►
From YouTube: 2021-08-05 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
Good
stuff
yeah,
what's
going.
A
A
B
Except
to
say,
oh,
that
yeah,
so
since
a
sampler
can
be
passed
to
starting
a
trace.
Yes,
you
know.
C
B
A
Yeah,
I
saw
the
prs
and
I
was
like
I'm
gonna-
be
lost
on
where
this
is
located
on
for
the
instrumentation
libraries
yeah.
A
Yeah
it'll
need
to
be
a
text
map
propagator
and
then
the
problem
with
the
rabbit-
I
guess
not
a
problem,
but
a
a
thing
with
the
rabbit
properties,
is
that
the
headers
are
all
tuples
with
the
value
type
in
the
middle.
A
So
that's
why
I
was
going
back
to
looking
at
stuff
and
just
being
like.
Oh,
you
can
just
pass
in
like
just
pass
in
what
the
carrier,
so
I
guess
an
array
in
this
point
or
a
list,
and
then
it
would
take
that
just
a
blank
context
or
a
top
level
context
object,
and
then
it
could
decide
what
it
needs
from
that.
A
So,
like
the
go,
implementations
and
all
that
stuff
have
like
a
top
level
context
and
then
span
context
under
it
and
baggage
context
under
it,
and
so
it
can
just
be
chained
together
and
each
one
can
just
say
like.
I
want
the
span
part
or
I
want
the
baggage
part
so
like
if
I
was
to
write
like
an
amqp
propagator
like
it's
like
really
unclear,
like
an
easy,
simple
path
forward
to
like
here's,
a
single
propagator
and
it
just
takes
a
test
map
and
isn't
able
to
inject
the
things
it
needs
to
inject.
B
You
want
it
to
inject
baggage
and
spam
context.
B
All
right,
yeah,
I
mean
you
can
there's
two
ways
and
you
could
write
two
propagators
like
how
the
right,
http
ones
are
done
or
just
write.
One
there's
just
helpers
right
now
that
accept
the
context
key
to
build
the
propagator
for
you,
but
you
can
do
it
yourself
without
the
helpers
and
then
you
can
pull
from
the
context
manually
yourself.
You
don't
have
to
so
it
I
mean
it.
Just
has
the
the
context
like
the
go
one,
but
there
are
helpers
that,
let
you
say
this
one's
a
span
context
propagator.
B
A
That's
where
I
was
getting
lost
to
was
just
that,
like
my
understanding
was
like
100
of
this
was
meant
to
be
in
the
api
and
then
then,
like
the
lack
of
helpers
and
whatever,
like
just
like,
if
it's
past,
a
context
that
it
works
versus
like
the
it's
a
pedic
implementation
or
something,
I
guess
that's
that's
where
I
was
getting
like
super
lost,
because
I
was
like.
Oh,
we
have
tests
in
here
to
like
inject
and
extract
the
w3c
trace,
headers,
but
not
baggage,
and
then
that
led
to
like.
A
Oh,
I
guess
I
should
write
a
baggage
test
in
here
and
it's
like.
Oh
that
doesn't
work
and
because
yeah.
It
also
said
like
on
the
propagators
that
everything
is
supposed
to
take
like
an
inject.
An
x
extract
function,
and
it
just
said,
like
basically
baggage
and
spans,
there's
no
difference
between
them,
like
they're,
just
contexts
in
general,
right.
B
Yeah,
that's
a
that's,
basically
how
it
works
this
just
right
now.
The
way
it
was
done
is
the
context.
Module
has
this
helper.
That
takes
the
the
key
you
want
to
pull
out
of
context
and
that
that's
not
required
to
be
used,
but
yeah
it
exists
just
because
otherwise
the
span
context
ones.
It
would
just
be
duplicate
code
to
write
that
in
each
one
pulling
it
out
putting
it
back
in
so
I
abstracted
it
but
yeah.
I
can
see
okay,
confusion.
B
A
Yeah,
it
says
like
text
map
propagator,
performs
injection
description,
blah
blah,
getter
and
setter
are
optional
helper
components
used
for
extraction
and
injection,
respectively
defined
as
separate
objects
from
the
carrier
to
avoid
runtime
allocations
by
removing
the
need
for
additional
interface
implementing
objects
wrapping
the
carrier
in
order
to
access
its
contents.
Okay,
so
it
is
for
the
carrier
so.
B
A
B
A
B
I
wonder
if
that
means
for
a
generic
propagator,
because
my
text
map
is
just
a
map
of
text,
so
I
don't
think
we
should
need
to
get
her
insider
there,
but
maybe
I
guess
it
would
make
the
oh
wait.
No,
that
wouldn't
work,
because
that
would
have
to
be
outside
of
the
propagator
themselves,
because
those
are
those
just
act
like
you
couldn't
have
a.
B
Rabbit
amqp
propagator
that
does
its
type
of
headers
combined
with
the
w3c
propagator
that
does
other
type
of
headers.
So
right
exactly
so.
That
has
to
be
something
higher
level
on
the
propagator
like
the
whether
it's
a
text,
map,
propagator
or
binary
propagator
or
whatever,
I'm
guessing
I'll,
have
to
look
at
that
spec
again.
A
Yeah
that
part,
the
thing
it
was
giving
me
before
was
that
the
composite
propagator
part
where
it
was
like
these
cross-cutting
concerns
can
all
be
put
together
and
then
under
global.
It
was
basically
saying,
like
platforms
like
net,
maybe
preconfigured
out
of
the
box
propagators
if
preconfigured
it
should
contain.
A
A
We
have
like
the
setup
and
like
the
thing
that
you
have
now
for
like
the
with
the
with
the
aggregator
of
the
propagators
that
that
is
essentially
like
what
the
innards
of
the
composite
propagator
would
do,
but
without
the
actual
interface
for
it.
So
because
they're
supposed
to
be
like
a
composite
propagator
with
like
a
create
function
and
get
set,
gotcha
yeah.
B
For
those
languages
like
java,
probably
go
that
aren't
as
expressive,
so
they
have
like
an
explicit
composite
propagator
because
they
can't
just
give
a
list
of
propagators
like
we
can
erlang,
there's
not
really
a
need
for
a
composite
propagator.
A
B
Definitely
I
still
wear
is
confusing
and
yeah
what
need
talks
in
the
module
that
says
like
because
it
yeah
it
would
be
verbose
and
unidiomatic
if
we
added
some
of
that
stuff,
so
it
would
be.
We
like
there
needs
to
be
docs
in
there
that
it's
like
in
the
spec
there's
a
composite
propagator.
A
Because
there
is
essentially
a
singleton
textmap
propagator,
it
seems
like
if
you
wanted
to
propagate
a
text
map
in
any
other
format
or
with
any
other
set
of
propagators
within
it
like.
If
somebody
wanted
to
only
send
trace
contacts
without
trace
date,
let's
say-
or
they
only
wanted
to
send
that
without
the
baggage,
whatever
like,
if
you
as
soon
as
you
set
it
up,
it's
kind
of
like
this
is
set
in
stone.
B
A
B
B
If
you
need
different
results
from
the
propagator,
then
the
two
tuple
list-
you
you
can't
you
either
can't
use
the
text.
Map
inject
and
extract.
You
have
to
create
a
new
version
of
that.
Just
like
with
binary
is
going
to
be
done
or
you
you
have
to
take
what
they
give
you
and
convert
it
after
the
fact,
but
yeah
text
map
propagator
it's
going
to
be
those
two
tuples
and
a
list
of
them
so
yeah.
I
think
I
think
the
oddity
there
is
going
to
be.
B
You
have
to
create
a
separate
type
of
propagator
which
not
really
a
good
idea,
because
you
want
the
user
to
be
able
to
say
they
want
trace
context
and
baggage
sent
in
their
propagation,
even
if
amqp
is
being
used.
So
I
think
the
best
bet
would
be
having
to
convert
it
after
you,
inject
or
or.
A
A
But
then
I
mean
that's
kind
of
the
challenging
part
like
it
could
attack.
It
could
take
a
text
map
and
then
essentially
like
or
it
doesn't
necessarily
have
to
be
in
text
map
format.
It
could
be
as
well.
It
could
also
be
like
in
the
normal
amqp
header
format
with
the
list,
but
we
just
ignore,
because
all
we're
doing
is
adding
to
it
on
inject
and
extract.
A
So
like
the
thing
that
I
have
now
like,
basically
like
you
can
inject
with
the
normal
http
headers
before
or
after,
and
it
just
loops
through
all
of
them,
and
just
says
like
is
the
just
pattern
matches
like.
If
this
is
a
two
parity
tool,
then
go
ahead
and
add
long
string
in
the
middle
and
then
yeah
it
just
returns.
It
so
and
it'll
be
the
same
with
the
extraction
or
whatever,
like
we're
just
looking
for
those
particular
things
in
the
first
position
and
then
extracting
the
state
from
it.
A
So
the
innards
of
like
the
hack,
it's
basically
just
to
try
and
wrap
it
up
into
one
thing,
though
yeah,
so
that,
like
it's
not
like.
Oh
you,
gotta
inject,
trace
contacts
with
the
special
thing,
and
now
you
gotta
do
baggage
as
well
like.
I
want
people
to
be
able
to
say
like
have
like
a
composite
version
right
where
it's
just
like.
I
want
all
of
the
w3c
headers
into
to
be
injected
whatever,
especially
because
they
have
to
call
it
explicitly.
A
B
B
Why
not?
Why
can't
it
go
through
the
global
part
if
you're
converting
after
it
after
it
injects?
Oh,
no,
I
guess
you're
right,
yeah
yeah!
You
would
need
some
function
in
your
library
that
does
that
conversion.
That
user
has
to
call.
A
A
A
Okay,
I'm
probably
just
gonna
continue
down
that
path,
then
of
just
it'll
be
a
little
happy,
but
I
know
people
kind
of
need
it
and
it
would
be
the
so
that
was
like
the
thing
I
was
trying
to
get
up
for
the
first
contrib
thing
so
that
I've
been
pressing
them
like
everything
is
set
now,
except
they
did
not
give
us
admin
rights
on
the
repo
okay.
A
A
Set
up
the
new
team
and
then
yeah,
we
just
need
the
admin
right,
so
I
can
actually
set
ci
up
okay,
cool
and
that's
why
I've
been
asking
all
of
them.
I'm
sure
you've
noticed,
but
I've
been
asking
all
the
ones
that
have
been
contributing
to
the
open,
telemetry
beam
yeah
to
sign
the
cla,
because
I
also
don't
want
to
have
to
like.
A
I
don't
want
to
like
import
all
the
get
history,
because
it's
going
to
be
a
gigantic
pain
in
the
ass
yeah,
so
we're
going
to
copy
it
in,
but
just
make
sure
everybody
signed
everything
and
then
we
can
modify
as
we're
bringing
each
repo
in
with
updating
the
copyright.
I
guess.
A
Because
yeah,
it's
like
really
confusing
for
the
the
one
thing
I've
noticed
for
the
people
that
are
all
these
people
that
have
jumped
in
in
like
the
last
week
and
a
half
is
like
they.
They
have
like
a
lot
of
good
intentions,
but
they
don't
know
like
the
basic
things
like
there
is
a
spec
and
there
are
conventions
and
you
have
to
adhere
to
them.
Yeah.
You
can't
add
anything
like
you
want.
A
If
you
thought
it
was
a
good
idea
or
or
leave
things
out
so,
but
then
they're
like
publishing
the
package
and
just
kind
of
like
oh
man.
I
don't
want
people
to
like
rely
on
this,
but
so
I
just
want
like
them
to
be
able
to
go
to
the
repo
and
just
be
like
you
have
to
go
through
this
gate,
and
we
can
tell
you
before
you
publish
a
package
in
the
name
space.
A
Yeah-
and
it
seems
like
pretty
good
at
this
point,
like
everything's,
pretty
locked
at
this
point,
but
yeah
on
the
amqp
thing
like
w3c-
has
actually
got
an
editor's
draft,
but
it's
for
1.0.
A
Oh,
so
we're
going
to
be
kind
of
rogue
regardless
because
I
guess,
like
the
1.0
format,
has
like
two
property
seconds
and
the
message
there's
like
a
properties
and
application
properties,
but
zero
nine
doesn't
really
have
that,
like,
I
guess
the
equivalent
of
message
properties
are.
The
headers
are
part
of
the
headers.
A
Like
ttls
and
things
like
that
right,
so
I
guess
in
one
they
separated
them,
but
so
anyways.
This
is
gonna,
be
kind
of
broke
like
we're
just
going
to
use
headers.
I
guess
because
I
think
correct
me
if
I'm
wrong,
but
amqp
always
has
the
notion
of
headers,
no
matter
whether
it's
rabbit
or
anything
else
right.
B
A
A
Get
somebody
on
the
technical
committee
to
get
that
admin
rights
set
up
because
we're
still
kind
of
blocked,
essentially
from
using
the
repo,
until
that
happens,
yeah
okay,
because
we
can't
set
up
any
of
the
checks
or
anything
right
you're
supposed
to
be
able
to.
But
I
guess
like
the
way
that
our
maintain
is
set
up
permissions.
It
doesn't
have
the
full
set
of
maintained
permissions
that
are
supposed
to
be
there.
A
A
So
yeah
like
I,
was
trying
to
get
that
up.
I
was
going
to
bring
phoenix
in
first
and
then
try
to
get
the
amqp1
in
there
because
we
needed
it
work
and
then,
at
that
point
we
would
have
at
least
some
of
the
base
paths
set
up
and
we
can
get
the
documentation
and
easy
cla
all
that
stuff.
A
B
Updated
by
next
week,
I
can
plan
to
make
the
meeting
to
bring
it
up
and
someone
will
probably
do
it
right.
Then,
if
I,
if
I'm
able
to
do
that.
A
You're
in-person
pestering,
and
not
just
in
slack,
pestering
yeah.
Okay,
I
guess
the
the
only
other
things
on
the
contributor
that
I
wanted
to
ask
you
about
were
one
is.
A
A
A
A
That
was
one
of
the
questions
like
that
dave
had
was
like
if
it's
a
smaller
library
like
who
all
can
put
stuff
in
there
yeah
are
you
just
gonna
have
everything
in
there
or
do
we
have
an
open,
telemetry
organ
hex.