►
From YouTube: 12 July 2018 Meeting
Description
The Rust WebAssembly Working Group meeting from 12 July 2018.
B
A
I
was
gonna,
say
one
thing
that
I
think
we've
found
is
well
at
least
it's
a
pattern.
I
see
in
rust.
Rfc's
is.
It
might
be
useful
for
some
of
us,
as
team
members,
to
write
our
comments
to
kind
of
like
get
the
conversation
going
and
then
tweet
it
so
maybe
aim
to
get
some
comments
in
today
and
then
tweet
it
tomorrow,
because
sometimes
when
people
start
commenting,
they
have
it's.
It's
like
difficult
to
know
exactly
what
to
chat
about.
So
it
can
give
people
more
to
say
if
there's
kind
of
already
a
combat
happening.
A
B
B
A
C
A
B
Okay,
see
I
like
since
we're
not
doing
that
much
stuff
actually
in
rust,
see
were
a
little
bit
less
affected
and
you
know
maybe
some
other
groups
are
but
the
the
second
preview
ships
in
early
August.
We
have
a
tracking
dashboard
and
then
basically
everything
that
we
want
for
to
ship
with
rust
2018.
We
have
this
new
blocking
rust,
2018
label
and
the
link
in
in
the
agenda
will
show
you
all
of
these
issues
labelled
this
across
the
whole
Russell
Assam
organization.
B
A
I
have
a
couple
comments
just
more
on
the
preview
stuff,
so
one
of
the
things
that
I
know
is
happening
is
they're
kicking
off
a
newsletter
to
help
talk
about
the
preview
and
the
work
that's
being
done
and
just
the
addition
in
general,
and
while
this
group
doesn't
have
a
ton
of
stuff,
that's
in
rust,
see
so
we're
not
super
beholden
to
the
deadlines
which
is
kind
of
Awesome.
We
might
consider
trying
to
ride
the
marketing
train.
They're
kicking
off
with
it.
B
A
Cool
so
I'm
on
the
hook
for
some
other
stuff
in
the
addition,
so
central
who
is
organizing
this
spoke
to
me
at
length
yesterday
about
all
this,
and
he
was
asking
me
about
some
of
my
other
projects
and
if
there
was
good
first
issues
and
I
was
like
not
yet,
but
for
us
there
definitely
is
so.
If
you
want
I
can
ping
him
or
you
can
ping
him,
but
I
think
there's
like
a
great
opportunity
for
us
to
get
a
little
bit
more
of
attention
or
eyes
on
on
this
stuff.
You
can
always
be
positive.
A
B
A
B
A
B
A
A
A
Okay,
so
the
1.0
question,
so
when
Nick
King
me
yesterday
about
blocking
for
2018
I
was
trying
to
think
about
like
what
do.
We
want
the
state
of
this
collection
of
tools
to
be
when
we
share,
and
so
inevitably
there's
the
question.
A
If
what
we're
trying
to
deliver
is
like
a
mature
like
developer
experience,
the
question
of
1.0
is
inevitably
comes
up
and
I
know
that
people
have
all
sorts
of
opinions
about
this
and
I
know
that
the
question
of
a
1.04
was
in
banja
and
it's
significantly
more
complicated
than
one
for
wise
impact,
but
I
wanted
to
get
folks
sense
of
how
and
how
how
important
of
a
signifier
is
a
1.0
and
do
we
think
that
we
could
1.0
was
impact
before
wasn't
Benja
and
given
that
wasn't
by
Drennen
is
proportion.
What
was
it
like?
A
D
D
Would
would
prefer
that
in
the
sense
that,
though
the
wizened
blend
stuff
is,
in
theory
like
to
me,
it
feels
1.0
when
it
includes
the
languages.
We
have
like
a
semi,
more
spec,
now,
I
think
in
the
meantime,
it's
all
it's
the
exact
same
stability.
You
have
it's
just.
It
will
change
in
the
future
once
we
actually
figure
that
stuff
out,
but
that's
all
like
internal
details
only
that
no
one
actually
has
to
care
about
so
in
that
sense,
I
do
think
that
we
should
strive
for
a
1.0
feeling
aspect
of
impact.
D
We're
like
the
idea
was
effect,
is
you're,
not
you're,
not
fiddling
with
all
the
whys
and
lines
and
pieces.
It
just
takes
care
of
everything
and
you
you
never
accidentally
use
a
debug
build
on
a
release
mode
stuff
like
that,
and
so
in
that
sense,
I
would
see
that
that's
pretty
important.
It's
pretty
critical
for
the
like
the
working
group
as
a
whole,
just
like
that's
very
central
to
our
experience,
is
having
a
really
good
experience
using
resin
pack
to
do
everything.
A
A
Like
that,
anyways
III
think
that
we
could
have
an
interface
that
we
consider
extremely
stable.
The
only
breaking
change
I
see
coming
down
the
pipeline
in
the
future.
It's
just
that
MPM
dependency
stuff,
and
that
would
be
like
because
currently
right
now
and
wasn't
binder
and
there's
something
landed,
that
I
think
we're
inevitably
going
to
end
up
changing,
but
I
haven't
actually
implemented
it
in
Westpac,
so
that
wouldn't
be
breaking
but
yeah
I
think
also
and
I'm
curious
what
other
people's
feels
are
on
this.
D
It's
not
entirely
that
and
the
sense
that,
like
cargo,
like
frees
us
from
worrying
about
that
sort
of
breakage,
for
like
my
build
randomly
broke
because
I
forgot
a
second
a
lock
file
or
something
like
that.
But
there's
still
like
they're
still
the
is
room
to
me.
It
depends
on
the
degree
of
breakage
where
it's
like,
if
nothing
works
from
before
now,
that
I
think
we
probably
don't
want
to
call
it
1.0.
But
if,
like
little
things
were
tweaked
and
there's
a
major
new
feature
which
required
breakage
and
like
that's
totally
fine.
D
So
like
it,
it's
kind
of
the
feeling
of
like
a--,
because
it's
not
just
that
people
like
things
will
not
break
if
they're
actually
version
correctly
and
using
like
they're
right
downloads,
and
all
that
it's
more
of
that.
We're
providing
a
very
strong
message
for
you:
do
need
to
upgrade.
You
do
need
to
change
and
like
we
as
a
community
are
going
to
start
using
the
new
thing.
We're
gonna
stop
using
the
old
thing,
and
so
then
it's
if
we're
requiring
everyone
to
make
the
transition
house
yeah.
A
I
totally
agree
about
the
upgrade
path
and
like
huge
breaking
changes
like
just
for
an
example.
The
breaking
change
I'm
thinking
about
right
now
is
like
in
the
og
days
of
Wes
impact
when
we
had
kind
of
a
smaller
scope
for
it,
I
kind
of
jammed
everything
in
the
init
commands
and
like
now.
The
unit
command
does
like
5,000
things
and
it
like
really
doesn't
mean
to
so
I'm
gonna
break
out
functionality
and
to
in
it
and
build,
but
I
I
can't.
A
A
B
B
This
this
reminds
me
you
have
actually
one
other
point,
which
is
that
I
think
almost
more
important
than
any
actual
changes
to
Azzam
pack
itself
for
wasm
pack.
Feeling
1.0
is
the
integration
into
all
of
our
tutorials
and
rest
in
webassembly
book,
and
all
that
and
and
so
I
don't
know
if
anyone
looked
at
the
actual
blocking
rest
2018
issues.
But
there's
a
couple
in
the
book
related
to
as
impact
that
I
marked
blocking.
A
It's
worth
one
of
the
like:
we
have
a
the
next
release,
which
is
oh
five
is
gonna.
Do
that
in
it
build
split,
and
then
it
was
my
goal
to
kind
of
turn
towards
the
dock
stuff,
because
I
haven't
I
haven't
touched
any
of
those
docks,
but
I
could
definitely
help.
Do
that.
So
I
want
to
get
the
refactor
done.
First,.
A
Like
no
apologies
all
circle
up
with
Alice,
also
nano
sorry
I,
just
realized
I'm
using
multiple
names
for
one
person
to
see
where
her
work
on
the
tutorial
is
because
I've
also
been
kind
of
holding
off
touching
it
because
I
know
she's
doing
a
really
big
refactor
and
so
merge
conflicts
trying
to
avoid
that.
But
it
all
circle
up
with
her
this
hopefully
today
and
make
sure
that
we
can
get
that
sorted.
A
Actually,
I
realized
it's
not
on
the
agenda,
but
I
am
kind
of
curious.
We've
been
talking
about
like
1.0
instability,
I'd
love
to
hear
from
Sven.
If
you
think
I
mean
1.0
is
not
really
how
you'd
label
it,
but
do
you
think,
like
bundler
support
stabilizing
in
any
of
the
butler's,
is
like
on
the
roadmap
for
thoughts
around
that.
F
I
mean
I
mean
so
they
will
press
the
web.
The
web
back.
Integration
is
quite
stable,
yet
but
I
mean
I,
don't
feel
like
we
should
label
it
1.0,
even
if
it's
stable,
because
it's
independent
anyway.
A
Yeah
I
guess
I'd,
but
I'm,
mostly
meant
to
say,
is
we're
marking
stuff
is
blocking
on
2018
in
the
stuff.
That's
in
our
repo
that
we
can
control
but
like
the
bundler
story
is
equally
important
and
I
don't
know
if
we
have
a
great
sense
of
how
we're
tracking
that
I
don't
know
if
it's
Jen,
if
you
had
thoughts
because
they're
not
really
in
the
repo,
so
we
can't
super
tag
them
or
whatever
yeah.
B
Alright,
let's
jump
into
goal
updates
so
web
engine
stuff,
we
landed,
the
websites
create
the
skeleton
of
it.
So
this
is
the
crate
that
is
mechanically
generated
from
web
IDL
and
there's
a
ton
of
issues
of
like
enabling
generation
of
different
web
IDL
interfaces,
because
right
now
we're
opting
into
generating
bindings
for
certain
interfaces
and
basically
there's
no
so
Court
there,
it's
very
thin.
B
A
Was
muted
I
feel
behind
this
week,
but
I'm
hoping
that
by
the
rest
of
today
and
tomorrow
it
should
be
going
pretty
quick.
B
Okay,
do
you
need
like
some
extra
help
on
issues
or
anything
Eric,
one.
A
Thing
that
would
would
help
make
it
able
make
us
able
to
move
a
little
faster
is
because
of
some
funkiness
and
a
cert
CLI,
which
is
apparently
changed,
and
it's
like
way
better.
Now
we
like
could
really
use
someone
who
really
wants
to
write
integration
tests
because
right
now,
I'm
like
moving
a
lot
more
cautiously,
because
I
don't
completely
trust
our
test
suite,
and
so
that's
slowing
me
down
a
little
bit.
So
if
you
happen
to
know,
somebody
who
loves
doing
testing
it'd
be
great.
A
I
know
that
--get
OC
was
working
on
it,
but
now
he's
gonna
be
gone
for
like
two
weeks,
so
I
think
it
I
think
the
testing
really
needs
to
land
405,
but
there
was
like
a
small
regression
and
so
I
have
to
do
a
patch
release
today,
which
sucks
and
so
yeah.
That's
that's
the
one
thing
that
I
think
would
be
really
beneficial
for
all
of
the
blocking
2018
stuff,
like
there's
already
PRS
for
basically
all
of
them,
and
so
it's
really
just
a
matter
of
me.
B
F
Ideally
I
would
like
to
add
more
integration
into
pass
alien
and
world
up
and
I
feel
like
a
bit
behind,
because
I
would
like
to
have
the
rest
bragging
a
bit
more
to
have
a
bit
more
functionalities.
It
has
not,
and
I
haven't,
had
the
time
to
dig
into
making
wasn't
by
gen
plugin
for
web
bike
yet,
but
there
will
be
super
cool
laughing.
B
Impact,
so
my
understanding
was
that
the
the
way
that
we
were
moving
was
that
Wes
impact
would
kind
of,
like
orchestrate
all
of
the
rust
side
of
the
the
build,
and
then
it
would
produce
this.
This
final
azzam,
the
web
pack,
would
then
it
would
hand
off
to
to
web
pack
or
like
web
pack,
would
invoke
wise
impact
to
get
that
and
then
like
fit
into
the
the
larger
bundler
and
and
the
dev
server
and
and
stuff
like
that,
and
and
from
that
perspective.
A
Just
a
root
to
expand
a
little
bit
that
I
think
our
goal
is
that
Voisine
pack,
especially
with
the
this
kind
of
like
three
build
modes
thing
that
we're
coming
up
with,
is
like
the
convention
path
and
then
like.
If
you
really
have
specific
customizable
stuff
that
you
like
really
want
to
do
on
your
own,
then
you
can
invoke
wise
and
bind
Jen
on
your
own.
F
A
So
those
there
are
PRS
open
for
that
right
now
and
that's
expected
to
land
in
oh
five,
which
is
the
next
release
or
at
the
very
latest.
The
release
after
that,
both
of
like
should
be
coming.
Pretty
quick,
like
the
PRS,
are
literally
already
open
and
mostly
dependent
on
us
kind
of
spec'ing
out
a
configuration
thing
to
add
the
cargo
tom.
Oh
okay,.
B
A
E
With
contributing
markdowns
that
I
had
proposed
and
that's
not
the
status
as
far
as
I've
heard
from
it,
I'm
gonna
get
back
to
Alice,
maybe
at
some
point
in
the
next
couple
of
days
to
find
out
if
she's
made
any
progress
on
the
stuff
that
she's
working
on.
But
my
sense
from
the
last
time
that
I
talked
to
her
was
that
she's
focused
on
that
refactor
of
the
book
tutorial
right
now.
B
B
Deal
is
not
here,
do
you
have
an
update
Ashley
on.
A
The
website
so
sindelle
sent
me
a
screen
cap
of
some
of
the
work
that
he's
done
so
far,
but
it
hasn't
been
filed
as
a
PR.
Yet
one
thing
that
I
think
is
very
interesting
is
that
he
lists
a
lot
of
numbers
in
the
page
and
I'm
intrigued
by
the
idea
but
I'm
worried
just
because
they
definitely
look
like
benchmarks
and
hosts
our
last
conversation
about
benchmarks.
I
we
we
don't
actually
have
any
so
I
would
definitely
want
to
hesitate
on
publishing
any
numbers
on
that
page.
B
B
A
B
I
just
turned
off
the
video
and
then
I
don't
know
it's
kind
of
coming
in
and
out,
but
yeah.
So
the
the
rest
of
the
the
goals
I
think
are
still
on
track.
So
that's
debugging
and
the
template.
The
template
has
some
more
issues.
We
need
to
finish
before
Rus
2018,
mostly
again
related
to
integrating
wasm
pack,
but
it
should
be
pretty
straightforward.
I
think
I
can.
A
A
F
Sorry
actually
I,
agree
and
I
think
we
talked
about
that
two
weeks
ago
was
abandoned
and
as
far
as
I
remember,
we
talked
about
having
two
templates,
one
that
ends
for
a
library
and
one
that
does
only
web
stuff
where
in
the
web
stuff,
you
only
have
what
I
can
in
the
library
I
don't
have
any
J
s
defenses
so
I
have
one
question:
would
the
cargo
Jam
generates
only
one
template?
Okay,
we
have
two
two
templates
so.
A
A
That
being
said,
like
I,
don't
like
I
guess
it's
possible
that
we
could
have
it
pull
down
both
I'd
want
to
like
talk
more
about
exactly
what
that
workflow
is
I.
Think
in
my
head
right
now.
The
way
I
would
imagine
it
is
I
want
to
create
a
wasm
library,
so
I
run
cargo
generate
and
use
the
like
ways:
impact
Lib,
template
and
I
make
my
library,
I,
publish
it,
and
then
I
use
cargo
generate
to
generate
a
wiser
pack
app
and
then
go
there.
Go
from
there.
A
I
don't
know
if
having
cargo
generate,
create
two
projects
for
you
at
the
same
time
make
sense,
but
I
also
do
not
have
strong
feelings
on
this
and
I'm
super
open
to
like
people
slots
everything
just
just
clones
down
one
repo
and
does
base
substitutions
across
any
of
the
files
you
wanted
to
do.
A
substitution
sauce
or.
F
Maybe
have
a
small
suggestion,
so
let's
say
you
want
to
do
a
libraries
or
you
use
the
cargo
gen
thing
and
if
you
want
to
build
a
web
application,
you
probably
also
have
like
NPM
empty
MPX
installed
locally,
so
you
could
use
the
NPM
creates
I.
Think
all
right,
NPM,
MPX,
great
I,
don't
remember
the
name,
but
they
could.
It
may
be
separated
both
in
two
separate
terms.
Yeah.
A
I
actually
think
it
makes
a
lot
of
sense
to
make
the
web,
as
in
pack
app
template
like
an
NPM
template
and
not
a
part
of
template.
I
think
that
makes
a
ton
of
sense
because
in
theory
a
Wesen
pack
app
should
not
need
to
know
rust
at
all
so
yeah.
That's
a
super
great
idea
and
I
also
forget
the
key
word,
but
I
think
it
might
be
create
because
it
was
based
on
create
react,
app,
I
think
but
yeah
I
can
investigate
that.
F
A
A
A
So,
given
the
rust
wasm
interest
in
debugging,
I
have
not
worked
terribly
closely
on
the
debugging
stories,
so
I'm-
probably
not
a
good
representative
for
that,
but
it
might
be
useful
to
have
someone
from
this
group
join
that
maybe
Fitz
gen,
maybe
someone
else
but
I.
The
community
group
is
open
to
join,
like
you
literally
just
go
to
the
w3
website
and
say
like
add
me,
but
then
in
order
to
get
invited
to
the
meetings
you
have
to
like
send
an
email
to
some
people
full
of
confusing.
A
B
B
My
internet,
yeah,
yeah
and
I'm
definitely
gonna
be
involved
at
a
bugging
group.
I
have
been
involved
in
all
the
threads
leading
up
to
it
and
commented
on
the
PO
requests
for
the
sub
charter,
I'm
just
waiting
for
it
to
actually
be
created
before
diving
into
their
meetings
and
stuff
cuz
I'm
not
really
interested
in
the
larger
community
group
meetings.
Yeah.
A
D
Away
sure
this
is
Bill,
it
would
be
on
the
status
updates.
This
is
kind
of
just
like
FYI
stop
business,
but
there's
a
couple
of
stuff
happening
on
the
the
toolchain
side
of
things
in
rust
itself.
So
there's
the
first
two
of
there
are
attributes
that
are
currently
reasons.
Why
or
reasons
why
wasn't
mine
done
lightly,
but
doesn't
necessarily
need
to
be,
and
so,
let's
just
heads-up
if
anyone's
curious
to
check
those
out.
D
So
basically
one
has
to
learning
important
namespaces
and
rust
doesn't
inherently
have
a
way
to
do
one
of
them
and
then
the
other
one
is
custom
sections
just
how
do
we
actually
define
that
in
rust?
Otherwise,
the
nut
to
today's
nightly,
but
tomorrow's
nightly
should
have
elevated,
and
this
is
a
really
big
thing,
for
us,
at
least
it
was
know,
has
some
space
in
the
sense
that
we
had
a
lot
of
features
from
LOD
now,
the
main
one
being
that
you
no
longer
have
to
use
LTO
to
get
small
binders
by
default.
D
So
this
is
something
where,
if
you
type
purple
just
release,
everything
will
get
GC
appropriately
is
the
terminology,
but
in
any
case,
just
keep
an
eye
out
for
regressions
or
anything
funky
looks
going
on
like
this
might
be
related
to
that,
and
otherwise
your
wiesen
binary,
your
build
should
be
faster.
Your
binary
should
be
smaller
as
a
result,
so
just
look
out
and
see
what's
going
on
there,
that's
pretty
much.
It.
A
A
A
A
A
That'd
be
super
useful
I
know
that
this
week
in
rust
does
a
whole
bunch
of
shenanigans
using
labels,
and
it
helps
automate
a
ton
of
the
content.
I'll
follow
up
with
you,
Fitz,
Jen
and
type
it
out,
but
I,
don't
know.
My
general
vibe
is
like
yes,
I,
think
automating
stuff
using
labels
as
long
as
that's
not
the
only
thing
we're
doing
some
stuff
won't
fit
with
labels.
I
think
it's
a
good
idea.
Anyone
else
have
any
vibes
on
that.