►
From YouTube: 19 July 2018 Meeting
Description
The Rust WebAssembly Working Group meeting from 19 July 2018.
A
C
D
C
D
This
is
ready
in
the
sense
that,
like
I,
basically
drop
a
one-liner
into
my
application,
and
it
does
everything,
so
it
automatically
runs.
Wasn't
by
Jen.
It
automatically
looks
everything
up.
Installs
all
unnecessary
dependencies
runs
away
as
a
mobster
ones
buys
them
all
those
things
really
so
do
we
want
to
make.
Is
that
something
we
want
to
put
on
as
blocking
worse,
20
teens
just
make
sure
we
have
it
done
by
that
way.
D
That
would
be
the
first
one
and
then
the
other
one
which
we
may
want
to
consider
is
along
the
lines
of
like
a
workflow
kind
of
given
to
you
is
tests.
So
we
don't
have
a
great
story
about
testing
right
now
and
the
main
thing
there
is
like
my
karma
test
doesn't
work.
How
do
I
actually
run
boy
run?
Headless
browsers,
I
wrote
node
all
that
various
business
so
flushing
out
that
story
and
making
sure
that
it's
a
basically
there's
a
guide.
A
D
B
D
Story
is
not
complete
if
the
test
fails,
and
you
have
no
clue
as
to
why
so
there's
just
some
level
debugging
where,
like
printing,
should
work
or
like
petting
messages
messages
you
come
out,
but
I
wouldn't
consider
that
the
like
the
testing
story
isn't
blocked
on
a
full
debugging
story,
where
I
can
actually
like
pop
up
a
web
browser
and
have
a
little
console
every
now
and
then
I'll
break
it
any
more
arbitrary
location.
So
for
me,
it's
like
minimal
debugging
is
necessary
to
get
a
good
enough
story
for
2018
I.
A
A
Most
of
these
things
are
pretty
much
assigned
at
least
like
at
the
yasm
pack
level,
like
moving
out
towards
the
integration
level,
a
bunch
of
stuff
isn't
as
assigned,
but
it
seems
like
a
lot
of
the
people
that
these
things
are
assigned
to
kind
of.
Don't
have
a
lot
of
time
to
commit
to
working
on
these
things,
like
kind
of
maybe
active
like
once
or
twice
a
week,
which
is
totally
okay,
but
maybe
not
something
we
want
associated
with,
like
the
blocking
issues
that
we're
trying
to
do
for
a
deadline.
D
Okay,
I
suppose
one
thing
I'll
say
is
that
I've
been
trying
to
make
sure
that
by
hopefully
by
the
release
preview
happening
in
August,
everything
is
stable,
that
we
need,
for
at
least
four
was
imagined,
and
so,
if
there
are
other
kind
of
wise
and
specific
unstable
things
that
I'm
not
aware
of
or
thinking
like
the
custom
sections,
import
modules
and
procedures.
So
if
there's
anything
else
like
specific
one
wisdom
but
nightly
features,
let
me
know
and
I
want
to
see
if
I
could
help.
A
A
A
D
I
started
kind
of
poking
at
it
as
well
and
I
think
a
lot
of
there
there's
some
aspects
that
are
blocking,
adding
more
features
or
enabling
more
video,
which
kind
of
more
deep
in
wasn't
benjin,
so
like
option
is
totally
flooded.
So
I
want
to
make
sure
that
those
are
taken
out
of
the
way
I
like
how
all
the
easy
work
is
left
over
to
just
gonna
plow
through
all
that,
and
so
I'm
currently
aware
of
options.
D
C
E
Okay,
Varian
thingy,
like
currently
the
grade
that
we
are
using
for
the
burial,
has
some
limitations
like
certain
parts
of
it
or
not,
parsed
correctly,
and
there
was
a
comment
that
just
comment
it
out
whatever
that
is
not
getting
patched
correctly,
like
just
an
interface
without
any
explanation
of
the
interface,
it's
not
passed
correctly.
So
do
we
have
any
ideas
of
implementing
that
and
massive
engine
or
go
and
enhance
the
ideal
gate
itself.
Any
plans
for
that
now
so.
A
E
For
interfaces,
we
had
an
issue
which
kind
of
is
not
working
like
if
I
just
give
interface
an
interface
name
together
without
any
explanation
to
that
interface
or
without
any
methods
that
interface
it
kind
of
like.
It
is
not
parsing
that
most
of
the
ideas
that
we
have,
for
example,
I,
started
with
indexdb
and
most
of
the
ideals
inside
that
has
just
a
interface
interface
name
inside,
which
is
kind
of
like
I,
have
to
mask
everything
out
so
I
think
and
also
I,
think
the
repository
Warner
of
the
buy
deal
said
no
for
that
feature.
E
E
They
a
part
of
my
body
inspect
because
okay
I'm
not
pretty
sure
on
providing
spec
but
I've
cross-checked
with
other
burial
likes
from
different
browsers,
the
same
with
ideas,
but
they
are
also
part
of
like
everybody
else
back
inside.
They
also
have
the
same
way
burial
stuff
inside
there,
which
is
why
I'm
kind
of
worried
now.
D
We
had
a
lot
of
success
with
kind
of
the
JIT
them
because,
although
like
all,
they
know
all
the
underlying
stuff
was
there
inputted
in
relatively
well
tested,
but
I
think
we're
having
less
success
with
the
Roboto
stuff
and
just
in
the
sense
that,
like
the
hello
world
works.
But
we
haven't
gotten
much
more
serious.
Just
because
there's
like
very
I
mean
there's
everything
with
already
oh,
create
files,
not
parsing.
D
So
many
types,
not
being
it
wanted
right,
and
so
it
might
be
that
we
just
need
to
like,
amongst,
like
the
people
like,
like
Nick,
haven't
done
a
lot
of
the
original
work
kind
of
do
like
a
couple
of
the
more
serious
web
audio
files
and
like
flush
out
all
the
bugs
associated
with
it
and
then
kind
of,
and
that
might
kind
of
rekindle
the
effort
to
have
anyone
come
in
use.
Whether
like
a
do,
what
I
do
at
some
test,
I.
A
A
A
A
A
B
B
B
C
A
Documentation
so
lots
of
documentation
folks
are
out
so
I
can
do
an
update,
I
think.
As
far
as
I
know,
it's
the
same
status
as
last
week,
there's
been
like
little
PRS
merged,
but
none
of
the
kind
of
more
fundamental
work
that
we've
been
aiming
for,
has
happened
and,
and
so
I
know.
There's
one
person
on
I
think
it's
the
rust
reach
team
who's.
A
A
E
Soon
do
so
I
had
an
initial
prototype
designed
and
then
I
shared
that
with
Ashley
and
I'll
just
create
a
work-in-progress
beer
which
actually
was
mentioning
in
the
group
other
day,
I'll
be
going
on
doing
that
and
also
I
think,
like
I
shared
the
cool
with
you
guys
or
shared
the
picture,
whatever
I
had
in
my
mind
with
you
guys.
So
if
you
have
any
other
feedback,
you
can
just
let
me
know-
and
yes,
that's
that's
pretty
much
it
on
that
side.
A
It
needs
a
little
bit
of
work
for
was
emphatically
of
those
2018
issues,
open
I
plan
on
getting
to
that
in
the
coming
week
and
then
Ashley
is
also
has
a
Wes
impact
template
that
she's
making
work
with
cargo
generate?
Oh
so,
I
guess
the
the
big
news
is
cargo
generate
shipping,
so
there's
a
tool
to
do
scaffolding
and
templates
and
stuff,
and
so
actually
he's
getting
she's
working
on
the
resin
pack
tutorial
in
the
book
right
now
to
get
that
using
cargo
generate
and
Wes
impact,
template
and
stuff.
A
So
I
guess
that
is
an
update
for
that
for
dogs,
but
then
so
once
she
does,
that
I'm,
gonna
kind
of
unify
the
templates
or
or
something
or
I
guess
we
can
figure
out
which
template
we
want
to
keep
or
whatever,
or
have
both
templates
I'll
just
dig
into
that
space
a
little
bit
and
and
try
and
knock
out
the
blocking
2018
issues
that
we
have
open
for
the
template
and
less
impact
integrations
and
that's
2018
goal
status
updates
Alex.
You
want
to
talk
about,
wasn't
benjin
yeah.
D
C
D
D
It's
not
necessarily
quite
ready
yet
we've
had
we
have
the
web
sis
creat,
there's
also
the
new
GS
module,
which
has
a
lot
of
the
gia
school
bindings,
and
so
they
may
not.
We
haven't
necessarily
ordered
those
yet
so
that
may
not
be
quite
ready
that
or
the
other
option,
though
we
have
that
we
could
publish
this
with,
is
we
could
just
go
back
to
the
old
commit
that
had
over
to
11
and
just
I?
Could
just
I
had
like
the
necessary
support
to
make
it
compile
and
stable?
It
was
republished
that
is
Oh
212.
D
This
would
effectively
be
a
branch,
but
it
would
just
be
a
tag
that
I
push
up.
It
wouldn't
actually
be
a
branch
or
anything,
and
so
that
would
be
that
doesn't
actually
pick
up
any
any
updates
from
the
master
branch
and
but
it
is
it.
C
D
A
A
All
the
boxes
checked
in
that
issue,
but,
like
I,
feel
like
it's
like
fairly
complete
now
I
think
we
have
like
something
like
75%
of
the
boxes
checked
and
like
the
ones
that
aren't
checked
are
generally
like
things
that
people
don't
use
very
often
so,
like
I,
think
if
we
wanted
to,
we
could
ship
the
j/s
module
and
it
is
behind
the
feature
that's
not
on
by
default,
or
is
it
on
by
default?
It
must
be
on
by
default.
I.
Think.
D
A
D
A
A
A
A
I,
like
the
idea
of
having
like
a
jss
crate
except
I,
think
I'm
pretty
sure
that
name
is
taken
but
I'm,
trying
to
remember
what
the
issues
were,
because
initially,
when
I
was
setting
up
the
infrastructure
for
all
of
the
J's
bindings-
and
there
was
a
reason
why
that
didn't
work
I
think
I
had
to
do
with
like
orphan
and
poles
and
things
like
that
or
because
we
want
nice
conversions
from
J's
value
to
all
of
these
things.
So
I
guess
we
would
need
to
do
extension
traits
like
FJ
s.
D
Well,
if
that
was
the
major
blocker
I
scan
and
seeing
if
we
can
just
separate
it,
because
I
don't
actually
think
there
should
be
anything.
Any
reason
I
think
like
right
now,
there's
a
method
on
J's
value,
called
as
J's
string
to
give
it
a
J's
string
out,
but
like
we're
gonna,
do
that
with
just
those
you
know,
conversions
we're
talking
about
the
ROC
anyway,
and
so,
if.
C
D
D
C
A
Okay,
I
guess
that
works,
so
I
mean
okay,
so
conclusion
is:
if
we
can
pull
Jess
sis
out
and
the
Jas
module
out
into
its
own
crate.
We
should
do
that
because
it
just
seems
easier
right.
We
can
version
it
independently,
people
who
don't
need
it
don't
need
to
download
it
right
all
the
usual
benefits
and.
D
A
More
and
I
think
we
should.
We
should
do
like
a
once
over
of
just
like
all
the
bindings
and
make
sure
that,
like
we
aren't
returning
like
number
objects
anymore,
we're
returning
s
64
and,
like
I,
think
we
caught
all
those
but
like
there's
just
like
a
few
places
to
like
make
sure
we're
doing
the
right
thing
and,
like
I,
think
we
want
to
make
sure
we're
returning
like
Jess
string
and
not
strain,
which
I
also
think
the
way
I
deal
front.
End
should
change
to
do
that
as
well.
B
A
I
think
so
that's
an
advantage
of
using
number
objects,
I
think
but
I
don't
think
that
our
bindings
should
be
returning
number
objects
because
they're
not
actually
returning
number
objects,
they're
still
returning
normal
number
values,
and
just
because
we
change
what
our
import
says
doesn't
actually
make
it.
So.
A
A
A
number
object,
then
the
bindings
will
work
because
they're
not
like
dynamically
checked
unless
you'd
passed
debug,
but
it's,
but
it's
not
actually
going
to
call
like
a
new
number
and
turn
the
number
value
into
a
number
object
right
right.
So
it's
it's
like
like
whether
or
not
a
number
object
is
a
nice
thing
to
have
is
orthogonal
from
like
our
bindings
staying
the
correct
types.
A
B
A
B
B
A
D
D
A
D
C
D
D
A
A
Okay,
so
let's
talk
string
versus
Jazz
string,
I
think
that
we
should
always
be
returning
in
taking
yeah
returning
and
taking
J
strings
rather
than
rust,
strings
or
stir,
because
the
user
can
always
convert
a
string
into
a
Jo
string,
a
rust
string
into
a
Jo
string
and
there's
costs
associated
with
that
converting
and
so
I.
Don't
think
we
should
do
it
ourselves
and
it
it
seems,
like
other
than
organ
onyx.
There's
no
downside.
D
Yes,
the
main
thing
that
I
would
say,
which
is
with
the
I
as
far
as
I
know,
the
idea
with
host
bindings
is
that
you
don't
do
anything
like
this,
where,
if
I
API
says
it
takes
a
string,
it
takes
a
pointer
in
a
length
and
you
just
hand
it
down
to
lies
of
memory,
and
it
just
deals
with,
and
so
in.
The
future
host
bindings
were
gold.
As
far
as
I
understand,
I
haven't
actually
I
haven't
concerns.
D
We
wouldn't
we
wouldn't
J
asserting,
would
be
the
wrong
thing
to
pass
and
then
actual
pointers
about
the
right
thing
to
pass.
So
that's
where
I,
given
that
belief,
I'm
were
and
III
would
press
the
prefer
to
shoehorn
towards
that,
but
one
option
we
could
certainly
have
which
I
was
just
thinking
of
is
we
could
have
like
if
a
function
foo
takes
a
string
as
an
argument.
We
could
generate
a
foo
function
that
takes
an
answer.
Oh
and
and
a
JSON
underscore
foo
function
that
takes
a
gesture.
D
Basically
has
the
jsf
I'd
versions
of
all
the
arguments,
and
so
that
might
strike
the
right
balance
between
like
today.
If
you
want
really,
if
you
want
high-performance,
you
wanted
to
be
make
sure
you're
using
J
strings
locks,
you're,
not
traveling
the
heap
that
often,
but
you
also
have.
If
you
just
want
to
kind
of
really
quick
and
easy
things,
you
have
like
the
ability
to
just
kind
of
pest
in
there
yourself.
A
Like
just
function,
a
function
that
takes
a
s,
that's
like
s
is
some
sort
of
string,
and
then
we
have
a
trait
for
that
function.
That's
implemented
by
rust,
strings
and
jaea
strings
and
goes
out
to
the
appropriate
thing,
and
so,
like
you,
get
the
nice
ergonomics,
but
you
can
use
either,
but
that's
also
like
a
bit
more
than
we
are
usually
doing.
I
mean.
A
D
D
Just
cuz
it'll
take
time
for
this
to
actually
be
used
in
those
contexts,
and
so,
in
that
sense,
having
nice
ergonomics
might
be
a
little
better
for
for
kind
of
uptick
and
kind
of
actually
using
it
small
small
things,
but
yeah
I'll,
definitely
ping
Luke
and
ask
him
about
like
how
this
is
expected
to
work
in
his
bindings.
Whether.