►
From YouTube: 2021-07-14 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
Sean
oliver
is
one
of
the
aws
interns
and
I've
been
helping
him
with
the
aws,
the
aws
portion
of
the
open,
telemetry
php
contrib
recently,
and
he
has
one
I
spoken
as
somebody
who
has
a
pretty
strong
hat
game.
Oliver's
hacking
is
very,
very
strong.
A
B
A
B
D
D
Went
kayaking
at
the
weekend
district,
so
that
was
a
nice
break
just
for
one
night,
but
where'd
you
go
kayaking
the
lake
district,
so
kind
of
yes
northwest
england,
a
bunch
of
lakes
and
mountains
and
stuff
about
two
and
a
half
hours
from
here.
D
A
My
university
is
about
two
hours
away,
so
I
know
that
length
of
drive
pretty
well.
We
have.
We
have
something
very
special
coming
up
in
the
fall
and
that's
football
and
as
you
as
you
brits
call
it
or
joke
about
to
me.
I
like
the
brits
that
I
know
they
joke
about
it
to
me.
They
call
it
hand.
A
Yeah,
that's
good
all
right,
so
I
think
this
is
all
we'll
have
today,
so
we
can
get
going.
Let's
take
a
look
at
the
open,
pr's
and
the
open
bugs.
So
I
think
they're
sean-
I
don't
know
if
you're
aware
yet
or
not.
We
created
an
open,
telemetry,
php,
contrib
repo-
maybe
not
this
past
week
or
the
week
before
so
we're
gonna
start.
A
I
just
wanted
to
make
sure
that
you
knew
about
it.
We're
gonna
start
putting.
We
have
contrib
stuff
in
the
like
the
root,
repo
or
whatever,
but
eventually
we're
gonna
move
all
those
things
to
the
contrib
repo,
but
slowly
and
truly
as
needed,
rather
than
just
like
forcing
people
to
do
it.
So
I
just
want
to
make
sure
that
you
had
contextual
awareness
there,
but
I
think
the
most
important
thing
that's
open
right
now
is
your
otl
http
first
pass
repo.
Is
that
ready
for
review
now.
D
D
D
A
Yeah
I've
left
that
for
the
minute.
Okay,
I'm
I'm
not
like
unbelievably
concerned
about
that
right
now.
I
don't
think
that's
like.
I
understand
that.
That's
you
know
it's
kind
of
itchy
to
not
be
able
to
write
a
test
for
something
that
you
implemented.
It
never
feels
good,
but
yeah.
I'm
trying
to
think
like
you
could
mock
the
response
and
make
sure
that
it
came
back
right
or
you
could
like
compare
a
binary
like
a
binary
output
from
something
but
yeah.
D
D
The
way
that
I
figured
it
could
work
because
you
basically
just
yeah
you
own,
you
get
the
response
that
was
sent,
you
unencode
it,
and
then
you
convert
the
binary
to
hex
and
just
check
that
it
vaguely
is
the
same,
but
I
didn't
find
that
out
to
be
too
clean
and
it
was,
it
would
have
been
a
bit
confusing
and
also
because
there
was
a
span
date
in
there.
Basically,
every
time
it
came
back,
it
came
back
with
a
different
encoding
yeah.
D
So
I
figured
I'd
leave
that
out.
For
now.
That
might
be
something
that's
better
tested
by
kind
of
an
end-to-end
test.
If
you
basically
span
up
the
collector,
because
what
I
found
with
the
collector
is
you
can
get
the
collector
to
output
directly
to
a
file
you
could
which
basically
dumps
the
or
tlp
as
json
into
a
file.
You
could
do
some
some
more
like
higher
level
tests
without
that,
especially
for
the
otl
or
trp
that
might
be
best
but
yeah
I
left
out.
D
There
is
a
comment
in
the
code
base
to
basically
think
at
least
like
this
could
do
with
testing
but
yeah,
then
so
yeah.
I
left
and
then
there's
two
other
to
do's,
which
is
supporting
kind
of
pass.
This
end
points
to
basically
add
on
the
v1
traces
if
it's
missing,
so
people
don't
have
to
necessarily
specify
that
that
seems
to
be
available
in
other
sdks
and
then
the
passing
export
config
through
the
constructor
would
be
also
nice
to
have.
But
I
wanted
to
kind
of
get
this
freed
up
and
out.
There.
A
Do
you
think
it's
worth
just
creating
an
new
github
issues
for
those
things.
A
I'll
make
you
a
deal,
I
will
review
this
today
and
I
will
create
issues
for
both
of
those
and
then
we
can.
We
could
probably
just
merge
it
unless
somebody
else
says
no,
but
I
doubt
it
that
will
be
the
case
cool.
So
thank
you.
Thank
you
for
all
your
work
on
this.
It's
great.
I
will
review
it
either
this
afternoon
or
tomorrow
morning,
depending
on
how
my
my
daughters
came
home
from
school
sick
today.
B
A
couple
things
I
I
just
uploaded
three
issues:
yesterday,
377
and
378-
are
I'm
sorry,
378
and
yeah.
377
are
the
ones
that
we
talked
about
a
couple
days
ago.
So,
if
you
can
assign
me
those
and
then
I'll
I'll
submit
some
pr's
later
today.
For
these
two
sure
these
are
kind
of
temporary
fixes
to
a
more
deep-rooted
issue
that
I
want
to
talk
about
later
today,
or
I
guess
right
after
this.
B
So
the
and
then
there's
issue
376,
which
is
the
semantic
convention
generator,
which
I
was
just
going
to
say
to
add
the
release
required
for
ga
tag
on
that
one
just
did
and
then
so
yeah
those
two.
B
So
these
two
kind
of
fix
some
hard
coding
it
chooses
in
span
context
and
the
tracer
and
span
options
when,
when
I
was
trying
to
run
the
examples
and
that
that's
when
I
ran
into
these
problems,
so
these
two
top
issues
once
they're
solved,
we'll
we'll
fix
that.
B
But
the
more
deep-rooted
issue
is
that
span
context
has
like
four
four
functions:
four
or
five
functions
that
should
just
not
exist.
There
they're
called
generate,
generate
sampled
fork
and
restore,
and
from
other
implementations
in
in
other
repos
in
other
languages
and
from
the
specifications
there
shouldn't
be
any
like
generate
functions
for
these.
B
It
should
just
a
spam
context
should
just
be
created
through
the
con
to
the
constructor,
like
just
like,
creating
a
new
span
context,
and
then
that's
also
because
these
generate
functions.
Use
like
yeah
like
a
random
hex
thing
when
it
should
be
using
an
id
generator.
B
So
that's
why
you
should
be
like,
like
whenever
you're
like
in
a
tracer
or
something
you
generate
your
traces
with
an
ide
generate
provided
by
the
trace
provider,
and
then
you
just
create
a
new
span
context
with
those
sit
like
those
generated
trace
and
span
ids
through
the
constructor.
You
just
like
place
those
in.
B
So
that's
that's
what
I
was
going
to
talk
about,
but
since
these
four
functions
are
used
mostly
they're
used
like
within
testing
like
unit
testing
like
just
to
generate
a
span
context.
So
it
would
be
a
little
bit
more
tedious
to
kind
of
remove
these
functions
and
then
replace
them
with
just
creating
a
constructor
but
yeah
these.
B
These
four
functions
are
like
very
circular,
also
like,
like
you
call
there's
like
generate,
and
then,
if
you
want
to
generate
a
sample,
that's
like
just
calls
generate
true
and
then
restore
calls
fork
and
fork
calls
or
store
it's,
it's
very,
like
kind
of
very
intertwined
when
it
could
just
be
all
one
function
or
should
actually
just
I'll
be
the
constructor.
B
A
Sure
so,
I'm
looking
at
those
and
I'm
seeing
a
lot
of
commits
from
dimitri
who's,
the
initial
committer
from
this
repo.
So
just
that
makes
me
think,
like
this
happened
when
the
spec
was
still
not
like
completely
solidified,
and
it's
probably
changed
it
a
lot.
So
thank
you
for
noticing
this
and
thank
you.
D
A
For
being
willing
and
able
to
work
on
it,
because
I
think
that
you're,
I
think,
you're
absolutely
correct
that
it
should
probably
be
part
of
the
constructor
or
it
should
probably
be
part
of
some.
You
know
something
else,
but
I
think
yeah
at
the
time
the
idea
of
a
constructor
pro
or
the
idea
of
a
well.
What's
the
word,
I'm
looking
for
blanking,
what
you're
working
on
was
not
a
thing
then.
B
Yeah
so
yeah
it
makes
sense.
Okay,
so
I'll
create
an
issue
for
that
and
then
once
I'm
done
with
my
project
I'll
tackle
that.
A
A
B
B
B
Oh
and
also
the
semantic
convention
generator-
I
don't
think,
should
be
assigned
to
me
because
I
I
oh
okay,
yeah
that
one's
just
like
a
general
thing,
because
I
don't
think
we
have
anything
for
just
semantic
conventions
at
all.
Got
it.
Okay,
I'm
sorry!
I
misunderstood.
B
A
B
Good
sean,
quick
question:
I
know
you
created
all
the.
I
think,
a
lot
of
the
examples
in
the
in
the
original
and
sorry
in
the
core
repo-
and
I
was
just
I
was
just
like
letting
you
know
that
I
sometimes
am
using
these
and
if
I
do
have
questions
I'll.
I
guess
I'll
reach
out
to
you.
So
just
look
out
for
that.
D
Help
me
with
that
so
yeah
I
knew
you're
out
those.
I
think
we
need
to
figure
out
a
nicer
way
to
run
those
examples
because,
like
editing,
I
kind
of
when
I
want
to
run
things.
I
start
up
and
take
the
first
line
from
the
docker
of
the
make
file
and
just
run
that
container
and
then
run
the
second
line
separately.
So
it
doesn't
kill
it
straight
away.
D
Oh
okay,
because
yeah
at
the
minute,
we
don't
really
we
don't
kind
of
output,
any
debug
information
which
might
be
handy
to
know
that
kind
of,
because
it
will,
it
will
sit
as
sent
when
it
hasn't
actually
correctly,
when
it
hasn't
succeeded,
really
which
yeah
isn't
isn't
great.
So
I
basically
break
apart
that
and
that
isn't
very
clear
to
anyone
else.
B
Actually
able
to,
I
was
actually
able
to
get
the
hotel
grcp1
example,
the
first
one
to
connect
to
the
aws
collector,
and
actually
I
saw
the
traces
on
x-ray
on
the
back
end
for
x-ray.
So
I
don't
know
if
I
told
you
bob,
but
I
got
it
to
work
yeah.
I.
B
It's
exciting
so
yeah
I
like
all
I
did-
was
just
download,
grcp
grpc
locally
and
then
just
ran
it
and
just
made
sure
that
all
my
credentials
were
okay
for
that,
but
yeah
I'll,
probably
write
one
or
two
example
apps,
because
I
need
to
just
for
to
like
run
it
on
an
eks
instance
as
well.
So
I
can
like
donate
those
as
well
for
for
the
examples.