►
From YouTube: 2020-11-10 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).
C
C
C
C
Well,
I
only
have
one
topic
on
my
site
and
one
follow-up
on
my
site,
so
we
mentioned
that
we
got
to
clarify
this
project
only
supports
c
plus
plus
it
does
not
support
c.
For
now
I
haven't
seen
the
player
yet
I'll.
Send
that
so
and
the
topic
I
have
I
put
on
the
agenda-
is
the
login
sdk
update,
but
before
that
I
won't
know,
like
any
other
topic
from
you,.
D
D
My
apologies
for
being
a
bit
school
with
the
cross
like
different,
compiling
support.
There
was
a
topic
in
the
chat
about
how
we
can
cook
different
sdk
profiles,
different
flavors,
with
vc
package,
one
of
the
community
members
proposed
how
we
can
use
feature
packages
for
that,
but
we
still
don't
know
how
to
do
it
best
for
base
of.
D
D
Yeah
for
vc
package,
I
can
probably
send
in
the
pr
like
where
we
say
vc
package
install
specifying
the
repo
and
then
there
could
be
an
option,
a
default
being
the
no
std
library
and
a
custom
option
like
std
lib,
which
builds
you
the
package
with
the
standard
library
classes,
again
not
being
the
default,
because
I
think
the
default
is
like
cross.
B
Yeah,
so
the
the
thing
with
bazel
is
there
it's
not
like
there's
a
default
and
then
like
a
different
one,
there's
just
going
to
be
split
targets
for
everything,
but
what
we
can
do
is
we
can
make
one
name
be
like
the
common
name
that
people
will
choose
and
another
name
have
like
an
extension
that
denotes
it's
with
standard.
B
Does
that
sound
reasonable?
I
should
probably
send
out
a
pr
with
what
I'm
thinking
about
doing.
Okay,
based
on
what
kind
of
works
so
I'll
see.
If
I
can
get
that
out
this
week,
for
you,
I'm
not
sure
what
time
it'll
come
out
but
I'll.
Send
you
a
pr
against
your
pr
with
bazel,
build.
D
George,
I
have
another
for
question
about
it:
w
exporter,
my
student
is
working
on
ntw
export
and,
like
this
is
event
racing
for
windows.
So
it
applies
only
to
windows
belt
right
now
how
he
handled
it
when
he
is
invoking
bazel,
he
just
specified
minus,
and
then
he
excludes
the
etw.
If
he's
running
unix
loops,
I
was
wondering
if
there's
some
construct
that
allows
the
blazer
level
specify
okay,
if
I'm
building
on
like
win32
then
exclude
this
this
this
like
for
of
tell
based
on
what
platform
we
are
building
so.
B
B
D
It's
a
separate
project,
it's
an
optional
component,
so
it's
not
a
required
component.
It's
not
the
default
exporter
per
se.
It's
just!
We
want
to
build
it
on
windows
only
and
if
somebody
needs
it,
they
can
use
it
like
plug
it
in,
and
I
think
maybe
just
to
alleviate
this
hustle.
D
For
now
what
the
student
is
doing
is
probably
okay,
like
I
personally
think
it's
kind
of
ugly
when
we
have
to
specify
minus
like
without
this,
rather
than
having
a
con
like
the
intrinsic
knowledge
that,
where
to
build
it
or
not
like
using
that
force,
opt-out
is
kind
of
a
bit
weird,
but
probably
that's
what
we
will
integrate.
B
So
that's
for
now,
okay
I'll
see.
If
I
can
get
you
a
solution,
quick
quickly,
but
yeah
like
like
that
pr
is
ready
to
submit.
Now
you
mean
he.
D
Still
needs
to
clean
the
export
api
implementation,
because
what
is
in
is
more
like
a
completely
right
is
the
key
not
like
it's
packaged
through
the
proper
exporter
interface.
It's
like
okay,
there's
a
there's,
a
tracer
which
fully
implements
the
full
sdk
end
to
end
without
actually
being
plugged
in
as
exporter.
So
he
still
has
to
work
on
that.
But
his
pr
is
compilable
and
passing
all
checks
and
build
checks
with
the
hack
that
this
component
is
excluded
on
unix.
B
Okay,
cool
I'll
see
if
I
can
get
a
better
solution.
That's
that
doesn't
use
minus
that's
good!.
C
Okay,
so
on
the
next
topic
I,
by
the
way
I
noticed
so
richard
hi,
do
you
like
introduce
yourself
and
say
hi
to
everyone.
E
C
Yeah,
so
I
I've
seen
richard
before
in
my
meeting
with
aws.
So
if
I
remember
correctly,
your
principal
dev
from
aws.
E
Yes,
that's
right,
I'm
fairly
new
to
aws,
but
I'm
a
principal
engineer
at
amazon.
I
used
to
actually
work
at
google
for
a
long
time
as
an
sre,
so
just
fyi.
So
it's
kind
of
novel
to
see
basil
build
stuff
again.
E
C
Got
it
cool
thanks,
so
we
we
can
talk
about
the
logging
sdk,
so
just
for
people
who
want
to
know
the
context,
so
the
login
api
part
based
on
the
initial
like
prototype,
that
pr
got
merged.
Although
there
are
a
couple
open
questions
like
do,
we
allocate
the
log
record
on
the
stack
or
the
heap
and
things
but
like
in
order
to
make
progress
currently
like
the
second
pr
is
already
there
on
the
logging
sdk.
So
please
take
a
look
and
mark
karen.
F
So
in
the
pr
that
we
just
submitted,
that
includes
just
the
implementation
for
the
logging
api,
the
rest
of
the
logging
sdk
will
be
coming
soon
and
that
will
include
a
simple
and
batch
processor
as
well
as
an
exporter
interface.
F
So
the
scope
of
the
current
pr
is
just
all
of
the
overloaded
methods
of
the
api,
as
well
as
some
methods
to
be
able
to
get
inside
a
processor
in
the
logger
provider
and
to
be
able
to
construct
actual
logger
and
logger
provider.
Instances
and
there's
still
some
comments
that
are
being
resolved
on
the
pr
right
now.
So
hopefully
we
hope
to
finish
that
soon
and
if
anyone
has
any
insight
on
that,
please
just
let
us
know
on
the
pr.
C
Yes,
I
put
the
pr
link
here
and
notice
currently,
like
I,
I
spent
some
time
and
also
I
see
some
feedback
from
max
and
lalit,
and
I
won't
want
to
see
like
if
josh
or
johannes,
if
you
could
put
your
comments
there,
because
I
think
this
login
part
is
important
for
everyone,
just
some
high-level
input.
For
example,
we
have
some
debate
where
the
logger
should
have
the
name.
C
D
D
C
Yeah,
I
I
agree
so
in
general,
I
want
to
get
more
pair
of
eyes
instead
of
like
just
like
three
microsoft
folks,
giving
like
two
interns
from
aws.
That
seems
to
be
a
dangerous
thing
for
if
we
believe
this
logging
api
is
important,
so
I
want
more
eyes
from
your
relic
from
google,
ideally
from
dynamics
and
other
companies
as
well.
That's
my
general
ask.
B
Yeah
I
I
can.
I
can
take
a
look
tomorrow
morning,
eastern
time,
the
quick,
quick
question
on
this.
What
is
the
timeline
for
when
this
is
expected
to
kind
of
get
used
by
people?
B
I'm
just
kind
of
curious,
because
we
so
we
internally,
we
have
people
who
want
to
use
the
on
the
wire
format
of
otlp,
and
so
I'm
really
really
kind
of
curious
to
see.
Logging,
sdk
work
and
log
spec
work
prior
to
ga
and
I'm
kind
of
curious
with
this
logging
sdk
like
what?
If
I
can
ask
what
kind
of
timeline
are
you
looking
at
for
having
this
kind
of
be
working
and
kind
of
stablish
like
what's
the
goal
there.
C
C
C
B
So
that
yeah-
that's
that's
fair.
My
my
fear
is
the
perception
that
I'm
hearing
from
other
people,
not
myself,
is
that,
oh,
if
amazon's
contributing
to
logging
sdk
right,
which
is
what
they
see
and
if
people
start
using
it
well,
they
won't
be
able
to
break
it
because
too
many
people
will
rely
on
it
at
release.
B
That's
the
concern
that
I'm
worried
about
here
in
the
sense
of
it
might
be
true
like
it
might
be,
that
this
is
a
unmet,
need
that
lots
of
people
want
and
a
lot
of
people
might
adopt
it
and
then
expect
it
not
to
break,
even
though
we've
told
them
it
will
break.
That's
that's
my
my
my
concern.
So
that's
all
I
kind
of
want
to
just
find
out
timelines
how
it's
going
to
be
presented
to
people,
what
the
expectations
are
from
us
and
then
just
make
sure
that's
communicated
effectively.
C
C
C
That's
not
the
case,
and
I
have
plenty
of
storage
to
share
from
the
open,
telemetry
downl
cleaned
up
a
lot
of
apis
that
currently
customers
are
using
and
we
work
with
them,
and
it's
not
just
from
from
my
team
and
also
I've,
seen
people
from
other
companies.
They
have
the
same
thing,
so
we
follow
exactly
what
we
specified
in
the
dock.
It
is
not
generally
available
and
we're
going
to
break
that.
C
D
D
This
is
not
what
will
be
in
the
ga
and
we
will
help
them
migrate
to
what
we're
working
on
right
now
like
on
that
current
unlocking
api
surface,
because
right
now
in
the
higher
level,
our
code,
the
customers
end
up
implementing
that
same
stuff,
like
log
levels,
verbosity
controls
and
all
that
and
that's.
Why,
like
there's,
clearly
a
need
for
that,
and
hopefully
this
year
we
should
have
something
functional
and
in
fact,
for
the
etw
export.
B
Okay,
yeah,
that's
fair,
I
my
fear
is
just
if,
if,
if
it
really
is
a
big
need
in
c
plus
plus-
and
it
gets
crazy
adoption
because
it
has
the
appearance
of
being
stable,
then
we
can't
really
change,
even
if
we
wanted
to
just
based
on
the
sheer
number
of
people
who
would
try
it
and
ex
and
like
start
relying
on
it
right
like
I,
don't
know
how
many
customers
you're
talking
about
migrating.
Hopefully
it's
not
a
lot
same
same
for
us.
B
If
we
try
it
out,
I'm
just
worried
that,
like
the
all
signs
point
to
I
used
to
be
in
the
scala
community,
so
scala
macros
were
an
example
where
we
tried
to
keep
something
experimental
and
tell
people
to
try
it
out
and
not
use
it,
and
then
so
many
people
adopted
it.
We
couldn't
remove
it
when
we
realized
it
was
a
terrible
idea
because
it
was
just
that
fundamentally
useful
to
so
many
people.
B
C
C
We
got
some
someone
angry
say:
hey,
I'm
looking
for
a
particular
like
company
xyz,
and
we
already
use
that
api
and
now
we
broke
up
and
we
send
them
the
description
and
we
need
to
negotiate
with
them
and-
and
I
can
tell
so
far
a
hundred
percent
of
the
cases
we
convinced
them
that
we
got
to
move
forward.
So
we
give
them
a
way
to
implement
what
they
want,
but
not
necessarily
that
we
keep
the
api
and
we're
not
motivated
to
keep
the
api
just
because
a
particular
company
says
they
have
a
dependency
on
that.
C
B
This
yeah,
this
is
just
a
question
of
the
weight
of
numbers
like
if
lots
and
lots
and
lots
of
customers
have
it.
That's
when
I
I'm
not
worried
about
one
company,
I'm
wearing
about
a
couple
hundred
or
or
more
right
that
adopt
it.
Now,
that's
probably
not
the
case
with
simple
plus
logging,
but
that's
what
I'm
suggesting
it's?
It's,
not
it's
not
a
worry
about.
B
One
big
company
either
like
if
it
was
just
google
who
cares
right,
but
if
it's,
if
it's
a
lot
of
customers,
that
that
would
be
my
concern
if
we
see
like
adoption
from
like
20
or
30
and
then
breaking
that
api
gives
them
that
bad
taste
in
their
mouth,
where
they
say,
I
don't
like
open
telemetry.
That
is
not
what
I
want
to
have
happen
right.
C
Yeah
so
so
far
in
open,
telemetry
community
I've
seen
that
once
in
the
open
telemetry
protocol,
we
used
to
have
some
legacy
status
code
and
later
we
want
to
remove
that.
But
people
are
saying
they
already
take
a
bet
on
the
product
and
because
they
have
the
like
the
client
side
and
server
side
and
some
part
is
already
deployed
in
production.
So
we
reach
the
agreement.
C
B
Okay
sounds
good,
sounds
good.
I
I
am
I'm
specifically
it's.
It's
logging
that
I'm
interested
in
and
it's
more
of
the
protocol
and
logging,
but
like
sc
plus
implementation
of
the
protocol
logging.
I
hope
to
share
more
details
relatively
soon
of
what
what
we're
planning
to
do
with
it
and
like
our
use
cases
but
I'll
yeah,
the
the
the
idea
that,
if
a
lot
anyway
it
would,
it
would
not
be
a
google
thing.
C
I
see
yeah
so
so
I
can
share
my
perspective
from
microsoft
side.
I
I
think
I
I
have
a
strong
motivation
to
move
my
production
customers
to
the
open,
telemetry
tracing
and
the
logging
by
end
of
june
next
month,
like
by
end
of
june
2021
and
and
that
number
is
based
on
my
assessment
of
the
progress
of
the
overall
project,
not
specific
to
c
plus
plus
the
the
spec
and
the
readiness
okay.
D
Okay,
I
think
we
we
might
have
a
few
other
options
example.
So
let's
say
there
is
already
some
api
services,
for
example
logs
log4cxx
or
other
login
api
services.
D
As
part
of
our
work
to
onboard
to
actual
open
telemetry
exporters,
there
could
be
a
sort
of
shin
layer
maintained
by
people
who
are
intimately
familiar
with
open
telemetry
that
way.
If
our
login
api
changes,
the
actual
customers,
don't
have
to
instrument
because
they
would
be
utilizing
their
familiar
api
surface.
D
B
Yeah
I
mean
for
from
our
perspective,
what
we
have
is.
We
have
people
using
open
senses
right
and
we
want
them
to
start
migrating
to
open
telemetry,
but
they
look
at
the
website
of
open
census
or
even
open
tracing,
for
example,
and
they
see
this
is
deprecated
and
you
should
use
open
telemetry
and
then,
when
they
come
to
open,
telemetry
they're
like
this
is
not
stable,
and
so
they
just
don't
use
either
right.
B
Yeah
right
and
that's
that's
my
big
fear
right
now,
as
I'm
actually
seeing
that,
and
so
we
have
people
who
actually
want
to
try
open
telemetry
but
are
like
not
even
looking
at
it
and
the
other
alternative
that
we
have
out.
There
is
deprecated,
so
yeah
like
I
understand
that
the
speed
of
the
community
can
only
move
so
fast,
I'm
just
it's
it's
it's
making
me
a
bit
nervous,
so
anything
we
can
do
to
help.
Let
us
know
I
will
help
with
basal
rules
for
sure.
B
C
B
C
Okay,
I
I
can't
put
a
a
I
could
a
road
map
there,
but
I'll
need
your
guys
to
like
put
more
feedback,
because
the
the
actual
execution
is
depending
on
the
investment
like
for
doughnut.
It's
like
75
engineers
from
microsoft.
That's
why
I
have
a
better
control
for
this
project.
I
have
a
probably
like
30
percent
of
the
contribution
here
from
microsoft.
That's
why,
like
it's
hard
to
decide,
what's
the
timeline,
unless
we
got
more
buy-in
from
from
people
like
you
guys.
G
I
also
think
an
additional
kind
of
problem
that
we
have
here
with
giving
timelines
is
that
open,
dynamic,
reciprocals
plus
projects
are
different.
Parts
are
not
very
homogeneous,
like
with
the
tracing
part,
that's
in
my
opinion,
pretty
stable
and
actually
for
neurotic
customers.
I
said
them
in
a
few
months.
I
would
be
comfortable
with
you
using
that.
G
G
So
I
think
we
have
like
these
three
moving
pieces
there
and
I
think
they
will
not
be
stabilized
at
the
same
point
in
time,
but
rather
come
one
after
the
other,
and
I
think
that's
also
a
challenge
for
us
to
communicate
that
to
users
saying
yeah
you
can
use
the
tracing
part,
but
the
matrix
and
locking
part
logging
parts
are
not
that
stable.
Yet
I
think
that's
that's
going
to
be
a
challenge
to
make
it
here.
C
Yeah,
so
one
thing
we
communicated
here
in
in
donet
is
in
the
milestone
we
have
to
clarify.
The
ga
goal
is
to
support
things
like
tracing,
and
the
logging
and
metrics
here
will
only
be
experimental.
There
are
some
glitches
and
misalignment
on
the
timeline.
So
if
folks
think
this
should
be
happy,
I
can.
I
can
create
something
similar
like
how
we
created
in
the
open
chemistry.net
just
to
create
the
milestones
and
clarify
what
we
can
do,
but
with
understanding
that
these
milestones
might
be
a
floating
target.
H
C
H
D
Can
we
have
like
more
agile
planning,
let's
say,
create
monthly
milestones,
then
overflow,
anything
that
has
not
met
the
target
into
the
next
month.
At
least
we
would
have
some
sense
of
urgency
and
like
importance,
I
think
we
should
just
create
the
masters
like
right
away
as
soon
as
possible.
C
Yeah
so
yeah,
so
a
monthly
milestone
would
help,
but
I,
like
initially,
I
didn't
put
there
because
my
worries
it
will
be
just
like
me
entertaining
myself
so
now
with
like
we're
doing
the
planning
for
the
next
six
months.
I
I
think
at
least
I
can
put
1.5
developers
time
for
microsoft
on
this
project,
so
that
gives
probably
like
two
developers
and
and
we'll
try
it's
all
depending
on.
D
Partially,
once
we
make
any
commitment,
let's
say
june
right,
we
still
have
to
create
the
milestone
and
we
still
have
to
chase
for
things
to
happen
right
yeah.
From
that
perspective,
it
doesn't
matter
even
now
we
can
create
the
milestones
that
gives
internal
pressure.
You
know
we
are
all
conscious
beings
and
responsible
people
here,
okay,
so
when
we
get
milestones,
that's
yet
another
motivating
factor
to
get
things
done.
C
D
C
G
Okay,
then
I
mean
with
for
the
milestones
I
mean
I
that
was
partially
my
intent
when
I
set
up
kind
of
this
project
board
for
the
tracing
for
the
trace
api
and
sdk.
I
mean
it's
not
really
a
milestone,
but
at
least
it
wasn't
then
just
to
capture
what
has
to
be
done
to
get
to
this
point,
where
we
have
the
trace
api
and
sdk
kind
of
ready
for
usage.
G
C
Okay,
so
this
actually
brings
out
the
next
topic,
so
I
I
remember
josh
mentioned
to
this,
and
also
I
I
think
this
is
something
I
would
like
to
try.
So
I
don't
have
the
initial
release,
it
doesn't
mean
it'll,
be
a
beta
or
something
it's
just
the
initial
alpha
release.
We
want
to
go
through
the
release
process
to
give
people
idea
what
a
release
could
look
like
and
we
can
take
the
process
to
like.
C
If
we
release
are
we
going
to
tag
something
and
put
the
document
and
change
log,
and
where
should
people
go
and
download
and
where
should
we
put
document
it
doesn't
have
to
be
complete,
but
it's
a
good
start
to
practice
and
see
what's
the
gap.
C
G
C
Is
yeah
what
I'm
thinking
is.
I
can
put
one
developer
from
microsoft
to
work
on
the
initial
release
like
the
changelog
and
I'll
work
with
the
developer,
because
I've
done
that
for
the
download
to
establish
the
process.
We
can
follow
something
similar
here
and
once
we
got
the
initial
proposal,
there
will
be
prs.
I
expect
help
from
you
guys
to
take
a
look
and
see
any
any
gap,
and
probably
that
that
will
be
like
several
prs
and
back
force
discussion.
C
C
C
Okay,
yeah
thanks,
so
so
with
that
I
I
would
see
the
the
most
like
important
prs.
I
would
say
from
what
I
can
see
the
login
part
which
we
like
instead
of
like
matrix,
we
ship
something,
and
we
know
that's,
not
a
good
quality
for
the
metrics
part.
We
want
to
ship
something
that
people
can
use
and
in
production
and
give
us
feedback
and
the
other
thing
you
want
to
do
initial
release,
so
people
can
try
the
tracing
and
logging
part
with
understanding
is
not
complete.
G
Yes,
I
mean
what
I
have
on
my
agenda,
maybe
for
the
I'm
not
sure
how
much
time
we
will
have
in
the
next
two
weeks,
but
for
the
tracing
part
is
definitely
to
verify
that
the
order
context
propagation,
like
the
w3c
stuff,
is
working,
because
when
we
have
the
stuff
there,
but
they
don't
have
any
example
program
and
that
I
think
is
really
like.
B
G
Suite
that
they
have,
because
I
think,
once
that
is
in
place,
I
think
the
trace
api
would
actually
be
kind
of
you
really
usable.
Okay,
because
I
mean
we
just
do
for
some
context:
children,
I
we
were
talking
within
relic
customer
recently,
and
I
mean
what
they
I
think
for
their
usage
like
for
their
needs.
The
tracing
apis.
The
tracing
of
the
nsd
cases
now
is
already
like,
is
what
they
need.
They
will
just
need
a
guarantee
that
this
will
tracing
is
working
like
wcc
and
they
would
also
need
resources
to
work.
G
C
D
Really,
like
I'm
just
thinking
about
my
other
customer
right
now,
they're
still
in
the
process
of
onboarding
to
tracing
api,
with
some
logging
alike
rapper
on
top-
and
it's
like,
we
could
have
gotten
this
done
in
november,
but
I
doubt
that
they
will
be
willing
to
try
it
right
away
because
they'll
just
ship,
the
first
part,
the
first
implementation
they'll
say:
hey
our
first
milestone
is
done,
so
it's
gonna
be
another
few
weeks
and
then
this
christmas
break
right.
D
So
I'm
just
thinking
like
I
don't
immediately
see
the
customers
to
try
it.
But
in
order
for
this
to
happen
in
january,
indeed,
I
agree
that
we
actually
have
to
plan
doing
it
this
year
like
because
then,
after
the
break,
people
can
have
what
we
have
done.
C
B
The
second
thing
would
be
to
take
a
look
at
the
actual
release
like
if
it's
a
tag
on
the
repository
or
whatever,
and
just
try
to
use
it
as
if
we're
a
client
and
say
we
think
that
this
will
be
a
good
release
mechanism
right
and
and
the
first
piece
we
totally
can
contribute
to
the
second
piece
like
from
my
perspective,
I
can
try
to
have
something
done,
but
actually
consuming
repositories
like
this
internally
is
really
expensive,
and
so
I
can
do
enough
work
to
tell
you.
B
Yes,
this
release
mechanism
will
work
for
us
or
not,
but
I
can't
actually,
I
won't
actually
be
able
to
give
you
better
feedback
than
that
likely.
B
C
D
D
I
can
volunteer
to
provide
the
method
how
you
can
consume
a
tag
with
vc
package.
So
again,
this
is
half
of
the
story.
This
doesn't
cover
blizzard,
but
I
can
definitely
contribute
that
right
up
saying
this
is
vc
package.
This
is
the
package
manager.
This
is
the
repo.
This
is
the
tag.
This
is
how
you
can
compile
hello
world
using
vc
package
install
open
challenger.
D
D
As
actually
finding
a
customer
who'd
be
willing
to
on
board
to
that,
this
is
more
like
a
dry
run
for
our
internal
use
only
and
for
illustrative
purposes.
D
Riley,
I'm
gonna
be
actually
here
in
december.
I
might
be
away
for
the
longer
weekend,
but
I'll
be
around.
C
C
Okay,
so
this
is
the
sdk
pr
that
I
like.
We
need
more
pair
of
eyes
for
this
one.
The
code
itself
looks
good
to
me.
I
just
want
to
have
more
folks
to
help
to
review
before
I
merge
that
it
hadn't
been
there
for
many
days
so
I'll
give
couple
days
before
emerge
and
if
you
have
like,
if
you
haven't
got
time,
please
take
a
look
and
for
this
one
I
think
tom
you,
you
changed
the
approach
right
now.
I
I
Yeah,
that's
right.
Even
you
know
I
added
two
two
lines
of
unit
test
to
make
sure
spam.
The
id
is
not
overwritten
by
anyone
else,
like
suggested
by
johannes.
C
Okay,
cool
max:
what's
the
status
of
this
one.
D
There's
a
comment
from
tom
that
I'd
like
to
address
about
a
parameter
for
when
we
build
for
win32
or
win64,
like
the
architecture
parameter
and
now
I'll
refresh
it
to
make
sure
that
it
passes.
I
think
there's
a
glitch
in
our
ci
that
broke
it
like
it's,
not
the
the
change,
it's
just
a
sporadic
failure
in
the
other
spot,
so
I'll
I'll
refresh
it
with
addressing
the
comment
from
tom,
I
don't
think
I
had
the
other
comments
on
that.
So,
if
you
guys
can
have
another
look,
it'd
be
great.
B
Okay,
can
I
go
back
and
ask
a
quick
question
about
the
curl
okay,
so
I
I
did.
I
did
look
at
that
and
my
question
here
was
actually
just
around
the
project
itself,
so
when,
when
it
adds
the
build
rules
for
curl,
it
looks
like
there's
an
option
to
include
curl
or
not
so
I
was
kind
of
just
asking
in
general.
Are
we
planning
to
have
curl
be
and
like
with
the
default?
B
D
So
I'm
just
thinking
like
here
about
my
customer
scenarios
on
windows
for
azure
services.
We
won't
use
it.
Obviously
we
will
use
something
else
on
linux.
This
is
more
like
when
you
build
a
test
setup
which
needs
to
push
data,
for
example
to
logstash.
D
I
guess
we'll
require
it
for
the
tests,
the
real
production
customers,
that's
a
good
question.
Most
likely,
another
alternate
stack
is
going
to
be
used
by
the
production
customers.
D
B
Yeah,
no,
my
main
question
the
reason
I'm
asking
is
again
around
bazel
right
of
if
I
set
up
basal
rules
for
that
no
standard
versus
standard
thing
is
this
another
thing
that
needs
to
be
accounted
for
in
bazel,
where
bazel
developers
can
opt
in
to
curl
or
not
right
or
like?
Is
it
a
thing
that
should
always
be
turned
on
for
linux
and
not
turned
on
for
windows
like
like?
B
I
just
wanted
to
know
generally
how
we
want
this
library
dependency
to
be
pulled
in
in
open
telemetry,
so
that,
because
there's
a
to
do
on
this
around
bazel-
and
I
was
kind
of
curious,
whether
or
not
a
that
should
even
be
implemented
or
if
this
is
a
c
make
only
option
right
or
if
it
if
it
is
implemented
in
bazel.
What
what
do
we
make?
It
look
like
that.
That
was
my
question
here.
So
it's
kind
of
around
how
we
take
independencies.
C
Okay,
so
josh
I
I
can
try
to
answer
so
number
one.
I
I
think.
Ideally,
we
want
to
keep
cmake
and
bazel
the
same
instead
of
having
an
option
in
cmake.
Only
number
two
is
I
I
I
think
in
this
case.
The
car
thing
is
not
something
that
would
require
because
number
one.
If
people
don't
use
http
or
any
like
network
interaction,
they
just
want
to
use
console
explorer.
C
They
don't
want
to
take
this
and
I
believe
the
general
philosophy
in
c
plus
plus,
is
you
only
pay
for
what
you
want
and
you
can
ask
for
it,
but
it's
not
given
to
you
by
default
and
also,
I
believe,
in
a
lot
of
production
scenario.
People
don't
want
to
take
this
dependency
because
the
core
library
might
have
actual
security
consideration.
It's
basically
a
text
surface
that
you
have
to
pay
attention
to.
So
that's
why
I
believe
it's
off
by
default.
C
I
Do
we
have
alternative
client
if
we
don't
turn
off
curl.
C
Yes,
so
in
many
in
many
complex
projects
or
in
hosting
environment
you're,
given
an
http
client,
let
let
me
give
you
a
couple
examples
like
you're
run.
Writing
some
ndk
like
application
running
on
android,
then
most
likely
you're
going
to
use
the
the
android
version
of
the
http
client
instead
of
your
own
core,
because
that
one
is
going
to
respect
the
proxy
and
other
settings
on
android
and,
if
you're,
using
some
windows
application.
C
Probably
you
just
use
the
underlying
one
http
or
like
onenote,
depending
on
the
platform,
so
most
likely
people
who
are
taking
open
time
tree.
They
already
do
something
about
interacting
with
http
server,
and
they
don't
want
us
to
give
by
using
open
time
trade.
They
don't
want
us
to
give
them
a
third-party
dependency
on
hdb
client,
but
for
people
who
don't
have
any
hdb
thing.
They
just
want
things
for
convenience.
I
think
this
is
nice
to
have.
D
Josh,
I
was
wondering
if
maybe
bazel
platforms
option
is
something
that
can
be
used
for
that,
for
example,
platforms
windows.
We
ignore
if
platforms
linux
dash,
I
know
like
local,
we
include,
but
in
other
cases
for
example,
if
you
are
talking
to
fluent
beat
fluent
d
over
local
unix
domain
socket,
then
we
don't
need
to
compile
and
include
the
client
at
all,
because
it's
all
local
kind
of
ipc
forwarding.
B
Yeah,
that's
an
option.
It's
just
the
you
can
end
up
with
a
combinatorial
explosion
across
all
options
right,
so
I
have
linux
with
this
with
this
with
this
linux,
so
it
just
kind
of
becomes
a
little
painful
to
deal
with
for
users.
I
I
guess
I
will
comment
on
the
pull
request.
That's
that's
what
I
needed
to
know
to
make
comments,
because
all
I
have
are
general
high
level
questions
are
like.
Can
you
document
how
this
will
get
used
so
I'll
make
comments
on
the
pull
request,
but
it
looked
fine.
B
C
D
To
get
the
build
those
first,
this
will
minimize
this
to
do
library,
a
change
okay,
because
a
lot
of
this
stuff's,
like
whatever
40
files
or
something,
is
the
belt
tooling
I'll
clean
up
the
build
tools.
Pr
first,
we
can
merge,
then
I'll
rebase,
this
studio,
every
change,
okay
and
the
the
the
failure
that
I
see
the
unit
test.
Failure
with
gcc
is
because
I
use
the
gcc
4.28.