►
From YouTube: Lang Team Meeting 2019.05.02
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
C
B
Some
I
tried
to
put
some
notes,
feel
free
to
interrupt
to
this.
If
you
have
questions
or
comments,
so
this
P
R
is
your
nominative
yours,
a
notification,
it's
a
PR
to
stabilize
aspects
of
maybe
uninit
and
the
primary
question
I
wanted
to
race
today
is
like,
and
we
need
to
schedule
this
for
deeper
discussion
and
there
are
people
feeling
like
they
are
ready
to
check
their
buses
or
at
least
ready
to
read.
First.
E
F
Ahead,
Josh,
oh
no,
it
was
me
boats,
but
my
only
concern
is
about
deprecating,
maybe
uninitialized
and
schedule
there
yeah
yeah
cuz.
Obviously
we
have
to
eventually,
but
it's
you
think
it's
perhaps
too
soon
or
yeah,
so
I'm
worried
that
it's
too
widely
used
and
there'll
be
really
disruptive
and
you
could
figure
out
a
way
to
communicate.
Well,
like.
F
I,
just
it's
just
one
of
those
things
where
it's
like
I
want
to
make
sure
that
we
are
communicating
well
before
they
like
the
ideally
people
with
who
are
attentive
would
know
about
it
before
they
I'm
Erin.
We
have
like
a
process
where
we
do
it
three
months
later,
purely
surrender
later
right
and
that's
like
mm-hmm.
F
B
I
just
want
important
ideas,
all
right,
central
dude
Kent,
maybe
it'll
be
good
for
one
of
us
to
prepare
a
sort
of
blog
post.
Saying
like
we
are
deprecating
uninitialized
starting
this
date,
yours
what
you
should
be
doing.
Instead,
you
know
maybe
I,
don't
know
exactly
where
this
should
be
floated,
but
basically
a
documentary
by
white
people.
Adam
we
can
post
it
in
advance.
I
would
propose
that
love.
Does
this
I
think
he
would
be.
C
By
it's
a
good
good
good
idea
that,
if
you're
interested,
he
would
be
a
good
person,
do
it
yeah
if
he
has
time
that
is
yeah,
as
as
for
deprecation
I've
proposed
to
two
releases,
which
is
like
three
months
from
it
being
stable,
which
is
I
mean
you
would
have
some
time
even
some
more
time
on
nightly,
but
I
think
that
if
we
notice
that
okay,
it
was
too
much
there
were
too
many
too
many
goo
regressions.
We
can
push
it
back
some
more
time.
We.
E
G
G
F
F
G
Right,
there's,
a
note,
I
think
is
worth
mentioning
by
the
way
I've
noticed
that
people
often
have
to
do
a
lot
of
work
pulling
together
the
release,
notes
close
to
a
release
and
figuring
out
what
things
are
relevant
and
that
also
seems
prone
to
causing
things
to
get
missed.
So
on
that
basis,
I'm
wondering
if
it
would
be
a
good
idea
in
general
to
adopt
a
more
general
policy
or,
if
you're,
making
a
change
to
the
compiler
that
is
going
to
introduce
a
release.
G
C
G
C
B
Issue
six
over
300.
This
is
a
proposal
to
basically
it
not
warn
about
things
in
my
boundaries,
if
I
understand,
with
the
with
the
lint
that
we
have
for
this,
unless
particular
to
allow
things
like
this,
which
are
supposed
or
represented
as
a
pointer
in
practice
and
which
the
FFO,
the
unsafe,
clear
guidelines
proposal,
also
sort
of
guarantees
in
some
cases-
and
the
question
is.
E
B
G
G
B
E
B
I
E
B
C
G
G
B
I
I
B
B
So,
let's,
let's
skip
down
to
ASIC
away
and
come
back
to
these
I,
don't
think
is
anything
to
burning
in
here.
Yeah
all
right,
I'm
gonna
hand
over
at
this
point,
two
votes.
F
F
Essentially,
it
seems
to
me
that
when
we
have
cats
and
we
can
express
streaming
iterator,
we
will
want
to
feel
the
process
streaming
iterate
for
loops,
and
so
that's
already
a
problem
we
need
to
solve,
and
so
we
can
just
make
it
possible
to
go
from
a
stream
to
a
streaming
iterator.
We
also
changed.
Name
is
streaming
iterator
to
a
stream
like
it's
a
different
concept
that
were
you
like
using
this
better
for
40
different
things.
Yeah.
F
Into
iterator,
because
the
item
is
supposed
to
be
the
next
future
and
the
next
feature
borrows,
even
though,
even
though
the
output
type
like
the
item
type
of
the
stream
doesn't
borrow
the
the
future
that
you'll
be
getting,
does
borrow.
So
even
in
regular
streams,
not
streaming
streams
are
streaming
iterator.
It's
not
iterators.
E
F
F
E
F
E
F
F
E
F
F
F
E
F
G
E
Kind
of
like
thing
but
I
don't
like
know
that
that's
what
you're
imagining
so
yeah.
It's
also
interesting
that
if
you
wrote
like
continue
without
awaiting
the
future,
the
next
thing
you
get
would
still
just
be
the
same
element
in
the
in
the
stream
yeah,
which
is
pretty
counterintuitive
yeah,
so
like
I,
can't
quite
make
this
work,
but
it's
a
cool
idea.
I'm
like
excited
to
buy
it
yeah
I
think.
G
Think
it's
pretty
clear
that
bounces
suggestion
would
allow
us
to
separate
the
concepts
of
integrating
this
in
a
for
loop
versus
the
syntax
and
I
agree
with
the
point
of
it's
not
that
we
have
to
commit
to
making
this
the
syntax.
It's
that,
if
we're
gonna
do
it.
This
sounds
like
a
reasonable
approach
and
we
should
try
to
achieve
the
property
of
let's
make
it
orthogonal
to
the
await
syntax.
C
A
C
B
Yeah
I
have
often
thought
for
loop
is
not
that
great
a
candidate
to
do
your
lowering
like
it's
nice,
but
it's
also
very
simple
to
type
check
and
you
could
do
it
in
your
tee
sugaring
time
not
just
as
using
I
think
that's
certainly
on
the
that
would
be
so.
We
could
have
different
distributing
for
yeah.
We
could
imagine
us
saying
we
it's
not
it's
the
kind
of
thing
I'm
not
like
wild
about
in
the
type
checker,
but
we
already
do
it
from
place
to
place
to
try
to
avoid
it
but
sort
of
saying
well.
E
E
F
A
D
G
G
F
So
but
I
think
then
I
feel
comfortable,
saying
if
we
ever
have
this
syntax,
it
will
work
this
way
and
not,
and
we
don't
need
to
worry
about
like
having
a
specific
408
syntax
that
we
need
to
worry
about
is.
E
I
F
F
F
F
Think
that
we
should
also
we
should
go
to
that
now
and
then
communicate
to
the
community,
that
that
is
one
the
decision
we
made
by
that
way.
There's
an
understanding.
Everyone
has
understanding
of
what's
gonna
happen,
and
then
that
also
gives
us
time
to
make
a
decision
and
then,
after
the
decision,
there'll
be
the
time
from
when
we
made
it
until
we
make
a
stabilization
report,
so
people
can
like
process
their
feelings
about
the
decision
I
made
before
we
make
a
stabilization.
C
F
C
B
C
B
B
The
same
thing,
yeah
that
seems
correct,
I,
mean
I,
think
that
I
would
like
to
talk
a
little
bit
about
what
are
the
things
like
these
other
stabilization
blockers
make
sure
we're
almost
into
what
pieces
we
need
something
I'm
sure
they
get
done
and
I
think.
F
Then
that
third
bullet
on
this
list
of
the
blockers
still
beating
like
I,
mean
I,
there's
been
some
stuff
up
I'm
reading
about
it
up
is
your
accost
abstraction
or
whatever
for
reasons
they're
like
basically
like
we
don't
have
the
perfectly
optimized,
for
that
we
know
is
possible
yeah
sort
of,
but
it
comes
down
to.
It
seems
like
so
I'm.
E
F
Yes,
I
think
that
we
should
just
they
just
we
all
we.
Basically
it's
like
to
send
the
same
things
with
this
number
of
misplace
things.
We
do
blog
posts.
We
need
to
be
actually
I.
Think
after
the
meadow
working
group
meeting.
I
think
that
this
is
the
theme
that
we're
finding
I
think
right
now
is
something
to
communicate
better
with
the
community
for
the
city,
the.
E
It's
the
same
thing:
animal
treat
it's
not
like
it
doesn't
manually
coerced
it
yourself.
It's
just
that
like
right
now,
the
way
I'm
talking
like
inference
doesn't
line
up
nicely
I,
see,
okay.
That
makes
it
get
pushed
back
far
enough.
It's
just
a
bug
we
can.
We
can
certainly
fix
it.
It's
not
even
like
that
hard
to
fix,
there's
already
an
open
issue
that
fixing
not
like
discussing
how
to
fix
it.
So,
okay,
good.
This
was
earlier.
E
You
use
like
try
combinators,
like
when
you
use
question
mark
in
a
in
acing
FN,
but
then
you
have
like
a
return
that
doesn't
use
a
result.
Value
it'll
like
in
furthers
the
return
type
to
be
the
incorrect
type
and
give
you
a
like
a
an
error
that
makes
less
sense
than
the
one
we
want
to
give
you
and
fixing.
That
is
the
same,
is
fixing
list
cool
one.
B
Thing
I
would
like
to
talk
about.
Is
this
documentation
aspect
I
think
it
would
be
really
great
if
we
in
general,
but
also
here,
had
sort
of
explainers
talking
about
the
future?
What
it
is
and
probably
laying
on
some
of
these
shortcomings.
I
know
there
was
like
an
async
book,
it
was
I,
don't
know
where
the
status
of
that
book
is
what
exists
here.
C
B
What,
when
it
comes
to
the
test
week
along
those
lines,
central
I
was
gonna
say
one
thing
I
found
is
kind
of
nice
is
company,
people
who
are
not
doing
implementation
are
often
putting
out
corner
cases
that
don't
occur
to
inventors
and
I'm.
Wondering
I
know
you
love
to
make
tests
to
the
idea
of
looking
over
accessories
and
time
you
see.
If
you
there
are
corners.
That
models
think
might
be
worth
testing
done.
B
C
E
E
E
C
B
So
and
that's
fine
I,
don't
know
if
it
needs
to
be
you
I
I,
just
thinking
that
in
past
I
have
sometimes,
for
example,
going
over
and
Orsi
and
read
through
the
RFC
and
said
okay
without
looking
at
the
code.
What
are
the
tests
like
that?
I
would
do
like
as
I
went
through
I
think
that
would
be
a
potentially
interesting
exercise.
B
But
I
guess
for
the
documentation:
I,
don't
know
what
what
I
think
we
want
I
think
we
want
like
something
that
says
the
documents
way
they
think
FN
does.
This
is
when
you
do
async
FN.
It
will
do
sugar
into
a
function
that
returns
a
future.
You
can
use
antique
move,
blocks
and
kind
of
shows
the
patterns.
We
expect
people
to
use
I'm,
not
really
sure.
If
we
have
clears
up
in
there,
you
can
read
that
tells
we've
got
yes.
E
The
never
existing
chapters
in
the
basic
book
that
do
that
they
would
probably
we
could
easy
things
to
pull
from
a
lot
of
that.
The
big
issue
with
async
book
grazers,
but
a
lot
of
it,
has
fallen
out
of
date
because
I
haven't
been
continually
updating.
It
response
all
the
various
changes
so
like
we'll
just
have
to
go
through
and
revamp
it
like
when,
when
we're
ready
to
stabilize,
then
we
know
what
final
things
it's
so
so
like.
F
G
F
E
Also
I
think
I
think,
since
it's
just
very
to
get
out
now,
I
might
be
interesting
to
too
quickly
if
we
have
like
five
minutes,
which
we
probably
won't
respond
to
the
whole.
Like
poll
that
people
said
it
and
the
results
from
that,
because
that
was
kind
of
an
interesting
thing
to
read
their
works
them
I.
Don't
know
that
I
would
drop
dudes
off
of
it,
but
it
was
interesting
there.
B
F
And
I
think
said:
I
would
like
to
limit
our
consideration
to
either
dot
o8
or
spatially
postfix
in
Texas
I
know,
and
even
probably
the
one
thing
that
just
referring
to
from
the
poll
is
dot.
There's
really
that
method
and
put
six
macro
syntax
were,
according
to
the
poll,
more
widely
preferred
my
penis.
F
This
is
just
people
are
wrong,
like
not
like
the
sensitive
like,
not
in
the
sense
of
what
my
opinion
but
I
think
it
is
like
yeah
I
think
it
is
a
good
thing
today,
like
it,
isn't
thing
that
I
think
we
should
not
do
but
with
Justin
it
isn't
actually
a
method
or
a
macro,
and
so
it
should
not
look
like
them.
Yes,
because
a
person
could
very
reasonable.
We
believe
it
it's
a
method
or
macro
or,
as
a
person
could
never
reasonably
believe
it.
F
It's
a
field
access,
then,
over
the
way
it
actually
does,
which
is
why
dot
await
doesn't
really
like.
The
idea
is
that
it's
not
a
field
access
but,
of
course,
there's
a
problem
that
it
does
look
like
that's
the
reason
why
we
should
consider
space
away
is
because
it
isn't
uni,
just
syntax,
and
so,
if
you
think
that
actually
it
could
peek
at
these
four,
because
X
is
important
and
we
should
go
with
absolutely
can't
confuse
with
anything.
I
think
those
are
the
two
that
are.
We.
G
There
are
some
tackling
issues
with
space
in
terms
of
grouping
that
I
recognize
and
acknowledge.
Boats
points
that
we
don't
necessarily
want
to
discourage
people
from
making
to
encourage
people
to
make
incredibly
complicated
expressions.
But
space
syntax
makes
it
awkward
to
do
any
chaining
at
all,
because
it
immediately
makes
it
look
like
it
groups
to
the
right
and
not
to
the
left.
Yes,
I
had
a
slightly.
A
I
I
F
I
F
I
E
It
is
obviously
not
a
perfect
macro.
It
is
obviously
not
a
method
call.
It
is
obviously
not
a
field
access,
so
that
kind
of
falls
out
in
anyway
have
to
wait
differently
based
on
whether
or
not
we
decide
to
add
Pacific
macros,
but
I.
Also
I.
Think
that
that's
one
of
it
big
detractors
right
is
that
it
sort
of
suggests
at
a
postfix
back
feature
which
doesn't
currently
exist
today
and
feature
which
I
I
am
a
little
skeptical
of
I.
Think
that
it
would.
C
H
G
I
would
like
we
call
attention
to
one
thing
as
well
that
I,
so
the
community
concern
regarding
this
looks
like
a
field
access,
so
it's
not
obvious
that
it
affects
control
flow
I.
Don't
I
want
to
register
that
concern.
I
want
to
document
a
good
rationale.
Why
we're
not
I'm
not
suggesting
that
that
a
blocker
for
us
I'm,
suggesting
that
we
need
a
good
answer
for
that?
Yes,
I,
think
there's
two
good
answers.
G
First
of
all,
I
agree
completely
with
boats
point
that
I
feel
you
know
other
syntax
that
conveys
the
idea
that
it
you
know
if
it
looks
like
it
can
run
code.
That
doesn't
necessarily
mean
it
looks
like
it
can
affect
control
flow,
which
is
entirely
different.
So
anything
short
of
full
postfix
macro
is
going
to
potentially
look
surprising
to
some
degree
and
given
that
I
think
favoring
simplicity
makes
sense,
there
is
one
other
argument.
I
think
is
appropriate,
which
I'd
like
to
get
people
thinking
about.
G
In
the
background,
which
is
I,
think
it
would
be
an
improvement
to
orthogonality
to
have
things
like
dot
match
or
dot
yes
or
similar.
I
think
that
it
is
to
be
able
to
say
complicated
expression,
dot
match,
open,
brace
and
that
would
feel
very
orthogonal
to
having
dot
await,
because
they
both
mean
exactly
the
same
thing.
They
both
mean
take
the
previous
expression
and
you.
That
is
the
argument
to
this
construct
that
takes
an
expression
so
that
I'm
not
suggesting.
We
approve
that
today,
I'm
not
suggesting
we
agree.
F
C
G
A
B
B
G
Observation
as
well
that
I
feel,
like
anything
that
involves
white
space
separation
with
no
punctuation,
makes
compiler
recovery
more
difficult.
Imagine
forgetting
a
semicolon
on
the
previous
line,
and
now
the
compiler
muddles
forward,
thinking
complicated
expression,
match
open
or
a
complicated
expression
away.
G
I'm,
not
saying
that
it
is
impossible
for
the
compiler
can
recover
compiler
to
recover,
but
I
think
that
the
thing
that
led
to
see
having
problems
with
I
can't
tell
that
you
didn't
put
a
semicolon
on
this
struck
because
well
you
could
be
declaring
a
function
that
returns
this
type,
that
you
construct
it
inside
and
return
type,
and
so
it
takes
a
long
time
realizing.
No,
you
just
screwed
up
no.
B
There
is
a
some
kind
of
inner
stuff
around
it.
I
think
this
is
actually
legal
today,
for
example.
Anyway,
let's
not
go
to
you
on
that
rabbit
hole,
one
of
the
ever
a
few
things
about
field,
I
just
want
to
say,
I
haven't
heard
them
said
I'm
a
little
confused
by
it
like
I,
think,
obviously
syntax
highlighting
which
is
like
this
is
simplest
of
syntax.
Highlighting
of
all
copper
keyboard.
Highlighting
is
clearly
gonna.
B
I
C
F
B
C
C
E
E
C
F
F
A
F
E
I
E
Was
gonna,
ask
we
Nico
and
I
had
discussed
I've
actually
already
implemented,
daughter,
wait
and
I
was
wondering
if
I
should
open
that
PR
or
not
hey.
E
Let
me
post
this
limb,
okay
and
and
my
current
implementation
makes
both
dot
await
and
the
current
the
wait
macro
syntax
valid,
although
it
makes
with
the
wait
Macker
syntax
based
on
keyword
rather
than
based
on
an
actual
macro
and
the
weight
Makarov
is
under
a
spell
feature,
gate
called
Oh,
a
macro
which
currently
happens
conveniently
to
be
the
name
of
the
library
feature
gate
so
that
existing
news
viewers
can
transition
over
that.
B
That
this
was
specifically
requested
to
me
by
several
people.
Actually,
when
a
then
I
heard
you
had
a
PR,
they
said
I
hope
that
I'll
be
able
to
transfer
yeah.