►
From YouTube: 2023-02-07 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
A
A
Three
minutes
password:
we
can
get
started
so
I
just
want
to
see
whether
we
did
the
update
about
the
release.
Okay,
we
did
the
release
password,
so
the
last
supposedly
last
release
Candace
is
already
out
and
have
like
one
issue
which
you
need
to
discuss,
because
this
might
affect
the
public
API
before
only
so
Banks.
Like
please
start
with
this
one.
B
It's
a
pretty
Heavy
Hitters,
David,
Fowler
finished
everyone
we'll
have
to
look
to
add
a.
Why
do
we
have
to
call
Starbucks
every
time
it
should
be
done?
Automated
so
I
think
we're
all
familiar
with
the
reasons
it
has
to
be
able
to
hosting
package,
but
there's
enough
noise
where
wanted
to
find
a
solution,
so
they
paid
out
and
I
think
the
variety
and
said,
but
there's
no
way
to
do
this
without
a
dependency.
Unless
you
do
reflection
or
it's
a
kind
of
dynamic
curve,
people.
B
That
so
that's
what
this
PR
is
doing,
so
it
puts
back
the
reflection,
dot
emit
package
in
SDK,
which
I'm
just
going
to
be
with
like
a
few
days
ago.
This
is
the
full
version
we
used
to
have
the
lightweight
version.
The
difference
is
the
light.
Weights
only
contains
Dynamic
method.
The
full
one
gives
you
methods
for
building
types
and
assemblies
like
full
on
artifacts,
which
we
need
here,
because
we
don't
just
need
a
method.
We
need
to
implement
an
interface,
so
that's
incredibly
difficult
to
do.
B
It
took
me
a
while
to
figure
out
how
to
get
through
this,
but
so
what
this
code
does
is
it
basically
reflectively
creates
this
class
that
I'm
showing
here,
which
is
just
the
implementation
of
I
posted
service,
so
that
we
can
drop
it
into
the
service
selection
so
that
the
host
starts.
This
code
executes
this
code.
Basically,
it's
exactly.
B
C
B
So
in
order
for
this
code
to
fire,
you
need
to
have
Dynamic
code
supported,
so
there's
some
runtimes,
some
execution
environments,
where
you
can't
do
this
think
like
mono
examine
but
there's
also
like
Windows
environments,
like
some
government
environments,
they
don't
allow
Dynamic
code
because
it's
security
risk
like
if
you
can
inject
code
into
a
process,
you
own
that
process,
so
there's
some
very
locked
down
environment,
so
I
added
some
checks,
so
you
have
to
support
Dynamic
code
and
then
you
have
to
find
the
service
or
line
77.
B
B
B
Up
a
little
bit
so
basically
line
221
and
227
is
really
all
the
hosted
Services
doing.
All
it's
doing
is
accessing
the
Tracer
provider
and
the
meter
provider
so
that
they
start
up.
So,
if
you're
in
an
application-
and
you
don't
have
Dynamic
code
or
you're
not
using
hosting
you,
either
have
to
do
these
operations
yourself
or
you
have
to
do
the
SDK
dot
create
style
where
you
just
call
build
and
start
it
up
manually,
which
is
sort
of
sort
of
the
existing
Behavior.
B
A
Okay,
so
if
the
user
is
on
the
like,
they
are
using
the
extensions
dot
hosting
and
like
we
previously
like
just
asked
them
like
service,
dot,
add
open
Telemetry,
and
then
we
asked
them
to
do
start
with
post.
Now
we
do
automatically
under
like
most
cases,
but
in
the
scenario
where
we
are
not
able
to
do
it.
We
emit
a
log
and
the
user
has
to
do
like
they
have
to
retrieve
the
tricep
provider
like
by
hand
yeah.
A
A
Reason
like
because
then
we
added
the
dynamic
code
earlier
for
activity
enumeration.
We
had
like
some
issues
with
the
mono
and
other
end,
so
is
that
the
reason
why
we
are
doing
that
pretty
much
yeah,
because
previously
our
Behavior
was
I,
think
we
were
crushing
ready
like
we
are
not
having
this
check
so
yeah.
We
didn't
check
those
before.
Let's
certainly
hurts
so
right
now
like
if
they
are
in
like,
for
example,
like
mono,
it
would
still
compel,
and
all
all
they
have
to
do-
is
there
to
do
this
extra
things
like
yep.
A
A
B
B
B
A
So,
can
you
remind
me
once
again
the
whole
reason
why
we
are
adding
this
to
still
trigger
it
just
to
trigger
the
like
building
of
the
entire
provider
stuff
right?
So
that's
the
reason
why
we
have
to
have
this
hosted
service
that
there
is
no
other
reason
right
and
nothing
is
changing
like
here
like
it's
just
that
it's
about
asking
users
to
do
that.
We
try
to
do
it
ourselves.
So
that's
basically
the
thing
we
still
have
that
hosted
Service,
but
not
in
the
extensions,
not
hosting
course
right.
B
So
really,
if
we
do
this,
we
don't
need
the
hosting
package.
A
I
mean
it's
not
I,
don't
think.
That's
left
right,
like
everything
is
absolute,
which
means
there
is
literally.
B
A
B
A
A
A
B
A
A
D
A
We
already
modified
this
okay,
so
this
packages
are
already
using
the
new
API.
They
are
no
longer
using
the
like
the
to
be
deprecated
once
anyway,.
D
I'll
just
say
that
I
support
this
PR
I
think
it's
kind
of
the
icing
on
the
on
the
cake
of
of
all
this
work
that
Mike's
done
to
make
the
configuration
of
the
SDK
unified
whether
you're
using
you
know
whether
you're
on
like
an
asp.net
core
kind
of
framework
or
something
else.
D
D
But
I
think
this
is
great
I
think
it
does
away
with
that
so
big
thumbs
up
for
me.
A
Okay,
so
like
Blanche,
can
you
again
remind
me
like
what
is
the
thing
that
we
are
going
to
ask
the
net
team
to
provide
like
we
are
going
to
ask
them
to
give
us
a
different
hook
without
using
a
poster
service
right
these
type
of
things?
We
suppose
you
would
plan
to
ask
so
that
we
can
get
it
off
this
thing
in
the
like
future.
B
Yes,
the
runtime
team
is:
can
they
give
us
some
API
in
dependency
injection
dot
obstructions
that
allows
us
to
plug
in
to
when
the
service
provider
is
created
so
that
we
can
warm
up
services,
so
it
doesn't
need
to
be
I
hosted
service.
It
could
just
be
like
an
event.
You
know
like
on
service
provider,
ready
I,
don't
really
care
what
the
API
is,
but
just
some
some
extension
point
where
we
can
say:
okay,
when
this
fires
we'll
do
these
two
things
and
then
we
can
drop
this
whole
mess
yeah.
A
B
B
A
Yeah
so
I'm
just
trying
to
see
like
get
a
feel
of
like
how
would
our
code
look
like
it
should
be
like
like
this
is
existing
code
like
for
in
espnet
core?
So
all
you
do
is
call
add
open
Telemetry
with
tracing
with
metrics,
and
this
is
all
our
stuff
is
to
like
pick
yeah
configure
as
source
and
with
metrics
here,
and
that's
it
yeah.
Okay.
So
that
looks
very
clean.
C
C
I
haven't
had
much
chance
to
play
with
the
stuff
since
it's
been
refactored
but
I
trust
in
you
all.
So
you
know
my
feelings
and
I've
not
exactly
been
keeping
them
to
myself.
Should
we
say
so.
Yeah
I
think
we're
in
a
relatively
good
Pace.
All
I
want
at
the
moment
is
the
release
out,
and
so
we
can
start
fixing
all
the
instrumentation
stuff
and
having
that
conversation.
A
Yep
and
deal
like
the
fact
that
we
won't
be
releasing
extensions
dot
hosting
at
all,
because
the
current
plan
is
like.
We
just
kill
it,
because
there
is
no
need
of
a
extra
package.
So
once
if
this
plan
goes
as
if
we
just
released
the
1.4
of
the
SDK,
which
means
you
don't
really
miss
package,
so
I
will
just
like
Mark
this
package
as
absolute
forever.
Not
not.
C
Good,
how
do
you
put
it
in
the
service
collection
now
because
we're
not
depending
on
service
collection?
That's
all
independent
in
abstractions.
Now
is
that
could
you
repeat
so
when
we
ask
somebody
to
add
it
to
an
asp.net
core
site
where
they
would
have
previously
added
it
to
the
service
collection?
Oh,
they
would
still
do
it.
The
certain
collection
is
part
of
the
abstractions
now
yeah,
so.
A
To
have
a
dependency
on
that
extension
start
hosting,
it
will
still
work.
So
if
you
are
an
asp.net
core
user,
you
would
get
the
udat
open
Telemetry
by
this
line.
Yeah.
That's
nothing!
No
need
of
extra
package.
So
once
you
install
the
SDK,
either
directly
owned
by
indirectly,
by
using
some
exporters
like
console
or
otlp,
you
don't
need
any
extra
packages,
so
the
onboarding
would
be
like
loads
motor
yeah.
C
Yeah,
that's
the
the
main
thing.
The
other
thing
is
obviously
the
package
versioning
stuff.
C
If
we're
getting
rid
of
the
hosting
package,
then
great,
because
we'll
just
have
one
release
package
that
people
can
use
so
we're
heading
in
the
right
direction.
I
think
there's
still
a
little
bit
of
work
that
we
need
to
focus
on.
Maybe
after
we've
got
the
1.4
out
around
the
instrumentation
stuff
and
re-bring
up
that
conversation
and
work
out
how
we
fix
all
of.
A
That
yeah,
so
that's
a
separate
one
I
think
we
should
definitely
have
time
to
talk
about.
This.
I
mean
talk
about
the
instrumentation
Library
version
right
after
this,
but
I
think
this
is
more
important,
because
this
would
definitely
have
impact
on
1.4,
because
proceeding
with
this,
we
should
do
the
like
real,
like
an
RC,
yeah
I
think
we
did
I
was
not
here
last
week.
So
did
we
do.
C
Thing
so
I'm
coordinating
with
Philip
Carter
around
documentation
yeah.
C
A
Yeah
I
did
reply
to
Carter
like
I
suggested
I
mean
I
replied
before
I
saw
this
proposal,
so
I
basically
asked
we
can
take
it
back
down
rc2
and
modify
the
examples.
But
if
you
are
going
to
remove
this
one,
then
let's
wait
for
this
to
happen,
because
that
way
like
it's,
it's
loadless,
so.
C
They,
the
the
problem
that
Philip's
got,
is
obviously
he's
getting
it
in
the
neck
and
I'm
getting
it
in
the
neck
from
customers
about
the
docks
and
that
little
section
that
you're
pointing
at
the
moment
does
not
work.
I
mean
we're
saying
we're
two
weeks
away
from
having
a
fully
fledged
released
version
of
1.4.
Then
it's
not
worth
it
if
we're
talking
a
month
ago,
a
month
away,
because
obviously
he's
been
trying
to
get
this
updated
since
November
and
it's
not
worked
since
November
and
that's
the
problem.
C
A
So
let
me
confirm
that
thing,
so
we
just
released
the
like
police
candidate
must
be
so.
The
ideal
plan
was
like
we
just
wait
for
two
weeks:
if
no
feedback
we'll
just
release
the
stable,
but
now
we
have
a
good
feedback
on
the
start
with
host.
So
this
means
we
have
a.
We
need
to
do
another
police
candidate.
If
this
PR
is
done
like
today
like
we
could
do
it
like
as
soon
as
tomorrow,
the
last
release
candidate.
A
A
But
I
think
one
one
like
one
alternate
option
is
we
can
go
ahead
and
update
the
dock
right
away,
but
like
with
the
rc2,
but
then
it
will
be
like
another.
It
will
be
update
to
remove
extensions
plot
hosting
and
start
with
host
call
again
within
the
industry
comes
out.
So
it's
going
to
be
like
a
little
bit
more
effort.
C
No,
so
what
we
were,
what
we
were
looking
at
doing
is:
we
will
literally
remove
the
pre-release
stuff
and
add
an
individual
version
number
in
there
and
pin
everything
to
an
individual
version.
Oh
yeah,
that
that
would
all
work
right
right
now:
yeah
yeah,
oh!
No!
We
can
make
it
work,
that's
what
I'm
saying
we
can
make
it
work.
We
just
don't
want
to
do
the
work
when
we're
so
close
to
getting
it
then
having
to
redo
the
work
we
can
make
it
work.
C
We've
already
got
our
own
documentation
inside
honeycomb
that
does
all
of
the
stuff
and
tells
them
specific
versions.
So
we
can
map
all
of
that
over.
That's
not
a
problem,
but
we
don't
want
to
be
doing
that
and
doing
all
the
work
and
getting
the
piaz
up
for
all
the
docs.
If
we're
two
weeks
away,
but
we've
been
perpetually
two
weeks
away
since
November,
which
I
can
get.
Why
he's
a
little
bit
annoyed
with
it,
but
I'm
keeping
them
at
Bay
as
much
as
I
can
yeah.
A
I
mean
I
totally
get
it
like.
We
originally
wanted
to
release
it
every
time,
I,
don't
think
we
ever
released
very
first
version.
Also.
We
were
planning
to
go
into
number,
but
it
slipped
to
January.
The
metrics
work
also
originally
planned
in
November
what
happened
only
in
January
when
this
one
originally
planning
downward,
but
it's
not
two
like
February
yeah,.
A
Next
time
we
will
Target
like
October,
so
we
can
actually
do
it
in
number
itself,
but
anyway,
so
like
any
like
plans
or
a
language,
can
you
comment
down
like
possible
timelines
like
if
we
get
like
a
couple
of
folks
to
extra
I
mean
outside
of
immediate?
The
maintain
is
like.
Maybe
if
you
can
like
look
at
this
and
lines,
if
you
can
get
David
folder
or
something
from
19
to
see
this
looks
okay,
then
we
can
proceed
with
this
and
yeah.
A
Please
do
write
that
in
the
description
saying
that
we
are
trying
to
get
rid
of
the
extensions
completely
I.
A
Think
blinds
can
directly
thinking
because
he
is
also
not
a
fan
of
this
one,
so
he
he
has
some
interest
in
between
like
this
PR
go
through,
so
if
he
can
get
an
explicit
act,
something
that
he's
improving
this.
That
would
be
great
because
our
Avenger
goal
is
not
to
have
any
hands
right.
We
just
want.
C
It,
but
what
I
will
do
is
I'll
tweet
out
that
say:
we've
removed
start
with
host
and
it's
now
start
with
David
and
see
what
he
says
which
might
prompt
him
is
coming
up.
Look.
D
A
Anyone
else
and
like
I
I
would
say
this
is
the
only
thing
this
year.
Sorry,
this
is
the
only
thing
which
we
are
waiting
for
like
1.4.
There
was
another
issue
about
resource
detector,
but
I
think
Blanche
and
Ellen
decided.
We
do
not
need
it
because
they're
like
some
migrainians,
you
need
to
handle
this,
but
that's
not
considered
a
broker.
So
as
of
today,
this
is
the
last
thing
which
is
standing
in
our
way
from
releasing
the
1.4
yeah
thanks
to
Overcoat
this
world.
A
Okay,
I
have
a
small
topic.
You
can
do
that
before
we
talk
about
instrumentation.
That
would
be
great
yeah,
so
I
mean,
as
you
might
have
some
of
you
might
have
noticed.
We
have
a
maintenance
issuing
control.
A
A
It
cannot
like
people
if
they
are
not
one
of
the
open,
Elementary
organization,
so
because
of
that
I
don't
think
like
we
get
any
active
engagement
from
the
people
who
contributed
so
that
basically
puts
a
burden
background.
The
media
needs
to
maintain
all
the
components.
So
one
of
the
proposal
in
this
guide
is
the
component
owner
should
be
an
open,
Elementary,
Community
member
which
has
its
all
the
government.
So
you
can
just
become
a
number
like
simply
by
saying
I,
don't
know
you
have
to
follow
the
process
documented
here.
A
So
that's
one
proposal,
so
the
idea
is
like
so
that,
like
if
someone
is
contributing
a
component,
we
kind
of
get
a
commitment
from
them
that
they
are
willing
to
own
it.
If
they
are
not
willing
like
they
should
find
someone
else
who
can
own
it.
That
would
avoid
the
issue
of
like
too
many
unattended
issues
and
protectors
being
opened.
A
So
that's
the
really
the
first
one
second
is
about
targeting
document
versions,
which
we
do
not
really
support
in
the
main
tattoo.
So
so
far
we
did
not
put
any
restrictions,
so
there
are
some
components
which
are
still
on
the
very
old
version,
and
we
want
to
like
make
sure
that
we
are
like
consistent.
A
We
already
support
whatever
is
officially
supported
by
that.net
team
itself,
so
two
real
proposals
and
everything
else
is
just
like
some
formatting
I
I
just
want
to
like
get
some
opinions
about
like
what
would
everyone
think
and
my
my
main
goal
is
to
make
sure
that
we
don't
have
like
a
lot
of
unattended
issues
or
unresponded
issues
and
responded
PRS
from
components
which
are
not
owned
by
the
maintenance
them
select.
Some
of
them
are
owned
by
thing
like
it
did
look
at
the
component
owners.
A
It
looks
like
some
of
them
have
owners
who
are
also
the
Brewers
or
maintenance.
So
we
get
like
pretty
active
response
to
that,
but
others
they
don't
even
get
notified,
so
I
have
to
like
go
and
individually
pick
them.
That's
something
which
I
want
to
avoid
doing
so.
I
will
leave
the
peer
like
give
it
like
few
days
for
everyone
to
think
through
and
come
in.
A
Sorry
Blanche
I
could
not
hear
you.
So
can
you
hear
me
now
I
can't
hear
you
but
a
little
bit
stable.
A
D
C
So
yeah
I
was
what
I
was
going
to
well.
I
was
wondering
whether
the
versioning
conversation,
whether
we
could
have
something
that's
more
amenable
to
a
person
that
lives
on
GMT
time
zone,
to
have
that
conversation,
because
I
think
it's
a
an
in-depth
conversation
that
I
don't
want
to
do
at
midnight.
C
B
A
So
since
I
am
on
phone
Alan,
maybe
you
can
share
the
screen.
D
A
Give
me
a
second
yeah
and
blind
today,
please
go
ahead.
I
think
you
were
saying
something
when
we
got
cut
off.
D
B
Okay,
I
had
to
open
my
laptop
so
on
the
proposal.
The
the
first
part
seems
totally
reasonable
about
the
membership.
The
second
part
about
the
supported
framework
versions.
So,
while
you
were
on
vacation,
see
Joe
I
went
and
re-enabled
the
CI
for
netcore
app
3-1,
and
you
karsh
and
I
talked
a
lot
about
this
issue,
our
guidelines
or
lack
thereof.
B
So
my
feelings:
are
we
shouldn't
ship
any
code
from
contrib
that
hasn't
been
tested,
so
I
would
say
the
policy
is,
you
may
only
ship
run
times
if
you
can
make
sure
there's
CI,
so
that
sort
of
gives
us
a
natural
like
once
we
can
no
longer
run
CI,
then
we
just
go
and
remove
those
targets.
We
don't
allow
new
code
to
ship,
I,
wouldn't
say
we're
going
to
go
in
and
just
remove
the
CI
and
allow
that
because
I'm
uncomfortable
having
my
name
as
a
maintainer.
B
A
I
closed
that
PR
and
instead
proposed
that
the
new
components
who
is
contributing
new
now
they
have
a
guidance
like
they
should
Target
only
the
officially
supported
one
so
that
we
can
Target
them
in
the
CI.
If
not,
we
don't
allow
it
because
we
cannot
test
it.
So
I
think
I
am
in
agreement
with
what
you
just
suggested.
B
B
D
B
A
Yeah,
we
already
done
this
in
the
main
repo.
Now
we
just
need
to
do
it
in
the
country.
So
I'll
do
one
thing:
I'll
split
the
pr
into
two
one
for
the
on
like
membership
part
and
for
the
Frameworks
I'll
first
update
the
guidelines
in
a
separate
PR
and
ask
comments
from
all
other
component
owners.
There
are
quite
diff
few,
so
we
don't
I
mean
I,
don't
expect
everyone
to
reply,
but
at
least
we
specifically
need
comments
from
the
AWS
owners,
because.
B
Yeah
cool
yeah,
because
I
I
could
see
them
being
very
upset.
If
you
know
we
remove
their
targets
and
then
they
come
and
say
like.
Why
are
you
doing
this
like
I
didn't
approve
this,
so
I
would
I
would
love
to
at
least
give
everyone
the
opportunity
to
sound
off,
and
maybe
maybe
someone
will
raise
valid
points
that
change
are
thinking
but.
A
Yeah,
so
I
will
create
a
separate
issue
to
let
to
let
everyone
comment
on
that
and
I
can
update
the
pr
to
remove
that
section.
Yeah
I
can
do
that
any
other
things.
This.
D
Is
the
question
about
this
statement
that
component
owners
are
expected
to
respond
to
issues
I,
don't
know
if
this
is
a
problem
today,
but
what
happens
or
do
we
need
to
articulate
at
this
point?
What
happens
if
somebody
is
not
responsive,
yeah.
A
I
do
not
know
the
answer.
I
was
hoping
we
could
learn
from
the
The
Collector
control,
because
that
seems
like
the
most
active
contribute
like
we
could
like
pretty
much
steal
from
what
they
are
doing,
but
they
do
have
the
requirement
about
like
you
should
have
a
component
owner.
I
do
not
know
whether
what
the
what
action
they
would
do.
If
someone
is
not
responding,
do
they
remove
the
component
or
do
they
like
ask
someone
else
to
take
ownership?
A
I
haven't
really
thought
through
the
whole
like
what
if
it
doesn't
happen,
in
fact,
foreign.
D
D
Sense
I
think
yeah
I'll
reach
out
to
the
Java
Community
too,
because
I
think
they
recently
did
a
reorganization
to
move
some
a
number
of
things,
so
their
contrib,
repo
and
I
think
they're
kind
of
thinking
about
this
as
well.
Yeah
yeah.
A
We
don't
want,
like
the
maintenance
to
be
like
responsible
for
all
the
compound,
because
the
components
are
like
growing
every
month,
so
we
won't
be
able
to
like,
like
physically,
like
deal
with
all
the
issues
from
all
the
components.
So
so
we
want
to
make
sure
it's
a
healthy
repo,
so
we
need
to
like
make
the
person
who
is
contributing
responsible
for
issues
in
their
component,
or
at
least
they
should
find
someone
else
whether
we
should
like
delete
the
component.
A
A
Yeah
I
think
that's
pretty
much
I
had
in
terms
of
agenda,
which
we
already
put
but
I
don't
know.
Martin.
Are
you
still
here
like?
Do
you
want
to
like
walk
us
through
like
your
idea
of
like
removing
the
pre-release
labels
and
what
benefits
it
would
bring
because
something
we
discussed
in
the
slack,
but
we
never
discussed
in
a
Sig
meeting
if
I
remember
correctly,
so
yeah.
C
I
couldn't
I
couldn't
do
a
little
five
minute
bet
on
what
we've
been
talking
about.
C
So
the
problem
statement
is
we
have
people
who
are
trying
to
install
various
different
instrumentation
libraries
on
top
of
our
current
code,
they're.
Also,
then
trying
to
use
the
pre-release
packages.
The
biggest
problem
was
the
hosting
package,
because
that
was
pre-release
and
people
were
using
that
in
production,
which
we
advise
people
to
use
it
in
production,
because
it's
stable
just
there
were
some
non
critical
issues
in
that
package,
so
we
were
using
RCS
as
not
exactly
the
right
thing.
C
So
there's
also
this
idea
in
the
community
that,
if
I
have
to
tick
the
little
tick
box
in
Visual,
Studio
that
says
pre-release
packages
or
I
have
to
add
a
pre-release
modifier
onto
my
nuget
install
commands
that
that
detracts
people
from
using
it
more
specifically,
security
teams
and
teams
in
organizations
where
they
go.
No
you're
not
using
pre-release
packages
that
hurts
adoption
but
most
importantly,
ends
up
in
a
scenario
where
we
have
binary
compatibility
issues
because
people
are
trying
to
use
one
pre-release
package
against
another
pre-release
package.
C
The
idea
is
that
the
when
we
are
happy
that
the
only
problem
that
we've
got
is
to
do
with
the
naming
of
attributes
and
how
the
the
spans
are
created,
that
we
use
the
zero
dot
notation
instead
of
using
a
pre-release
modifier,
and
we
Mark
all
of
those
things
as
0.5
onwards.
C
So
I
think
somebody
suggested
using
0.5
as
a
starting
point
rather
than
0.1,
which
don't
really
care.
But
we
start
at
a
number
that
is
less
than
one.
We
will
not
get
to
a
1.0
release,
but
we
will
use
the
full
release
rather
than
a
pre-release
to
back
that
up.
I've
been
working
with
Philip
on
a
document
of
what
is
versioning.
C
C
And
that
says
this
is
what
we're
doing,
because
at
the
moment
our
release
candidate
is
one
big
hammer
that
we
use
for
every
single
thing.
That
says
we're
in
a
cycle
of
doing
work
when
realistically,
what
we've
had
for
the
last
three
months
is
actually
a
beta,
maybe
an
alpha,
because
we
were
trying
things
out
so
I
mean
I,
suppose
more
recently.
Yes,
maybe
as
a
release
candidate,
because
we
were
fairly
confident
but
hey.
How
so
that
that's!
C
The
proposal
is
that
we
move
at
least
our
core
instrumentation
packages,
the
ones
that
are
part
of
the
non-contra
repo,
that
we
move
all
of
those
to
0.5
as
a
full
release.
C
And
then
we
can
still
use
our
pre-release
stuff
for
creating
stuff
in
the
interim,
which
will
allow
people
to
use
a
normal
standard,
nougat
install
to
install
packages
and
if
we
are
doing
some
pre-release
stuff
that
will
not
affect
them
at
all,
because
yes,
there'll
be
a
pre-release
of
the
API
package
and
the
SDK
and
our
abstractions
package
and
all
of
those
things
will
be
in
pre-release.
C
C
C
A
Should
definitely
be
able
to
remove
the
four
o'clock
one
because
that
was
added
like
originally,
when
there
were
people
from
different
time
zones
who
wanted
to
attend,
but
at
least
one
of
them
indianers
at
that
time.
Added
that
but
I
don't
think
like.
We
really
need
that
alternating
one.
If
you
can
do
with
11
o'clock
PST
every
week,
that
should
be
fine
with
most
of.
A
Yeah
so
maybe
like
this
is
something
which
we
can
you
say
do
like
something
in
the
slack
channel
to
figure
out.
If
we
can
find
an
alternate
time
right
and
it
should
be
fine
with
any
other
time.
I
can
like
really
rearrange
things
so.
C
Mostly
just
I'm
looking
at
a
one
off,
you
know
just
for
this
particular
conversation.
The
4
pm,
one
is
fine.
I
can
make
that
happen
just
because
this
is
a
really
in-depth
conversation
and
I
want
to
dedicate
the
time
it
needs.
A
We
can
do
it
at
work,
one
then
like,
then
we
don't
need
to
change
the
Sig
meetings.
We
can
probably.
C
Do
it
absolutely
I
mean
I
want
to
talk
about
the
Sig
meetings
at
some
point,
because
you
know
I
like
being
involved
and
I
really
would
like
to
find
a
better
time
that
works
a
bit
better
for
me,
but
I'm,
one
person
out
of
enough
people
that
it
shouldn't
matter,
but
for
this
particular
conversation
because
I'm
really
invested
in
trying
to
make
this
work.
C
A
So
do
you
look
like
like
I
think
we
did
have
like
some
one-off,
like
I,
think
Allen
you
set
up
the
zoo
meeting
or
we
stole
some
of
the
other
Zoom
meetings
right
we
did.
We
did
have
that
like
in
the
past
as
well.
So.
D
C
So
this
time
it's
really
good
for
me
at
the
moment,
except
I'm
supposed
to
be
in
the
mean,
but
the
what
I
will
do
is
on
my
way
back
on
Friday
I
will
try
and
see
if
I
can
finalize
the
versioning
proposal
document,
so
obviously
not
final
text
or
anything,
but
just
what
I
think
we
should
be
pushing
out
to
customers
and
telling
them
about
so
and
then
that
can
be
the
basis
of
a
a
meeting
about
versioning
and
that
we
can
all
have
some
pre-site
of
what
all
of
that
is
and
then
have
a
proper
conversation
about
it.
A
Yeah
yeah
so
once
you're
back
like
two
things
like
you'll
write
like
a
mini
proposal
or
draft
about
the
library,
versioning
and
second
you'll,
give
a
time
zone
or
actual
time
for
us
to
schedule.
A
one-off
ad
hoc
meeting
and
Allen
can
schedule
the
thing
in
the
calendar.
So
we
can
have
that
ad
hoc
meeting
yeah.
C
I
think
it
would
be
really
important
for
Riley
to
be
part
of
that
conversation
because
they
obviously
were
really
involved
in
that
and
I
wouldn't
want
us
to
not
have
Riley
involved
in
that
conversation
and
then
have
a
a
Twitter
GitHub
issue
battle
again
about
that.
So
I
think
Riley
is
an
important
person
into
that
meeting.
Yeah.
A
Yeah
I'll,
let
him
know
I,
I
I,
don't
think
we
can
always
proceed
like
that.
We'll
always
have
to
give
people
an
opportunity
to
comment.
So
it's
not
just
really
like
anyone
else
can
come
in
right
he's
not
a
maintainer.
So
we
cannot.
A
Yeah
so
I'll,
let
Riley
know
to
attend
that
once
we
have
the
time
I.
C
A
A
Like
for
us,
we
are
in
the
same
time
zone
like
Alan,
Blanche
myself.
We
are
all
in
the
Pacific
time
zone,
so
it
should
be
like
really
easy
for
us
to
find
a
common
time.
Yeah.
C
C
D
C
They
will
that
may
well
happen
in
March,
if
you're
all
going
to
be
there,
because
I
will
be
hopefully
in
Redmond
for
21st
of
March
for
MVP
Summit.
As
soon
as
I
was
made
an
MVP
last
week,.
C
Yeah,
it
was
a
little
bit
too
far
ahead
for
me.
I
want
to
get
inside
a
little
bit
before
the
end
of
March.
So
but
anyway,
I'll
send
you
over
some
times.
I'll,
hopefully
get
that
proposal
done
before
next
week,
and
then,
let's
see
if
we
can
have
that
conversation
and
get
that
sorted
sooner
than
later,.
C
A
C
A
B
A
I
think
we
can
end
so
like
plans
like
immediate,
like
only
thing
we
need
to
do
right
away,
is
to
update
the
pr
like
with
the
description
saying
what
what's
the
end
result
going
to
look
like
once
this
PR
is
in,
and
if
you
can
ping
David
folder
to
comment
on
it.
That
would
be
great
as
well
and
yeah,
we'll
all
review.
It
sounds.