
but
like
Sergei,
do
you
think
we
should
like
move
this
into
an
MD
file
and
put
this
in
the
github
repo?
Because
now
that
the
commenting
period
is
like
somewhat
over,
it
would
make
sense
to
do
an
MD
file
and
how
old,
like
properly
formatted?
C
E
I
think
it
may
be
easier
to
read
and
comment
as
in
the
document,
as
a
Google
Doc,
at
least
from
my
experience
for
such
a
big
document.
It
may
be
easier
and
it's
worse
to
present
a
specification
meeting.
Some
unfortunate
I
cannot
make
any
other
certification
meetings
because
it's
because
of
Cobie
situation
but
I
think
many
people.
There
will
be
interested
to
learn
how
we
adapted
dotnet
api's
to
open
to
limit
the
specification
and
give
a
feedback,
maybe
from
past
experience,
maybe
from
Java
experience.
So
I
think
that
may
may
be
very
valuable.
C
C
C
C
Okay,
make
sense,
yeah,
I,
think
the
one
of
the
main
reason
why
people
did
not
follow
up
is
it
was
difficult
to
tag
so
there
was
nobody
got
the
notification
that
there
was
pending
questions
on
them,
so
yeah
we
just
mentioned
like
yeah.
Ok,
so
let
me
do
this.
Let
me
follow
up
on
find
answers
and
try
to
get
answers
updated
here.
G
A
C
It
like
couple
of
weeks
back
so
it
should
be
in
the
previous
meeting
notes
as
well,
but
like
now,
they're
like
I,
have
no
I
here
like
we
can
probably
ask
about
like
couple
of
questions
right
away,
because
these
are
things
which
we
want
answers
from
so
yeah
before
that
just
AQ
cup
date.
So
there
were
like
two
designs
which
I
referred
here.
C
So
there
was
some
discussions
today
and
like
pretty
much
everything
discussed
here
or
proposed
here
was
approved
except
one
thing,
which
was
making
activity,
dot
kind,
settable,
vs
still
like
meeting
tomorrow
to
figure
out
what
would
be
the
alternate
option.
It's
not
really
a
big
deal,
consider
for
the
new
API,
because
anyone
writing
new
instrumentation
or
emitting
new
activities.
They
have
the
activity
kind
to
the
right
one
when
they
start
the
activity,
it's
just
for
the
existing
instrumentation
using
old
activity,
which
they
would
how
the
activity
kind
say
to
internal.
C
So
we
need
a
way
to
like
change
it
afterwards.
So
that's
an
open
open
hue,
but
everything
else
is
covered
so
I,
don't
know
whether
the
dock
itself
is
reflective
of
today's
discussion,
but
as
soon
as
I
have
that
here,
I
think
this
dough
could
be
complete
and
I'd
be
ready
to
move
it
defy
that
depart.
Let
me
see
if
Noah
can
help
with
a
couple
of
questions
here.
So,
okay.
C
Okay,
this
one
yeah,
so
this
part
is
done
so
this
is
a
question
was
from
follow
a
boat
like
how
do
we
do
like
out-of-band
release
like
if
they'll
look
at
new
things,
much
faster
than
dotnet
dotnet
releases,
typically
once
a
year
so
questions?
Are
we
open
to
releasing
like
out
of
five
on
each
typical
time
frame
to
address
some
new
features
so.
D
I
think
it's
it's.
It's
certainly
technically
possible,
like
we
have
the
capability
to
do
it,
whether
or
not
we
actually
did
do
it
on
any
particular
occasion
would
probably
get
evaluated
case-by-case.
It
would
probably
be
you
know:
what's
what's
the
new
work
being
added?
What
timeline
are
we
trying
to
do
it?
What's
the
impact
of
getting
it
out
sooner?
What's
the
impact
of
waiting
but
I
I
mean
I?
Think
it's
def,
it's
definitely
a
conversation.
D
So
if
it's
just
sort
of
like
hey
here's,
the
next
set
of
feature
additions,
can
we
get
these
out?
I
could
I
could
pretty
easily
imagine
the
pushback
of
sort
of
like
yeah.
We
will
but
like
how
like
how
much
is
it
really
gonna
matter
if
it's
now
versus
six
months
from
now
that
aligns
was
sort
of
the
official
shift
date,
but
I
I
I,
don't
know
for
sure
I
mean
I,
I
wouldn't
be
the
one
I.
Would
probably
have
some
influence
in
that
decision,
but
I
wouldn't
be
the
one
make
the
call
officially.
D
I
mean
def.
Definitely
the
like
all
the
pre-existing
sort
of
servicing
criteria
for
the.net
still
applies,
and
there
is
there
are
servicing
releases
that
I
think
I
think
it's
the
fastest
that
they
come
is
about
once
a
month
like
once
a
month,
maybe
once
every
other
month.
Something
like
that.
So
any
anything
that
qualifies
just
as
a
as
a
servicing
fix,
certainly
so
I
think
where
it
gets
and-
and
we
have
in
the
past.
D
We've
basically
talked
about
diagnostic
features
in
particular
that
there
is
some
leeway
to
sort
of
put
new
diagnostic
features
into
the
servicing
pipeline,
even
though
they're
not
specifically
a
fix,
but
there
would
need
to
be
relatively
good
justification
for
you
know.
What's
the
urgency
like
why
can't
it
wait
until
the
next
normal
planned
major
release,
yeah.
F
I
think
they
answer,
at
least
for
me-
is
very
satisfactory.
I
just
wanted
to
confirm
that
is
the
Chipko
release
for
dotnet
libraries.
From
from
the
dotnet
theme,
you
know,
III
think
for
a
lot
of
people
will
be
good
to
have
explicitly
there.
You
know
I,
it
simplest,
but
I
think
it's
good
to
have
explicit.
C
F
C
D
I
mean
as
long
exactly
as
long
as
it
meets
the
sort
of
the
criteria
for
doing
a
sir.
You
know
like
we
don't
fix
every
bug
in
servicing.
We
we
only
tend
to
fix
the
ones
that
we
think
are.
Are
you
know
reasonably
impactful,
but
as
long
as
we
can
make
the
argument
that
a
given
bug
or
you
know,
is
reasonably
impactful
that,
yes,
we
can
bring
it
through
servicing.
Just
like
anything
else.
Okay,.
D
C
Sample
code,
which
I'm
hydrating
is
from
pure
opened
elementary
concept.
You
use
a
provider
with
some
configuration
and
from
the
provider
you
get
the
tracer
line.
Using
this
tracer,
you
start
stop
image
spends.
The
key
thing
to
note
here
is
to
get
the
tracer.
You
need
access
to
this
provider
and
come
back.
I
mean
on
contrast
to
that
in.net.
The
equivalent
of
tracer
is
activity
source.
The
thing
which
is
the
entity
which
can
produce
spans
right
here.
There
is
no
requirement
just
new
topic.
C
You
can
just
create
it
out
of
thin
air,
like
just
new
Liam
and
the
way
like
we
specify
in
the
current
implementation,
how
you
are
interested
in
a
particular
activity
source
is
by
using
the
name,
slash
library
of
activity,
so
you
tell
which
libraries
you
are
interested
in
and
we
just
get
notification
from
those
libraries,
so
I
think
seg
is
asking
for
the
reason
like
why
it
is
like
this.
They
are
supposed
to
the
way
in
which
Obama
tracing.
E
D
I
mean
off-the-cuff,
the
main
thing
would
just
be
that
none
of
the
all
of
the
state
or
configuration
that
goes
into
a
tracer
provider,
none
of
it
seems
obviously
relevant
for
activity
or
we
haven't
yet
made
it
relevant
like
like
the
fact
that
we
can
write
new
activity
source
and
we're
not
missing
some
other
parameter
that
you
need
to
provide
for
the
framework
to
do
what
it
needs
to
do
effectively
from
the
frameworks
perspective.
There's
no
purpose
to
having
the
other
type
I
mean
we
could.
You
know
like
if
we
created
it,
it
would
be.
D
You
know
new
tracer
provider,
empty
object,
it
has
no
state,
and
then
you
call
create
activity
source
on
it
and
you
get
an
activity
source
but
like
from
a
framework
perspective.
There'd
be
no
value
and
using
it
as
it
didn't
do
anything.
It
was
just
an
extra
object
you
had
to
create
before
you
could
create
your
activity,
source.
E
G
D
G
D
G
D
G
D
A
D
That
had
its
own
set
of
issues
either
you
could
do
Singleton's,
but
then
you
either
had
to
decide.
If
you
want
the
the
more
useable
API,
you
pay
a
distinct
performance
cost
for
it,
like
you
have
to
start
having.
You
can't
have
like
a
global
single
to
now.
You
need
like
thread,
static,
Singleton's
or
even
async,
local
Singleton's,
and
if
you
do
try
to
and
if
you
try
and
get
rid
of
the
performance
cost
by
using
a
global
singleton,
then
you
run
into
a
bunch
of
usability
issue.
G
G
D
E
Think
one
thing
that
we
may
not
hear
is
one
way
to
avoid
this
allocation
is
to
return
the
same
object
over
and
over
again.
The
kind
of
no
no
operation
object
like
empty
object
and
the
reason
why
we
didn't
do
that-
and
I
think
it
worth
mentioning
in
the
document
by
the
way
is
that
we
want
to
preserve
the
behavior
when
agent
attributes
actually
stores
the
attributes,
no
matter
what
and
then
you
can
extract
this
attribute
back.
E
E
If
you
feel
really
strongly
about
it,
it's
worth
bringing
up
and
say
like
with
areas
like
reevaluate,
whether
we
actually
need
this
pager
on
or
attributes
behavior,
and
maybe
we
can
like
maybe
eliminating
question
marks-
is
more
important,
epigenomic
attributes
and
activities
that
are
not
actually
active.
I
see.
G
C
D
Was
just
gonna
say:
I
mean
it
did
come
up
in
discussion
before
there
were
yeah
that
there
were
definitely
I
mean
I've
had
similar
feelings,
that
it
is
an
unusual
pattern
and
I
think
ultimately,
where
we
sort
of
concluded
was
that
the
odd
sort
of
usability
issues
and
the
that
this
that
the
singleton
or
the
no
op
pattern
sort
of
off.
You
skated
what
we
thought.
D
G
G
G
G
C
E
There
is
a
full
activity.
Mention
is
recorded
that
exists,
an
activity
but
socially
for
this
part,
I
can
see
any
mention
of
the
comparison
of
to
api's,
wondering
like
it
exists
was
like.
While
we
didn't
put
a
map
in
talk
to
spend
api
to
this
flag,
so
I
would
assume
that
spam
context
that
sampled
or
like
sampling
flag
will
be
a
mapping
for
is
recorded
back
and
I
can
see.
This
may
think
anywhere.
So
I
guess
it's
missing.
I
see.
C
E
C
Mean
that
general
feedback
round
like
restructuring
this,
to
really
talk
about,
like
like
the
reefs,
if
using
the
API
and
SDK
for
applications
or
SDKs
for
applications
who
he
choose
to
use
open
telemetry.
So
that's
something
which
I
should
be
able
to
like
structure
better
think
follow
and
someone
else
also
shared
this
same
feedback,
because
some
of
the
examples
are
mixing
the
concepts
of
AP
and
SDK.
So
why
I
try
to
exploit
them
so
that
it's
obvious
that
for.
C
E
Yeah
I
think
you
on
the
simple
library
instrumentation.
It
may
be
interesting
to
know
whether
what
standard
process
we
can
implement
like
if
you
implement
sampler,
do
you
have
access
to
all
the
same
fields?
If
you
have,
if
you
want
to
implement
your
like
I,
don't
know
expert
or
do
you
have
access
to
all
the
same,
so
this
also
may
be
important
point
to
like.
It
may
be
another
point
of
view
of
a
person
reading
this
document,
so
you
shouldn't
just
contact
on
a
single
part
of
it
on
library,
instrumentation.
G
Five
preview
API,
not
not
at
the
repose
that
could
be
missing
something
and
I
was
thinking.
So
if
there's
no
text
at
all
strings,
so
if,
if
my
back-end,
once
you
have
certain
types
that
it
cheats
preferentially
so
I
am
just
spent
information
on
my
back-end,
but
then,
if
a
tag
has
some
kind
of
type
boolean
number
whatever
something
that
is
not
string,
I
want
to
actually
like
do
something
special.
There
can
I
have
like
what
what
is
the
right
way
to
do
type
like
strongly
typed.
You.
C
Don't
have
it
today,
but
it's
it's
like
the
current
will
be
its
do
not
exist
today,
but
the
upcoming
versions
of
the
preview
would
contain
something
which
will
allow
you
to
do
more
than
string
and
string,
because,
right
now
it's
just
string
and
string
that
is
already
captured
in
the
design
discussion
so
like
in
that,
whenever
it's
a
next
release,
I
think
it's
preview
age.
It
should
be
like
string
and
object,
so
it
should
be
able
to
matching
the
awkward
elementary
speaker,
cities
open
telemetry.
C
C
It's
it's
any.
It
will
be
the
extreme
object,
which
means
we
should
be
able
to
support
anything
because
right
now,
one
of
one
of
the
open
peers
in
open
telemetry
is
about
like
having
that
like
the
tight
can
be
like.
Another
attribute
of
itself
can
be
checked
chained,
but
those
things
were
not
yet
much.
So
we
want
to
make
sure
like
we
leave
it
like
string
object.
So,
depending
on
how
our
respect
would
line,
we
should
still
be
able
to
support
it.
I
think
the
merge.
E
B
Hi
Sergei
I
know
I,
don't
mind
right
now
because
see
you
pointed
out,
I
still
didn't
add
new
relevant
code
reviews
for
the
current
repository.
So
for
now
I,
don't
think
that's
relevant,
but
maybe
later
I
guess
I
think
see
Joe
added
a
time
frame
like
one
one
month
or
lake
or
more
to
add
me
or
check
me
and
if
to
see,
if
I'm
able
or
not
to
to
be
a
code
reviewer
so
for
now,
I
think:
that's
okay,
I'm
I'm,
doing
almost
code
changes
and
trying
to
start
with
doing
things.
But
it's
fine
yeah.
C
So
I
shared
a
similar
opinion
because
I
want
to
like
see
if
we
can
do
some
more
code
reviews,
because
we
desperately
need
approvers,
that's
given
me,
we
always
are
like
7
to
10
psi
T
of
all
the
time,
so
we
definitely
need,
but
I
just
want
to
make
sure
we
follow.
The
leg
stated
guideline
that
there
should
be
like
significant
number
of
active
participation
in
court
reviews
airport
from
contribute.
Oh.