►
From YouTube: 2021-12-08 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).
B
B
I
don't
I
don't
know
if
you
are,
I'm
not
sure
if
I'm
correct
meeting
is
this.
C
B
A
B
F
Yeah
my
wife
likes
to
rotate
the
the
decor.
F
Yeah,
she
finally
brought
her
attention
to
to
my
little
corner
of
the
room.
H
Yeah
I
I
I
would
like
to
I
I
put
the
background
for
a
meeting
and
then
I
forgot,
I
never
remove
it,
but
actually
it's
good.
I
don't
need
to
clean
up
anything.
H
E
I
I
can't
give
a
title
right
to
the
meeting.
Anyway:
apologies.
We
are
php
we'll
leave
and
find
the
correct
link.
H
Okay,
I
think
I
think
for
my
my
time
here
we
are
good
a
few
new
issues.
I
want
to
separate
this
last
one
for
us
to
talk
later.
I
think
it's
a
good
suggestion,
but
let's
discuss
that
separately.
H
There
is
no
reason
for
us
to
keep
supporting
this.
We
didn't
do
a
release
before
so.
That's
removing
use
the
the
stuff
from
resources
and.
H
G
G
D
B
H
D
C
H
H
F
Yeah
raj
has
has
a
good
comment
on
there
and
zach.
I'm
curious
if
you
can
take
a
look
too,
because
I
recommended
the
events
that
new
relic
subscribes
to
in
order
to
flush
everything
when
it
can
and
so
would
be
interesting
to
see
what
what
theta
dog
does
too.
H
So
we
we
did
a
bunch
of
reviews
and
stuff,
so
this
one
is
the
one
that
we
just
discussed
it.
This
is
the
document
one
we
okay.
I
think
this
is
just
a
matter
of
reviewing
right
so
and
we
keep
getting
this.
I
think
we
should
do
either
an
update
or
close
this,
because
right
now
they
just
pollute
our
our
views.
Here
we
are
not
actually.
B
First,
I
was
looking
into
creating
more
integration
tests
for
other
other
instrumentations
or
even
for
our
existing
samples,
and
then
I
will
feel
more
stable
to
make
some
bumping
to
avoid
problems
that
we
did
had,
for
example,
astronaut.
I
think
integration
before.
H
So
did
you
already
started
on
that
because
then
perhaps
we
do
these
upgrades
afterwards?
No!
No!
I
have
okay.
B
B
C
B
H
I
don't
know
I
I
to
be
fair,
I'm
I'm
gonna
be
on
vacation
until
next
year.
I
start
today.
I
don't
know
that
I'll
have
chance
to
look
at
this.
You
know
so.
F
E
F
H
I
think
I
think
it
let's
merge
as
it
is.
You
know
this
is
just
more
of
kind
of
a
question
about
kind
of
improving
and
best
practice.
I
I
was
just
worried
about
this
sketch
kind
of
hiding
stuff,
because
I
don't
see
the
exception
in
the
normal
path.
You
know
so
that
that's
why
I
kind
of
approve
and
kind
of
I.
I
don't
think
we
need
to
do
this.
We
can
look
after
it
so
feel
free
to
merge
right
after
the
meeting.
H
Okay,
so-
and
we
all
already
know
about
the
this,
one
is
basically
the
skeleton
waiting,
the
other,
the
other
one.
So
I
think
then,
the
the
most
interesting
thing
here
is
chris
proposal.
Here
I
think
chris,
you
are
the
best
one
to
go
over
it
and
tell
exactly
what
you're
proposing.
F
So,
in
short,
there's
a
few
things
pushing
me
in
this
direction.
One
the
startup
hook
allows
us
to
take
advantage
of
the
performance
improvements
in
the
runtime.
When
you
don't
need
the
bytecode
instrumentation.
F
Two
we've
got
the
timing
issues
with
bootstrapping
the
sdk.
If
we
use
the
the
profiler
the
way
we
are,
it
seems
too
early
in
the
process
and
we're
getting
errors.
But
if
we
use
the
startup
hook,
the
dependencies
are
there
they're
loaded
and
things
are
working,
and
so
my
recommendation
is
going
forward.
For.Net,
core
and
there's.
B
F
My
recommendation
is
going
forward
eventually,
there'll
be
less
and
less
need
for
bytecode
instrumentation,
at
least
that's
my
theory,
especially
for
dot
net
core
and
net
five
plus
we'll
see
more
and
more
source
code,
instrumentation
or
at
least
more
instrumentation
packages
supported
by
the
sdk
itself,
and
so
I
think
also
I'm
recommending
a
go
forward
strategy
for
net
core
to
separately
or
to
just
rely
on
the
startup
hook,
but
also
allow
the
optional
enabling
of
the
profiler
as
as
well
for
the
cases
where
bytecode
instrumentation
is
desired
and
in
practice
I
I
want
us
to
document
setting
up
both
sets
of
environment
variables,
as
opposed
to
having
the
profiler
during
initialization
inject
the
startup
hook,
environment
variables,.
H
Yeah,
I
I
agree,
I
think
it's
a
a
good
path.
One
thing
that
I
have
in
my
mind,
so
they
start
up
hook
for
known
net
frameworks,
as,
as
you
are
saying
there
for
the
framework
we
keep
the
profiling
and
injected
via
that.
H
So
I
kind
of
just
just
to
mention-
and
this
is
something
that
I
mentioned
to
chris-
even
there
are
some
apis
exposed
by
net
monitoring
that
actually
allows
to
from
manage
code
to
attach
a
profiler.
H
This
is
something
that
we
could
even
explore
in
the
future
kind
of
having
the
startup
hook
and
injecting
the
profile
according
to
configuration.
You
know,
I
think
this
long-term
perhaps
is
something
that
looks
more
flexible
and
easier
for
the
users,
but
I
I
I
think
right
now
we'll
be
kind
of
adding
just
another
path
that
we
don't
need
to
go
deeper
now,
but
I
do
think
it's
something
interesting.
So
I
I
really
like
the
proposal
chris
had
shown
me
kind
of
doing
this
from
the
profiler.
H
Then
I
I
I
get
very
kind
of
animated
about
that
prospect
also
makes
makes
things
kind
of
between
the
framework.
If
you
load
from
the
profiler,
you
call
the
startup
hook.
H
For.Net
framework
in
that
sense
gets
very
similar
code
without
us
be
trying
to
do
something
kind
of
super
different
for
the
code
when
it's
dotnet
framework.
So
I
really
like
that.
B
Just
have
one
question
because,
like
I'm
not
a
subject
matter
expert
here,
but
I
remember
that
even
raj
had
some
idol
concerns
because
I
do
not
remember
right
now:
there's
one
possibility
that
you
can
build
like
the
application
together
with
net.
I
don't
know
what
was
like
self-contained
application
or
something
like
that
was-
and
I
remember
something
was
not
working
for
this
scenario.
I
don't
know
it
was
hooks
or
profiler.
F
Chris,
I
don't
remember
what
the
specific
thing
was.
I
know
that
for
at
least
at
the
additional
dependencies
loading
that
depended
on
how
the
application
was
published,
but
yeah
there's
raj.
Maybe
he
can.
A
So
startup
hook
will
work
everywhere,
so
there
is
an
additional
dependencies.
Whatever
we
had
runtime
store,
that
is
not
supported
by
the
for
self-contained
apps,
which
is
not,
which
is
not
supported,
could
repeat,
it
won't
work
for
self-contained
apps.
The
recent
work
which
we
did
for
additional
dependencies
and
the
runtime
store
that
part
will
not
work
for
self-contained.
B
A
Yeah,
that
is
the
reason
I
have
my
other
pr,
because
we
we
cannot
expect
user
to
go
through
everything.
That's
the
reason
we
have
that
the
diagnostic
source
proxy,
to
avoid
those
kind
of
issues.
If
additional
depths
does
not
work,
we
should
fall
back
to
some
other
mechanism.
A
We
track
the
activity
which
is
created
by
the
lower
version
of
diagnostic
source
and
convert
into
the
higher
version
and
have
it
some
other
means
like
we
don't
lose
that
activity
or
throw
the
user
right
away
there.
So
that's
the
reason
for
the
other
pr
we
should
address
in
in
that
way.
Instead
of
saying
putting
a
note
saying
that
hey,
we
don't
support
it,
that
that
won't
be
a
right
thing
to
do.
I
believe.
H
Yeah,
so,
basically
is
that
for
self-contained,
the
story
is
very
different
and
that
kind
of
puts
us
in
in
the
need
for
the
procs
yeah.
I
I
I
keep
going
back
when
whenever
we
mentioned
that
and
because
of
where
we
are,
I
think
that
in
practice
I
did
encounter
many
self-contained
applications
that
people
want
to
instrument,
because
typically
they
are
running
a
server
and
they
are
building
on
top
of
or
either
images
or
vms,
that
have
their
own
time.
So
people
are
not
concerned.
A
H
The
self-contained
so
perhaps
kind
of
hammering
the
point
that
I've
been
making
a
few
times
is
that
I
think
we
should
look
and
be
concerned
about
that.
But
I
don't
think.
For
instance,
we
are
trying
to
kind
of
move
in
next
year
to
try
to
have
this
beta
and
we
should
not
hold
because
of
that.
H
B
That's
the
reason
why
I
suggested
only
to
just
write
it
down
in
documentation,
so
it's
it's
visible.
We
can
even
create
an
issue
to
add
support
to
it
and
mention
intoxicant
just
to
avoid
you
know,
problematics
or
I
don't
know
someone
even
asking
on
slack.
What's
going
on
or
creating
issues,
just
removing
some.
You
know
potential
questions.
If
we
have
a
release,
yeah.
H
C
A
I
have
some
questions.
I
think
we
just
went
a
little
off
the
track
just
to
understand
whatever
the
conversation
that
we
have
been
having
with
chris.
This
is
a
good
like
a
proposal
what
he
has
very
good
proposal
along
with
that
you
were
also
speaking
about
like
changing
the
order
of
the
load
right
to
the
between
the
profiler
and
the
startup.
Now
profiler
comes
in
and
load
startup
hook.
A
You
are
saying
there
is
an
option
where
we
can
use
a
startup
fork
to
load
the
profiler
you
know:
do
you
think
we
should
get
anyone
from
like
the.net
to
explore
that
options?
So
if
so,
let
me
know-
because
I
don't
have
expertise
in
that
area.
We
can
any
time
bring
someone
from
dot
net
team
directly
here
to
explore
those
stuffs,
if
need
be,.
H
Yeah,
I
think
I
think
that
is
a
interesting
possibility.
I
say
that's
too
early
to
bring
somebody
here,
because
we
don't
have
kind
of
a
concrete
kind
of
path
that
we
want
to
pursue
there,
but
I
I
think
it's
a
possibility,
because
I
was
looking
to
to
decoding
the
dot
net
monitor
tool
and
they
have
this
possibility
of
attaching
the
profiler
and
it
immediately
occurred
to
me
because
then
our
configuration
gets
much
simpler.
H
You
have
this
startup
hook
and
if
the
configuration
that
the
user
is
is
using
kind
of
say,
hey,
I
have
bytecode
instrumentation,
for
instance,
then
we
load
use
that
api
to
load
the
profiler.
You
know,
I
think
it's
too
early
for
us
to
involve
some
somebody
from
the
runtime
time
on
that.
I
think
we
should
play
a
little
bit
and
kind
of
mature
this
idea
before
you
get
that,
but
I
do
think
it's
something
that
we
want
to
look.
H
H
There
are
other
scenarios
that
we
also
want
that
ability
of
rigid
and
also
providing
in
general,
but
one
thing
that
I
think
for
the
project
in
general
because
of
we
use
these
things,
we
should
be
trying
to
look
of
the
stuff.
That's
provided
by
their
own
time.
So
hope
telemetry
has
is
not
part.
Right
now
is
not
going
to
be
something
that
the
other
sigs
are
looking,
but
there
is
talks
about
collecting
profile
information
via.
A
H
And
then
I
think,
then,
that
natural
approach
will
be
to
look
what's
provided
by
the
runtime
and
the
dotnetmonitor.net
trace
tools
use
you
know,
but
I
think
greg
has
a
point,
that
we
have
to
keep
the
mind
open
that
the
clr
profiler
right
now.
We
think
about
byte
code
instrumentation,
but
it
can
be
used
for
other
things.
You
know
so
for
one
side
I
feel
very
kind
of
tempted
to
take
this
path
and
the
clr
profile
disappears
completely,
but
I
still
have
kind
of
besides
the
bytecode
instrumentation.
B
There
was
I
I
know
I
paul
was
talking
a
few
times
with
you,
but
maybe
it's
worth
sharing
also
with
others.
B
Some
time
ago
I
had
an
idea
to
replace
the
bytecode
instrumentation
instead
of
using
dotnet
profiler
using
just
ilv
writing
during
build
time
using
cecil
or
foddy
or
whatever
just
creating
some.
You
know
something
that
can
cook
to
ms
build
and
can
be
used
to
replace.
I
don't
know
if
someone
was
investigating
if
it's
possible,
but
so
far
I
heard
that
most
of
the
customers
have
possibility
to
just
cook
to
the
build,
or
at
least
at
splunk.
We
didn't
have
so
far
any
issue
that
they
built
couldn't
be
changed.
B
A
I
spent
some
time
in
those
areas
also,
the
biggest
challenge
in
that
area
is
like
maintaining
our
code.
Actually,
if
we
cannot
have
a
generic
solution,
if
the
library
changes
their
version
and
if
there
is
a
slight
change,
whatever
the
implementation
we
have
might.
A
We
need
we
may
need
to
write
a
lot
of
custom
code
here
being
a
clr
profiler,
it's
a
little
easier
to
manage
when
compared
to
the
other
things,
but
I
was
also
when
I
started
my
work.
I
was
more
inclined
towards
the
il
rewrite
than
using
a
profiler,
because
it
is
easy
and
we
have
a
lot
of
control
at
the
managed
layer,
but
it
is
what
I
would
say
to
try
it
out
and
check
it.
A
A
We
had
a
talk
with
like
the
dotnet
team.
They
were
not
more
inclined
towards
that.
They
pushed
me
towards
the
profiler
direction.
If,
if
we
want
to
get
that
done
more,
they
push
me
towards
clr
md,
not
even
the
way
what
we
use
it
here,
so
they
they
never
recommend
using
an
ilv
right.
That's
why
it
was
not
never
documented
anywhere
by
the
net
team.
That
part.
H
Yeah,
I
I
I
think
that
we
mentioned
this
also
some
time
ago
and
to
repeat
myself
is
what
we
consider
the
devops
scenario,
that
there
is
no
build
at
all.
You
know.
So
when
there
is
no
build
at
all,
then
we
have
to
consider
that
the
clr
profile
is
the
path
that
the
runtime
recommends.
You
know
so
yeah.
So
once
more,
I
think
sometimes
it's
very
tempting
to
teach
the
profiler,
but
I
think
in
practice
we
still
have
very
good
reasons
to
keep
it
so.
H
Okay,
I
so,
I
think
chris,
I
think
we
have
a
broader
agreement
to
your
proposal.
I
think
nowadays
is
starting
kind
of
trying
to
move
that
ahead.
I
don't
know
how
folks
are
planning
for
the
end
of
the
year
and
next
year,
so
I
don't
know
what
to
expect,
but
I
think
it's
a
matter
of
coordinating
kind
of
hey.
I'm
gonna
work
on
implementing
this
part
and
bring
this
and.
B
I
have
one
question
because
I
prefer
to
know
your
preferences
before
I
do
so,
if
I
will
have
time
to
work,
create
more
integration
tests,
we
have
seen
some
patterns
and
I'm
not
sure
if
what
is
our
preference
like
for,
like
you
know
the
architecture
or
even
the
so
the
the
solution.
B
How
should
like,
because
chris
show
that
this
approach,
that
you
have
one
application,
basically
which
can
be
used
to
in
a
lot
of
tests
right
in
right
now,
we
have
like
there
was
also
there's
right
now
in
data
docket,
also
in
current
code
base.
There's
this
notion
of
test
applications
in
another
folder.
B
There
was
also
an
idea
to
to
have
basically
this
sample
application
test
closer
to
each
other,
just
to
have
it
closer
together,
and
I
don't
know
if
we
have
any
if
we
had
any
conclusions
there,
because
simply
I
do
not
want
to
land
into
situation.
If
we
have
some
ideas
right
now,
it
will
be
better
to
to
discuss
it
right
now.
If
we
do
not
know,
then
it's
also
okay,
but
maybe
there's
already
some
strong
opinion
that
I
can
change
something.
G
H
Yeah,
I
I
also
got
the
same.
You
know,
so
I
think
we
have
to
use
this
moment
kind
of
to
experiment
with
something.
So
if
you
have
an
idea-
and
you
want
to
implement
the-
I
know
that
it
can
mean
that
later
you
will
have
to
do
some
real
work,
but
if
you
have
some
idea
that
you
want
to
put
forward,
I
think
putting
in
a
pr
and
and
the
test
is
the
best
way
for
people
to
kind
of
evaluate
that
you
know
sure
us
deciding
here.
It's
gonna
be
kind
of
very
abstract.
H
You
know
so
having
a
pr
is,
I
think
the
best
best
way
it
works.
H
All
right,
so
I
was
saying
about
kind
of
let's
follow
up
with
chris
proposal.
I
think
it's
good
and
I
just
don't
think
at
least
for
me
personally,
that
I'm
gonna
be
doing
any
more
work
this
year
or
on
this
area
or
anyone.
H
But
I
I
think
it's
a
matter
of
coordination,
I
think,
can
be
done
in
few
pies
or
even
just
one,
I'm
not
sure,
but
coordinating
between
us
to
be
sure
that
we
do
that
work.
F
F
So
if
I
don't,
if
you
don't
see
it
done
by
next
tuesday,
then
it's
up
for
grabs.
H
Okay,
by
the
way,
I
think
good
moment
to
talk
about
the
holidays
upcoming
day,
so
I'm
going
on
vacation
after
today,
so
I
will
just
be
back
to
see
meetings
and
this
stuff
next
year.
I
think
you
already
discussed
this
that
this
is
the
last
meeting
of
this
year
that
we
schedule.
H
H
If
there
is
something
I
can
check
is
like
here
and
there
if
people
need
some
of
the
maintenance,
for
instance,
to
change
something
the
ripple.
I
can
check
this
like
here
and
there
so
bring
us
there
if
we
need
to
close
stuff
or
do
something
I
think
from
the
maintainers
here
me
zach
and
chris
are
the
ones,
and
I
think
the
two
of
us
are
taking
back
a
shower
taking
vacation
to
zach.
G
Yeah
I'm
taking
a
short
vacation.
So
just
like
christmas
to
new
year's.
H
E
F
There
was
a
question
about
environment
variable
naming
and
robert.
Thank
you
for
looking
into
what
java
did
as
far
as
naming
goes,
and
so
I
think
the
difference
there
is
naming
it
dotnet
agent
versus
dotnet
auto,
and
I
don't
know
if
we
have
a
strong
preference
or
if
we
need
consistency
or
if
we
also
need
to
have
this
or
get
gather
the
opinion
from
the
other.net
sig.
H
So
this
was
for
that
environment
variable
that
had
the
dotnet
prefix
right
and
there
was
a
comment
there
to
put
dotnet
auto
and
is
java
doing
java
agent.
B
H
Oh,
we
can
do
a
I.
I
think
we
should
not
get
stolen
this.
We
can
do
a
vote.
I'm
gonna
vote
my
conscience,
I'm
gonna
vote
auto.
A
B
B
B
H
B
So
I
will
just
in
in
the
issue
I
would
just
make.
I
would
just
ask
the
dot
net
maintainers
from
the
sdk
if
they
have
any.
If
they
have
any
opinions
here.
Okay,.
H
All
right
so
from
my
side,
happy
holidays
to
everyone.
Happy
new
year
see
you
guys
next
year,
I'm
I'm
not
going
to
the
meetings
anymore
and
actually
I
removed
them
from
the
calendar
by
the
way,
all
right,
bye,
guys.