►
From YouTube: 2020-10-26 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
A
C
A
D
Yeah
cool,
so
we
have
been
having
fun
with
formatting
those
on
different
versions
of.
A
A
Okay,
so
I
I
see
two
topics
for
this
week.
The
number
one
thing
is
max.
I
saw
your
big
pr
about
the
standard
library.
D
I'll
fix
the
ci
issues,
yeah
I'll,
probably
target
the
work
in
progress,
but
I
should
be
done
with
it
like
end
of
day
today
we
can
probably
start
like
resume
our
conversation,
yes
yeah
and
I'm
trying
to
document
that
like
because
there
are
so
many
questions.
I
thought
that
would
be
best
if
we
capture
this,
like
in
the
mcdonald
dock,
if
you
guys
agree.
D
Yes,
yes,
indeed,
I
already
tried
reducing
it.
This
is
roughly
half
of
what
I
originally
said.
I
can
split
it
into
non-functional,
just
headers,
let's
say
just
to
illustrate,
what's
in
without
any
of
the
rest
of
the
build
alpha
which
may
trip
the
code
coverage
warning,
because
we're
still
building
everything
the
way
how
where
we're
building
and
the
code
is
still
going
to
be
operating,
the
way
how
it
was
operating
like
I
can
split
it
into
two
parts
code
and
building
for
if,
if
that
makes
sense,.
A
So
my
suggestion
would
be
having
relatively
smaller
prs
and
for
anything
else,
try
to
make
sure
they
work.
They
pass
the
ci
for
code
coverage,
no
need
to
worry
at
this
moment.
D
It
used
to
be
passing
ci,
I
pulled
the
latest
master
and
now
I
need
to
redo
things.
Yes,.
A
Okay
and
I'll
I'll
take
a
look
and
josh
will
probably
need
your
input
as
well.
So
we
talked
about
this.
The
idea
is,
we
want
to
be
able
to
support
like
scenarios
where
people
are
saying
they
just
want
to
use
the
standard
library,
no
need
to
use
the
non-standard
one
and
give
a
better
compliance,
security
and
smaller
binary
size,
and
they
don't
worry
about
cross
compiler,
ebi
compat.
D
I
tried
addressing
some
of
these
in
the
dark
and
there
are
a
few
tricky
things
with
respect
to
variant,
upsell
variant,
the
lack
of
variant
in
msgsl
library,
the
fact
that
our
current
code
that
uses
embark
variant
doesn't
actually
even
compile
with
visual
studio
2015.
D
So
the
variant
that
we
have
today
is
not
going
to
work
on
15.
We
need
something
else
and
then
a
lot
of
interesting
things
with
the
way
how
we
build
static
library,
it's
just
like.
If
we
build
static
library
with
15,
we
can
use
it
with
17
and
19,
but
not
the
other
way
around.
If
we
build
a
library
with
19,
then
it's
no
longer
usable
by
older
compilers,
because
there
are
certain
default,
build
flags
like
with
respect
to
exception,
handling
it's
even
described
in
the
documentation.
D
Pretty
much
from
all
the
you
actually
need
to
use.
The
latest
linker
of
all
of
the
leaves
static,
leaves
you
use,
which
means
that,
in
order
to
be
more
universally
compatible
like
for
the
static
flavor,
we
probably
have
to
be
shipping
like
15
base,
build,
which
then
triggers
another
slew
for
other
issues,
such
as
changing
the
variant
from
empire
variant
to
upsell
variant,
because
that's
the
only
one
that
I
got
working
on
old,
compiler
anyways,
let's
probably
discuss
it
in
the
pr
it's
tricky.
I
can
elaborate
in
a
shot.
D
If
you
guys
have
questions
on
this.
E
Yeah,
the
the
main
thing
that
I
think
was
needed
was
was
just
the
understanding
the
what
happens
if
I
use
the
standard
library
version
of
the
build
and
like
what
is
no
longer
kind
of
supported
so
like
as
long
as
it's
clear
to
people
who
know
you
know
I
I
plugged
in
the
standard
library
or
I'm
using
the
2015
version
right,
you
have
a
very
clear
documentation
for
windows.
D
And
I
think
we
need
to
make
sure
that
when
we
build
packages
we
by
default
provide
pre-built
binaries
with
no
std
option.
So
there's
no
question
about
this,
like
if
somebody
seeks
for
easy
and
good
and
building
like
in
a
shared
run-time,
shared
libraries,
environment,
they'll
get
no
std
flavor.
But
if
somebody
wants
to
tinker
with
this
and
statically
link
in
their
product
or
maybe
build
a
header
on
the
implementation
of
the
library,
then
they
would
actually
have
to
go
to
that
advanced
topic
document
and
understand
what's
possible
and
what
yeah.
F
Yes,
yeah
so
mark
you'll,
probably
need
to
that
was
last
week,
actually
nope
we're
not
on
vacation,
or
that
was
two
weeks
ago.
I
think
oh.
D
F
A
Okay,
so
for
the
for
the
logging
part,
I
if
I
look
at
the
current
status,
I
I
think
we
started
by
having
a
like
a
smaller
scope
between
aws
and
microsoft.
It's
just
because
they
have
interns
who
are
pretty
new
to
this
project
and
they
have
a
lot
of
questions
regarding
to
spec.
But
looking
at
the
current
progress,
I
I
think
they're
ready.
So
my
ass
is
to
switch
all
the
discussions
and
prs
to
the
public
report
to
give
others
more
visibility.
So
I
I
would
expect
starting
from
this
week
we
can.
D
Yes,
that
would
also
help
with
the
ttw
exporter,
because
michelle
part
of
linux
foundation
project
the
student.
He
is
remote
and
best
timing
for
him
would
be
like
to
join
the
scroll
the
other
week
where
we
have
it
in
the
morning.
A
F
We
had
a
meeting
with
alelita
about
that
this
morning
and
there's
still
a
couple
of
changes
that
she
wanted
us
to
make.
But
for
sure,
we're
aiming
for
later
this
week
to
have
the
pr
merged
with
upstream.
E
F
Do
you
mind
if
I
ask
my
manager
if
she's
okay
with
that
first,
because
I've
actually
shared
a
document
before
and
gotten
some
like
gotten
in
trouble
about
that?
So
I
want
to
make
sure
that
it's
okay
to
share
the
pr
that
totally
fine,
totally
fine
I'd
love
to
see
it.
But
yeah
like.
A
G
G
The
pr
I
think
should
be
finalized
by
tomorrow,
so
hopefully
you
could
take
a
look
at
it.
Then.
A
G
Yeah,
we've
also
uploaded
a
copy
of
the
design
doc
onto
one
of
our
team
repos,
but
we
won't
be
submitting
that
in
the
pr.
Yet
it's
just
going
to
be
referenced
to
know
that
that
exists
and
we'll
be
finalizing
it
and
printing
that
later,
but
the
api
pr
is
probably
going
to
be
ready
for
tomorrow.
It's
just
elia
needs
to
take
one
final,
look
at
it
and
then
we
can
ship.
A
It,
okay
cool,
so
josh,
just
a
quick
question,
probably
not
related
here.
So
when
you
mention
like
on
your
side,
you're
very
interested
in
logging.
Does
that
cover
only
the
c
passport
or
someone
from
your
team
or
company
would
be
also
interested
in
the
chart.
Part.
E
E
Regarding
logging
yeah,
I
mean
we're
interested
in
all
of
open
telemetry,
but
specifically,
we
might
have
people
willing
to
invest
in
the
implementation
as
well
just
trying
to
get
understand
what
the
status
is
and
and
then
kind
of
get
a
look
at
it
see
if
we
need
to
help.
A
Okay,
so
just
for
the
information,
the
open,
telemetry.net
part,
the
login
support
and
integration
is
almost
there.
The
only
thing
that
we
don't
have
is
the
plumbing
tool
otlp.
So
what
we
have
is
the
integration
with
the
donet
native
logging
api,
the
I
logger,
so
people
can
use
I
logger
and
we
have
the
exporter
processor
and
we're
working
on
the
sampler.
A
What
we're
lacking
is
the
ability
to
take
the
logs
and
send
to
the
otlp,
so
we
don't
have
the
log
exporter
for
otlp
and
that's
probably
fine,
because
the
otlp
logging
part
the
pro
like
the
protobuf
definition
is
still
in
the
air.
So
we're
also
interested
in
finalizing
that
if
you
know
any
anyone
who
might
be
interested
in
this
or
habitual
collaborate
as
well.
E
Yeah
sounds
great
hope.
I
I
was
yeah,
hopefully
in
the
next
c,
plus
plus
meeting
we'll,
have
a
few
more
people
from
google
we'll
see,
okay,
cool.
Thank
you.
C
So
tom
yeah,
I'm
asked
so
many
questions
beforehand:
yeah,
I'm
still
interested
in
the
ga
gap
for
our
three
and
three
trees
and
api
and
sdk
as
we
I
think
we
discussed
we
were
targeting
to
reach
in
ga
state
for
trees,
api
and
sdk
at
the
end
of
this
year,
so
which
is
probably
two
months
left
right
and
yeah.
C
I
looked
at
the
project
page,
which
lists
the
to-do
work
items
and
some
some
items
which
is
in
progressing,
but
mostly
I
saw,
is
that
there
are
two
exporters
and
a
few
bags
there,
or
I
mean
I
mean.
Do
I
need
to
look
at
the
specification
matrix
to
get
the
exact
gap
or
like
to
pick
up
what
item
to
work
or
just
focus
on
the
project.
A
Page,
okay,
so
so
this
is
the
single
source
of
truth
and
the
spec
could
change
so
we're
having
multiple
moving
targets.
The
spike
could
change
and
people
could
add
new
stuff,
they
could
change
the
behavior,
and,
even
if
today
we
look
at
c
plus
plus
we're
saying
the
tracing
part
is
done
like
everything
is
plus
that
doesn't
mean
it
will
be
the
reality
tomorrow.
People
could
change
and
add
a
new
role
or
they
could
change
the
plus
two
minus
and
that's
why
we
have
this
coordination
and,
in
the
open
time,
respect
and
maintainers
meeting.
A
Nothing
will
try
to
nothing,
will
be
ga
for
open
time,
3c
plus
fast
we're
in
the
early
stage.
So
we
want
to
get
the
list
of
the
ga
language.
There
is
a
maintainer
meeting
and
there's
public
blog
post
on
the
open,
telemetry
official
document.
I
believe
these
are
the
languages
we
try
to
create
like
go
java,
javascript
python
and
dotnet.
A
A
So
in
opentimetree.net
I
figured
out
it's
impossible,
because
if
we
claim
that
open
climate.net
is
0.5
alpha,
that
means
we
got
to
do
a
lot
of
work
that
is
no
longer
existing
in
the
latest
spec,
which
doesn't
make
sense.
So
we
we
talked
with
the
the
community
and
decided
okay.
We
can't
have
a
beta
as
long
as
we
catch
up
with
most
of
the
items,
but
for
the
ga
we
have
the
agreement,
what's
the
what's
the
bar,
so
in
c
plus
plus,
I
I
think
most
likely
we're
going
to
follow
that.
A
So
to
answer
your
question,
no
strict
definition
from
the
spec,
we
have
the
the
freedom
to
define
what
it
means
for
alpha
and
my
proposal
is
the
initial
release.
Should
have
the
the
tracing
part
down,
so
it
has
to
have
a
hundred
percent
compliance
with
the
the
g8
version
of
the
tracing
spec
and
for
the
matrix
part.
We
need
to
make
it
very
explicit
that
this
is
only
experimental.
The
isdk
has
a
lot
of
global
logs.
It's
not
meant
to
be
used
by
any
production
services
and
the
login
part
is
a
preview.
E
Clarification,
so
can
I
ask
a
question
about
that?
Then,
yes,
please,
it'd,
be
it'd,
be
launching
an
alpha
right.
So
so
I
guess
two
questions.
One
is
what
does
a
release
of
sequels
plus
look
like
because
since
we
don't
have
one,
I'm
just
kind
of
curious
like
what?
What
do
you
envision
that
actually
like,
physically
or
mechanically?
Being
you
know,
is
it
a
tag
on
the
repo
or.
A
Is
it
actually
artifacts
okay,
so
my
understanding
initial
release
can
be
as
simple
as
a
tag
on
the
ripple,
because
we
we
haven't
decided,
what's
the
scope
for
the
binary
like
do
we
have
static
link
libraries?
Do
we
have
dynamic
link
libraries?
Do
we
offer
windows,
dlls
or
linux
iso
files,
those
are
not
yet
figured
out,
or
do
we
allow
any
like,
like,
like,
like
iso
files,
to
be
like
to
have
the
freedom
to
move
to
any
address?
So
basically,
then
have
absolute
jumps.
A
A
I
would
be
very
careful
to
promise
any
binary
release,
because
once
you
have
binary,
it
has
a
whole
bunch
of
combinations,
and
I
I
think
it's
just
hard
for
us
like
like
how
many
number
of
compilers
and
which
platform
dependency,
those
things
and
and
once
you
have
static
link,
you
got
to
provide
symbol
files
and
you
got
to
provide
script
versus
the
full
symbol.
So
there's
a
lot
of
implication
and
with
the
current
setup,
like
just
a
few
of
the
members
here,
I
think
it's
impossible
for
us
to
support
all
of
them.
A
E
Okay,
I
have
a
second
question:
that's
somewhat
related,
so
in
the
past.
So
just
so
you
know
my
history
with
open
source.
I
used
to
be
the
scala
language
like
release
manager
and
we
attempted
to
mark
something
as
not
ready
for
usage
by
the
general
population
and
then
every
single
freaking
person
and
their
brother
relied
on
it.
E
Okay,
to
the
point
where
we
were
basically
unable
to
make
changes
to
that
ever
again,
because
even
though
we
had
tried
to
protect
ourselves
and
say
this
could
break
at
any
time,
please
don't
rely
on
this.
Please
don't
rely
on
this.
Our
users
did
so
anyway.
E
E
I
think
that
might
be
the
case
for
some
of
the
stuff
here
in
cepa
plus,
so
I
just
want
to
know
like
what
what
we're
thinking
there
and
like
if
we
have
any
plans
to
deal
with
that,
if
it's
just
gonna
be
we're
gonna
break
you
and
you
can
complain
as
much
as
you
want.
But
you
know
we
need
to
be
able
to
break
this
or,
if
we're
going
to
do
our
best
to
not
make
breaking
changes
as
often
as
we
can
even
for
things
that
are
preview,
I
see
so.
A
I
hope
that
we
can
clarify
here
and
my
my
thinking
is
for
tracing
part.
I
would
be
careful
even
if
we
released
the
first
version
I
I
would
try
the
best
to
not
break
any
of
the
interfaces,
because
the
tracing
part
of
the
spec
is
almost
done
and
when
we're
releasing
the
the
initial
like
tracing
sdk,
the
spike
should
be
already
finalized
and
we
like
at
least
I
I
got
experience
working
on
open,
telemetry.net
and
python,
so
I
I
think
we
can
achieve
that
for
long
in
part.
I
I
I'm
thinking
that.
A
E
That's
fun
so
what
where's
the
zero
percent
confidence
coming
from
just.
A
E
A
So
I
know
what's
what's
coming
so
I
I
know
that
the
metrics
I
stick
at
that
time.
We
got
several
interns
and
they
were.
They
were
running
short
of
time.
So
as
agreement,
we
let
them
go
just
to
finish
the
intern
project
and
get
things
hooked
up.
So
we
have
a
great
demo,
but
underlying
we
look
at
the
memory
allocation,
the
risk
condition
and
the
the
contention
is
it's
just
not
something
that
you
can
use
for
any
serious
product.
A
E
Yeah,
yeah,
cool
cool.
Okay!
That's
that's!
That's!
That's
a
good
answer!
So
you're
80,
confident
with
logs
and
100,
confident
or
99,
confident
with
traces
and
metrics,
is
the
thing
that
we
know
will
change
cool.
A
D
That
not
a
problem
at
all.
I
was
just
trying
to
say
that
for
the
vc
package
package
manager,
which
some
of
our
internal
and
external
customers
are
using
so
we'll
see
make
base
as
long
as
we
have
cmake
set
up
up
and
running,
you
can
also
add
various
flavors
from
the
belt
like
the
stuff
that
we
discussed,
for
example,
a
default
flavor,
which
is
now
an
std
versus
the
other
flavor,
which
may
be
what
which
was
for
using
standard
library,
header,
only
sdk
usage.
D
Then,
basically,
the
users
would
decide
and
it
works
like
fetching
it
hub
location
for
which
we
will
have
to
maintain
tags
and
releases.
So
we
will
have
to
drive
tags
and
then
for
every
port
file
that
we
submit
in
the
mainline
vc
package.
D
We
refer
to
some
tags
and
then
pretty
much
build
from
source
happens
with
customers
compiler,
and
that
is
the
case
where
api
is
less
relevant
because
most
likely
they're
going
to
be
statically
linking
a
header
only
including
that
thing
I'd
be
more
worried
about
the
pre-built
binaries,
so
for
something
like
envoy
or
engines
or
other,
where
devops
would
prefer
to
get
it
and
not
build
it.
D
D
D
Well,
one
doesn't
invalidate
the
other,
but
maybe
if
you
make
it
work
through
lua
scripting,
if
engine
supports
it
and
then
have
some
other
bindings
from
that
lua
script
to
c,
plus
plus
cool,
then
if
you
really
want
perf,
probably
you
have
to
build
some
c
ffi
and
then
from
that
c,
ffi
figure
out
how
to
invoke
this
support
for
sdk,
because
we
don't
have
c
api
right
now.
A
Yeah,
I
remember
the
python
sig
has
mentioned.
They
might
have
some
interest
in
like
doing
foreign
function
interface
with
this
project,
because
in
python
there
there
are
certain
libraries
that,
like
90
percent
of
the
code,
was
actually
in
c
plus
class
and
python
is
just
a
form,
function,
interface,
writer
on
top
of
that,
and
they
want
to
be
able
to
stitch
things
together.
A
D
Maybe
we
should
start
a
separate,
seek
for
open,
telemetry
c
and
have
two
separate
tracks.
One
is
just
a
wrapper
over
c,
a
plus
plus
core,
where
you
have
c
api
spec,
how
it
looks
like
then
binding
to
existing
sdk
implementation
that
we
have,
maybe
somebody
would
say
no.
I
don't
even
need
like
a
pure
c,
no
stl
at
all,
and
then
those
guys
would
have
to
do
like
from
scratch
right.
D
And
even
like
for
kernel,
I'd
say
no:
how
about
use
something
like
user
mode
helper
or
something
or
build
some
other
different
pipe
into
a
service
which
is
already
written
in
c,
plus
plus
and
then
from
there
use
the
c
plus
plus.
A
A
Yeah
for
for
this
one,
I
think
it's
almost
ready
I'll
I'll,
take
a
look,
and
for
for
this
pr,
I
think
lalit
has
received
several
comments
and-
and
it's
not
closed
any
outstanding
thing-
that
you
think
this
is
going
through
the
wrong
direction
or
or
the
direction
seems
correct.
It's
just
like,
like
the
implication
details.
E
I
think
I
think
I
had
a
comment
on
distribution
just
the
only
if
this
is
the
one
yeah,
but
my
concern
is
mostly
around
if,
if
we
are
depending
on
this
library
and
we're
pulling
it
in,
how
do
we
know
what
version
of
the
library
where
we've
pulled
in?
E
How
do
we
know
if
it
goes
stale
and
then
how
do
we
deal
with
somebody
who
wants
to
compile
with
that
library
and
our
library
and
there
might
be
a
version
conflict
sku
where,
like
it
breaks
everything,
that's
the
only,
and
that
is
probably
99
documentation
and
zero
percent
implementation.
But
that's
that's
the
only
bit
that
I
was
was
concerned
about
with
that
pr.
D
We've
also
briefly
chatted
with
lalit
on
this.
Maybe
we
should
pull
in
the
lib
event,
because
lib
event
appears
to
be
more
mature
in
general
and
more
often
more
frequently
used
and
also
more
friendly
less
than
the
license
than
this
guy.
D
I'm
a
bit
skeptical
about
the
brand
new
fresh
library
designed
by
some
dude
in
2020,
with
due
respect
it
is
functional.
But
what
about
going
back
to
that
lib
event,
which
I
think
orion
used
in
another
pr
for
the
async
networking
that'd,
be
a
much
better
option
if
we
use
that-
and
it
also
has
its
own
http
client.
D
So
we
can
bind
from
our
let's
interface
to
web
event,
http
client
and
also
we
discussed
that
he's
going
to
look
at
a
posix,
and
I
can
provide
the
draft
when
I'm
net
based
client
in
our
repo
licensed
as
copyright
open,
telemetry
authors
not
for
more
or
less
for
unit
tests,
because
I
think
for
production
we'd
probably
be
recommending
with
web
event.
D
D
D
Libra
vent
is
a
cross
plan
and
for
windows
we
can
also
have
our
own
http
client
integrated
again
just
for
unit
tests,
because
I
think,
even
on
windows,
we
should
be
able
to
use
lib
event
minor
concern
I
had
while
researching.
This
is
criminal,
for
some
reason
does
not
seem
to
be
using
it
on
windows,
even
though
it
is
about
going
complex.
Labor
event,
just
like
leap
event
has
a
steep
client
everywhere
and
then
on
windows.
We
have
alternatives.
A
Yeah,
so
I
I
do
share
similar
concerns.
As
josh
mentioned,
like
in
open
telemetry.net,
we
we
used
to
have
a
dependency
on
the
on
the
third
party
business
item.
There
were
a
lot
of
questions,
for
example,
if
they
changed
the
behavior.
For
example,
when
we
got
notified
there
is
a
security
issue
and
we
got
update
to
the
latest
version,
but
during
the
update
we
notice
they
have
a
breaking
change.
Some
behavior
will
will
just
be
different
from
from
before.
What
are
we
going
to
do?
A
D
There's
a
way
out
of
this
so
for
bazel,
I
think,
like
I,
I
see
that
it
has
references
to
exact
version
of
commit
id,
whereas
for
cmak
we
can
also
rely
on
sub
modules
infrastructure.
So.
A
D
A
A
A
Then
we
can
roll
forward
with
an
update
over
third-party
dependencies
as
well.
There
there
will
be
a
cross-language
sdk
guidance
on
the
security
and
the
support
before
vga,
but
I
like
so
just
try
to
summarize
max.
We
we
need
to
cover
security
considerations,
sure
yes,
yeah,
absolutely
in
case
there's
zero
day
we
got,
we
got
to
update,
we
got
to
move
forward
and
by
up
updating
to
the
latest,
like
rendering
version,
we
might
break
the
like
the
behavior,
and
in
this
way
we
got
to
decide.
A
Even
if
we
didn't
change
any
api,
and
that
seems
to
be
scary,
so
we
got
to
be
very
careful
and
in
this
way,
depending
on
the
effort
like
if
it's
done
at,
I
would
rather
go
and
implement
something.
Instead
of
taking
the
dependency.
D
But
for
bigger
projects
we
see
their
velocity,
for
example,
for
lib
event.
It
is
mature
and
stable
and
there
shouldn't
be
a
significant
api
or
api
breaking
changes.
A
E
Yeah
I
I
would,
I
would
also
vote
for
lib
event
over
the
this
header,
only
one
just
from
the
standpoint
solely
of
it's
been
around
longer.
It
has.
It
has
a
good
track
record
and
we
don't
know
about
this
new
library
right.
We
have
no
idea,
there's
not
enough
data,
so
it's
just
it's
a
little
bit
too
much
of
a
risk.
D
And
I
think
lalit
also
agreed
on
this
in
a
chat
we
had
this
side
conversation
in
particular
literature.
A
Yeah,
I
agree
okay,
so
I
remember
like
at
some
point
we
took
a
dependency
on
an
npm
package
that
was
very
famous
and
it
was
developed
by
by
a
college
student
and
and
later
he
removed
that
npm
package.
I
have
no
idea
that
people
could
remove
npm
package
after
it's
published
and
they
put
a
message.
I
got
a
girlfriend.