►
From YouTube: 2021-10-13 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
A
So
all
other
news
here
there
was
some
aurora
that
was
visible
from
here.
I
think
monday
night
and
I
miss
it.
I
don't
know
how
I
I
was
awake.
I
just
didn't
see.
A
C
C
Yeah,
supposedly
it
reached
oregon
too,
but
every
time
I
stepped
outside
it
was
clouds.
C
I
I
saw
a
picture
of
it
from
the
columbia
river
gorge.
A
A
That's
not
the
thing
that
I
wanted
here.
We
are
so
first
thing
hotel.
Sdk
now
is
our
main.
We
have
to
update
a
bunch
of
stuff
regarding
the
readmes
and
the
files,
but
now
all
the
work
is
there
should
be
happening
there.
So
kind
of
we
perhaps
should
do
a
little
celebration.
I
don't
know
anything
but
cheers
to
us.
We
switch
it
fully
to
the
sdk.
A
You
know,
so
our
work
now
should
be
kind
of
going
through
that
direction
I
will
join.
I
will
join
the
the
tadas
and
the
party
and,
let's
celebrate.
A
I
would
like
to
have
a
discussion
about
the
error
handling,
so
we
could
set
up
at
least
have
a
decision
for
that.
I
know
that
robert
opened
issue
with
the
spec.
I
don't
think
there
was
last
time
that
I
checked
yesterday.
I
didn't
see
update
on
the
speccy
issue
that
robert
opened.
A
But
I
think
basically,
I
I
tend
to
favor
kind
of
failing,
faster
and
start
up.
I
know
that.
That's
not
the
approach,
for
instance,
that
upstream
use
it
and
some
libraries
on
open
telemetry
even
to
a
bit
different,
but
I
tend
to
think
that
bad
configuration
kind
of
I
don't
know
something.
That's
supposed
to
be
an
integer
and
the
user
puts
a
text.
It
should
fail.
They
start
up.
A
You
know
before
I
I
already
give
the
the
headline,
but
let's
wait
a
few
minutes
to
discuss
that
until
raj
you
can
join
so
then
we
can
kind
of
try
to
if
not
solve
at
least
kind
of
settle
the
questions
that
we
want
to
answer
to
decide
once
for
all
that
thing
so
not
designed
today,
but
perhaps
at
least
having
the
work
items
for
us
to
follow
up
and
finish
that
sorry,
and
I
think
then
the
the
next
thing
is
the
project
board
that
robert
did
update
and
close
a
bunch
of
issues.
A
I
think
robert.
If
you
want
to
go,
I
think
we
should
go
over
it
and
see
if
we
can
shave
even
more
items
that
are
there
yet.
You
know
so
we
have
a
very
lean
and
kind
of
really
matching
the
the
work
that
we
plan
to
do
all
right.
Let
me
stop
sharing
and
if
you
can
share
and
go
over
it.
B
B
B
I
think
two
or
three
one
is
this:
remove
big
glow
purple
giddy
repository
and
the
second,
I
think,
is
this
hard
coded
name
from
the
output
file,
because
right
now
we
are
not
supporting
mac
os
and
then
the
third
one
I
think
was
regarding
this
code
ql
workflow
scanning,
because
I
thought
that
simply
it
might
be
important
in
the
future,
but
I
don't
think
it.
I
think
it
will
just
collateral
board
here
any
comments
here
or
is
it
okay
for
you.
A
I
I'm
just
thinking
about
the
blobs.
I
know
that
is
a
kind
of
personal
piv
for
you,
but
I
kind
of
can
live
with
those
blobs
there
on
the
ripple
for
some
time,
but
also.
A
Yeah,
I
I
I
think
we
have
to
rewrite
the
the
history
right
right.
If,
if
people
want
to
get
rid
in
their
own
forks,
then
they
need
to
create
new
forks
right,
correct
yeah.
So,
from
my
perspective,
I'm
fine
living
with
the
blobs.
A
If
somebody
wants
really
to
remove
them,
I
I
don't
have
anything
against,
but
on
the
other
hand,
I
kind
of
do
you
really
want
to
do
it.
B
B
I
think
I
I
have
also
I'm
not
sure
if
that
that
one's
yeah,
I
have
not
put
this
support
for
monitoring
sharepoint.
I
don't
know
chris
if
you
want
to
put
it
on
the
board
or
not
really
so.
C
That's
something
that
we
might
eventually
want
to
get
to
as
far
as
testing
and
validation
goes,
but
I
don't
think
it's
required
for
an
alpha
or
beta
necessarily.
C
B
B
So
this
is
the
ones
that
I
decided
to
keep
like
before
beta
release.
So
basically,
here
are
the
items
about
documentation
which
I
think
are
kind
of
important.
During
poc,
initially
we
haven't
created
any
integration
tests.
We
have
started
adding
them
later,
but
still,
I
think
there
are
some
samples
that
are
do
not
have
any
integration
tests.
I
think
even
these
are
the
basic
integration
tests
that
simply
are,
for
example,
creating
a
span
using
sdk
rasmus.
Correct
me.
If
I'm
wrong,
because
I
may
be
misaligned
here.
B
B
A
Yeah,
so
so
probably
we
need
kind
of
to
put
this
in
in
a
list
right,
so
we
need
to
to
build
the
artifacts
isolate
back
then
how
we
are
gonna,
make
them
available
to
people
to
try.
You
know
I
think,
from
the
very
very
beginning,
just
have
links
to
the
artifacts
in
github.
A
Is
it's
fine,
but
as
you
move
forward,
we
need
something
better,
but
I
think
for
for
now
just
producing
the
artifacts
and
have
needs
that
we
can
perhaps
put
in
the
readme
or
something
for
people
to
give
a
try.
You
know
exactly.
B
The
first
one
was
is
about
just
updating
use
hmd,
I
think
we
are.
We
have
some
misalignment
there.
I
think
there
are
some
environmental
vibers
from
from
native
code,
which
are
not
listed.
Maybe
some
should
be
removed,
so
it's
about
double
checking.
B
This
error
handling
is
something
that
we'll
probably
discuss
later
regi
is
already
here,
but
we
will
discuss
it
later.
Also,
during
previous
meeting
we
were
dis.
I
think
that
zach
was
describing
this
directory,
how
they
are
used
and
based
on
discussion
and
how
I
checked
the
code.
I
believe
that
only
two
directories
are
needed
right
now.
These
two,
the
the
rest
of
ones,
could
be
removed.
D
Definitely
net
four
five
is
not
needed,
and
this.
B
D
We
can
test
net
standard
if
we're
going
to
be
definitely
running
on
net
core
app,
3.1
or
higher,
then
I
don't
believe
that
standard
2
is
going
to
be
necessary.
F
F
B
B
D
D
C
Yeah
the
one
question
that
I
have
where
we're
currently
setting
the
flag
to
allow
communicating
over
grpc
without
using
tls.
I
I'm
wondering
if
we
need
that
sooner
rather
than
later,
just
so.
B
I
think
there's
also
we
should
create
an
issue
like
a
general
one
about
the
environmental
variables
which
are
needed.
We
have
created
internally
jira
some
list
of
basically
things
that
should
be
supported,
like
this
hotel
tray,
hotel,
tray
trace
exporter,
blah
blah
blah.
So
this,
like
this
set
of
basically
advertising
variables
that
should
be
supported
for
life
for
mvp
to
select
you
know,
for
example,
the
otp
exporter
or
jager
exporter
or
zip
exporter,
etc.
B
A
So
we
have
the
environment
variables,
but
we
talked
about
some
time
ago
having
things
like
turning
on
and
off
exporters,
for
instance,
this
we
don't
have
yet
right
or.
A
B
A
Okay
and
they
stuff
that
yeah,
basically
that's
what
we
need
to
do
to
track
this
stuff
like
propagators
and
things,
because
we
added
that
humane,
but
now
that
we
are
using
the
sdk,
I
don't
know
if
the
environments
variable
are
already
available
for
us
to
to
to
use
it.
A
Yeah
correct,
okay,
one
one
question
related
to
that:
we
need
to
update
the
internal
build
of
the
sdk
right.
When
was
the
last
time
that
we
did,
that.
A
Yeah,
we,
I
think,
I
think
the
first
one
will
be
to
catch
up
to
hotel
sdk,
whatever
it
is,
and
they
are
probably
getting
ready
to
publish
something,
but
I
think
they
published
something
very
recently.
A
A
Yeah,
so
we
should
do
one
update,
because
we
are
red.
Pickup
can
pick
up
some
of
the
environment
variables
and
then
we
have
a
more
accurate
picture
about
when
when
you
are
gonna.
What
is
this
still
missing
regarding
environment
variables,
any
volunteers
you
do
that.
A
Okay,
I
I
will
try
to
do
this
update
during
this
week.
So,
let's,
let's
see
where
we
are
in
that
regard,
our
update
to
whatever
they
released
last
week
and
then
we
we
check
the
environment
variables
I'll,
try
to
use
and
and
see
how
how
it
goes.
Do
you.
A
One
thing
that,
but
this
is
not
for
for
the
beta-
is
more
for
hours
in
internals
and
kind
of
help
with
people
getting
to
the
project.
A
And
there
is
documentation,
read
that
tony
wrote.
I
remember
looking
when
we
are
switching
from
call
target
to
calcite,
but
we
should
kind
of
document
how
the
things
work.
You
know
I
started
that
overview,
but
we
should
keep
going
deeper
on
that
not
required
for
the
beta,
but
we
should
keep
going
on
that,
so
people
can
understand
when
they
arrive
at
the
repo
how
things
work
you
know.
A
So
I,
as
I
said
not
not
for
for
this
beta
but
later
we
need
to
do
that.
A
All
right,
so
I
think
I
invite
people
to
kind
of
let's
try
to
use
the
board
now
to
guide
us,
and
I
think
we
can
move
the
discussion
about
the
error
handling.
B
Okay,
so
basically
raj
thanks
a
lot
for
like
starting
the
discussion
here.
I
I'm
not
you.
You
mentioned
that
trusk
java
instrumentation,
though
they're
not
impacted
user
application,
but
I
was
also
talking
internally
with
other,
like
developers
of
auto
java
and
they
say
they're
not
impacting
the
user,
but
in
the
runtime
they
told
me
that,
as
far
as
they
know,
if
there's
some
misconfiguration,
they
think
that
they
are
crashing
the
application.
G
Oh
so
yeah,
when
hired
a
discussion,
he
was
saying
that
we
should
never
at
least
they
the
design
principle.
What
java
auto
instrumentation
has
is.
They
should
never
crash
an
user
application
at
any
cost
and
they
should
do
like
log
only
to
the
console
like
they
have
a
console
logging
in
built
console
logging,
so
they
wanted
to
lock
it
there
and
especially
he
provided
that
the
principal
somewhere
from
their
repo.
That's
that's
a
pr
that's
being
in
progress.
So
that's
what
the
the
resilience
they
want
to
follow.
B
So,
regarding
this
pr,
I
also
wanted
to
double
check
there.
I
don't
know
if
my
comment
is
already
resolved,
so
I
asked
here
about
the
resilience
it
is
deployed
to
the
community,
about
the
values
of
open
telemetry,
and
here
is
even
the
outer
even
author,
again
said
that
he
didn't
had
in
mind
the
initia
the
initialization
time
he
was
just
thinking
about
the
runtime.
B
And
I
do
think
with
your
chance
so
based
on
it
just
to
make
sure
that
we
will
have
any
kind
of
consistency
and
ability
to
hear
from
others.
I
have
created
this
and
to
close
the
tabs.
I
have
them.
There
are
a
lot
of
them.
B
It
was
this
one,
so
I
have
created
this
issue
in
the
specification.
G
Sure
so
what
we
could
do
is
like
we
could
stick
to
whatever
we
are
doing
now.
We
are
throwing
like
an
exception
from
like
the
instrumentation
now,
so
we
should
continue
to
do
that
and
bring
the
clarity
in
the
community
on
this.
So
whatever
we
agree
on
that,
we
will
do
the
changes
later
anyways
we
have.
We
are
tracking
this
one
exactly.
A
Yeah
sounds
good
one,
one
thing
related
to
that.
I
think
erasmus
and
raj
already
mentioned
this
one
one
of
the
pr's
it.
It
does
seems
that
we
really
can
unify
the
startup
hook
with
the
loader
kind
of
right
now
it's
using
it,
but
we
can
even
go
one
step
further
and
just
have
one
code,
I'm
not
because
they
hold.
This
is
small.
I
don't
think
that
is
something
that
we
should
kind
of
just
oh,
let's
we
need
to
do
it,
but
it's
something
that
we
we
could
look
in
the
near
future.
A
You
know
just
because
having
just
a
single
piece
then
makes
things
I
think
easier,
but
it's
a
it's
a
side
comment
on
that,
so
we
are
agreeing
that
for
now,
let's
throw
exception
on
initialization,
let's
be
noisy
in
this
case
in
case
of
misconfiguration.
A
I
think,
especially
now
that
we
are
trying
this
beta.
I
want
to
avoid
that
case
of
kind
of
hey.
There
is
no
data
and
then
asking
people
to
go
to
the
logs
and
files
is
kind
of
not
ideal.
You
know
so
if,
if
we
can
give
kind
of
clear
error
messages,
especially
on
this
beta,
I
think
I
think
you'll
be
you'll,
be
more
we'll
get
more
feedback
in.
In
that
case,
you
know.
A
B
That,
basically,
as
part
of
this
issue,
we'll
just
create
some
some
little,
I
don't
know
sentence
about
how
we
are
handling
errors
in
our,
for
example,
contributing
docs
or
some
design
docs.
Basically,
maybe
some
maybe
a
figure
out
somewhere
and
once
we
merge
it,
we
can
close
it,
and
I
will
just
try
to
follow
up
on
the
specification
of
these
proposals
as
a
basic.
It's
a
separate
issue
that
I'm
basically
working
on.
A
Yeah
we
have
that
document
with
the
high-level
goals
right
yeah.
I
think
in
some
place
in
that
dark,
perhaps
after
that
high
level
session
rules.
We
we
add
that
as
our
current
behavior
and
we
describe
and
even
perhaps
link
it
makes.
B
A
B
C
B
C
To
I
was
gonna
say
robert:
have
you
gone
to
the
specifications
sig
meeting
to
kick
start
this
discussion
by
any
chance?
No,
you
want
to
do
it
chris,
I
gotta
brush
up.
I
just
know
that
at
least
for
some
of
the
specification
gingers,
my
teammates
have
have
done.
They've
they've
only
been
getting
traction
during
the
sig
meeting
and
outside
of
the
sick
meeting.
They've
gotten
almost
no
traction.
B
A
A
Yeah
no,
but
I
I
if
you
think
that
as
the
project
as
a
as
a
whole,
you
want
to
kind
of
unify
as
much
as
possible.
You
know
it's
not
never
going
to
be
100
possible
with
all
these
languages
and
everything,
but
the
closer
that
you
can
stay
the
better.
I
I
I
think
for
us
doing.
Uh.Net,
of
course,
is
java.
That
is
kind
of
the
the
behavior
that
if
people
see
java,
they
expect
the
same.net
and
vice
versa.
You
know,
but
there
are
differences,
but.
A
Raj,
do
you
have
any
next
steps
planned
because
we're
talking
about
about
the
rapper
that
is
kind
of,
I
think
in
my
mind,
the
next
step
regarding
the
startup
hoop.
Let
me
check.
G
Yes,
brother,
so
the
currently,
what
I'm
doing
is
I'm
working
on
the
integration
test
to
finish
it
off
like
so
this
kind
of,
like
whatever
the
implementation
we
have
in
this
project,
is
a
kind
of
my
new
learning
curve
for
me.
So
just
taking
a
look
and
trying
to
finish
off
that
this
week,
so
probably
from
monday
onwards,
my
plan
is
to
start
on
the
the
diagnostic
sources,
which
I
I
provided
a
small
demo
a
few
weeks
back
on
that
so
create
an
issue
and
start
the
work
on
it.
G
I
might
have
one
of
my
other
colleague
michael
also
joined
with
me
here
shortly
to
provide
like
like
implementation
on
it.
G
Yeah
one
more
question
like
ram
was
last
time
he
bought
me
on
that.
Pr
is
about
like
we
having
both
like
the
like,
clr
managed
library
is
also
built
as
native.
He
got
a
very
good
point
at
that,
and
I
also
like
whenever
I
build
something
I
just
take
a
look
into
the
module
list
and
everything.
G
So
I
I
kind
of
see
like
it's,
both
native
and
managed
layer
stays
there
and
optimization
and
all
is
being
turned
off
for
our
libraries,
even
when
we
do
a
startup
hook,
because
the
build
earlier
it's
done
with
the
native
one.
So
there
is
some
change
required
there,
or
probably
I'll,
be
doing
a
little
more
research
in
that
area
to
see
how
to
build
only
the
manage.
G
Maybe
we
I
don't
know
whether
we
should
plan
to
have
two
nugets,
or
only
with
one
new
gate
with
different
parts,
because
I
see
the
optimization
is
turned
off,
even
if
we
don't
need
a
byte
code
instrumentation
in
that
case,.
A
This
is
just
kind
of
first
thought
because
you
mentioned,
I
think
we
need
to
pack
this
separately
because
they
start
up
who
can
be
much
smaller
in
that
sense
and
kind
of
relatively
easier
to
deploy.
You
know
if
you
think
you
windows,
kind
of
the
the
profiler
basically
requires
install.
A
G
I
also
had
a
similar
thought
process
like
you,
what
you
were
explaining,
but
I
I
think
this
is
like
we
need
to
like
think
towards
it
and
then
take
a
call.
C
G
F
G
That
time
not
check
that,
but
if
provider
is
starting,
why
do
we
need
to
worry
about
like
we
have
already
a
profiler?
We
don't
need
to
worry
about
startup
book,
setting
up
that
environment
variable
from
there
right.
F
I'm
was
thinking
that,
like,
if
profile
is
setting
the
startup
environment
variables,
then
it's
just
simple
out
of
the
box
setup.
So
user
doesn't
need
to
care
about
this
variable.
A
That
that
that's
dynamic
research.
F
A
But
we
still
need
the
profiler
to
be
able
to
to
call
and
load
the
assembly
right,
so
we
are
going
to.
Basically,
you
are
saying
we
are
going
to
do
by
the
path
right.
A
A
F
Now
we
go
and
you
so
basically
you
mean
the
only
totnet
fx
way
so
that
that's
the
only
exception
that
needs
embedded
loader.
F
A
Yeah,
basically,
what
I'm
saying
is
that,
because
the
native
profiler
has
the
resource,
it
doesn't
need
to
know
anything
about
the
paths
at
that
stage.
A
But
if
we
switch
to
kind
of
using
an
assembly
outside
the
profiler,
then
we
need
to
be
very
careful
about
location.
I
think
for
net
core
in
general.
That's
not
a
problem
where
you're
gonna
load
the
assembly,
but
I
think
for
a
framework
that
could
be
a
big
trick.
A
A
F
F
A
problem
you
know
yeah,
so
basically,
if
you're
staying
with
the
same
folder
structure,
then
it
will
be
like
really
easy,
so
you
need
to
set
up
anyway,
the
environment
variable.
That's
stating
the
full
bet.
A
Yeah,
I
think
we
need
to
test
this
with
the
framework.
To
be
sure,
I
I,
as
I
said,
I
think,
with
dot
net
core
is
not
a
problem
at
all,
but
I
I
always
get
a
little
bit
worried
about
dotnet
framework,
because
I
think
the
rules
for
that
are
a
bit
different.
F
So
yeah
the
framework
one
is
tricky
so.
F
There
are
yeah
two
things
I
seems
like
it's
easy
to
replace
with
with
the
bat
and
it's
in
a
core
profiler
get
assembly
and
symbols.
Byte
is
the
method
that
does
all
the
things
and
if
it's
not
like
easy
to
replace
with
the
external
dllt-
and
we
can
just
chop
it
up
to
two
native
dlls
and
in
case
of
tottenham
framework,
then
you
load
embedded
loader,
that's
embedded
to
the
another
native
gll.
F
A
So
basically,
what
you
are
saying
is
kind
of
okay.
We
remove
the
the
loader
from
that,
not
that
dotnet
core
in
general
and
the
windows
build
of
the
native
has
it
to
be
able
to
load
in
the
framework.
F
A
I
see
just
a
general
feeling:
do
you
guys
think
this
is
something
that
we
should
try
right
now
or
something
that
we
should
postpone
for
kind
of
after
we
have
this
beta.
B
I
can
say
from
myself
that
yesterday
I
started
to
think
if
basically.net
framework
is
an
important
like
framework
for
us,
given
that
we
are
already
saying
that
we
are
supporting
only
the
sdk,
which
is
like
four
four
four
six
one
which
already
says
that
we
are
kind
of
not
handling
quite
a
good
legacy
applications,
and
I
started
to
think
that
maybe-
and
basically
you
know
in
dot
net
core,
you
have
this
hooks
and
other
stuff
and
maybe
to
also
simplify
a
lot
of
stuff.
B
A
Yeah
but
but
yeah
yes,
we
still
would
bite
cold
instrumentation.
That
is
only
via
the
profiler
right,
not
only
if
you're
the
profiler,
but
we
are
not
talking
about
alternatives
to
that.
At
this
point
you
know
kind
of
yeah.
We
could
use
some,
I
would
say,
for
instance,
cold
coverage.
We
use
cover,
let
right
cover.
Let
does
at
build
time
the
rewriting
but
you're.
A
Think
we
we
for
now
and
as
that
old
conversation,
I
think
at
least
five
years.
We
still
need
bite
code,
instrumentation,
also
the
stuff
that
greg
mentioned
about
other
capabilities
that
the
cll
profiler
brings
profiling.
That
I
think
open
telemetry
has
been
mentioned
in
this.
That
have
been
conversations,
so
I
think
we
still
have
to
keep
the
profiler
for
quite
some
time
or
maybe,
depending
on
the
the
direction
that
open
telemetry
takes.
A
F
Yeah
and
it's
not
just
a
profiler,
so
basically
it's
a
domain,
natural
loading
and
everything
like
this.
There
david
wants
to
speak.
A
E
So
go
ahead,
yeah
I
was
just
so
now
we've
kind
of
moved
on,
but
they
I
looked
at
the
code
and
it
should
be
possible
to
set
the
startup
hook
from
the
profiler
yeah.
I
mean
I
don't
want
to
say
100
sure
without
actually
trying
it,
but
it
looks,
but
I'm
pretty
sure
it's
possible
because
the
all
the
startup
stuff
isn't
managed
code.
A
A
Okay,
I
think
I
think
that
that's
good
for
us.
We
are
not
rushing
for
that.
I
will
say
that
related
to
the
question
that
I
was
saying.
My
personal
take
is
that
we
don't
do
this
suggestion
that
erasmus
was
talking
about
right
now.
So
I
think,
having
the
answer
in
the
time
frame
that
you
can
help
us
with.
That
would
be
great.
Then,
in
the
future
we
can
do
the
change.
I
think
for
this
beta,
we
kind
of
don't
shoot
for
that.
A
All
right,
we
in
turn,
you
still
have
15
minutes,
but
anyone
wants
to
bring
something
up.