►
From YouTube: wasmCloud: S3 Blobstore Provider, Interface Evolution Discussion, Community Updates - 03/09/22
Description
wasmCloud is a platform for writing portable business logic that can run anywhere from the edge to the cloud, that boasts a secure-by-default, boilerplate-free developer experience with rapid feedback loop.
https://wasmcloud.com
A
All
right
welcome
to
awesome
cloud
wednesday.
For
march
the
9th
2022.
we've
got
a
lot
of
fun
stuff
to
show
today,
and
I've
got
a
call
coming
in
just
a
second.
I
send
somebody
to
voicemail
all
right.
Hilarious
couldn't
plan
that
any
better.
I'm
gonna
go
and
start
with
a
quick
look
at
our
new
training
stuff
that
we're
getting
ready
to
release
sort
of
like
an
early
preview.
Oh
jordan,
you're
on!
Do
you
want
to
do
the
demo
of
the
other
stuff
through
the
walkthrough.
B
Yeah
sure
give
me
me
30
seconds,
I'm
running
I've
lost
my
mind
today.
That's
okay!
I
hope
you
find
it.
It's
me
too.
A
All
right
so
today,
so
we're
gonna
start
with
a
demo
of
the
new
training
platform
we're
getting
ready
to
launch,
which
I'm
super
excited
about
sort
of
what
we
always
hoped
we
would
have
with
katakota,
and
then
steve's
got
a
great
demo
for
us
today
around
some
of
the
new
capability
providers
that
we
have
out
a
new
demo
that
we're
releasing
out
to
the
community
as
a
poc.
B
C
B
B
Updating
I
need
to
do
it
still
references
a
bunch
of
the
pre
.50
things,
but
we're
getting
there
very
quickly
and
in
a
matter
of
days
we're
going
to
have
you
know
the
whole
thing
updated
like
there's
more
rebels,
so
things
like
that
have
to
go
away,
but
one
of
the
great
things
that
we've
gotten
from
this
is
one
of
the
hard
the
hard
points
early
on
was
it
took
like
five
minutes
to
bootstrap
this.
Luckily,
the
way
this
platform
handles
things
we
get
the
whole.
B
You
know
environment,
bootstrapped
and
you
know
just
under
30
seconds,
and
that
includes
like
wash.
It
includes
wasm
cloud.
It
includes
gnats
and
all
you
know
everything
you
need
to
be
a
happy
developer
and
yeah
and,
and
the
next
great
thing
about
this
is
after
I
do
a
little
bit
of
magic
with
the
github
actions.
This
will
be
another
open
source
repo
into
the
wazon.org.
B
That
anyone's
going
to
be
able
to
you
know,
contribute
to
right,
because
this
is
a
github-based
type
thing
and
a
few
of
the
things
you
know
we
we
see,
you
know
we
see
wash.
You
know
it's
a
very
familiar
feeling
terminal
very
familiar
filling
instructions.
The
washboard
looks
great
in
it.
It's
fully,
you
know
functional
so
yeah.
This
is
this
is
where
we're
at
and
as
again
right
now
we're
at
the
point
where
we've
done
a
direct
port.
B
It
took
me
a
few
days
to
get
the
actual
functionality
a
few
days,
a
few
weeks
to
get
the
actual
functionality
working,
but
now
we're
at
the
point
where
we
can
iterate.
We
can
iterate
fast
on
content
and
pretty
much
what
this
is
going
to
look
like
soon
is.
You
will
be
able
to
go
to
just
the
cosmonic
page
and
we
will
have
a
full
wasm
cloud
topic
here
with
all
the
fun
stuff
that
we
can.
B
You
know
figure
out
what
we
want
to
help
train
people
on,
and
I
assume
at
some
point
we
will
have.
We
will
have
to
repo
up
and
we'll
start.
B
So
I
I
don't
want
to
commit
to
any
hard
time
lines,
but
I
would
say
that
we
will
be
in
a
place
where
we
can
start
testing
these
and
filling
these
out,
probably
in
the
next
week
or
so,
but
yeah.
That's
really
sorry
that
was
a
little
ad
hoc.
I
didn't
have
my
thoughts
yeah
online.
A
No,
absolutely
not
jordan.
Thank
you
so
much
and
you
had
so
many
criteria
coming
into
this,
that
it
could
be
embeddable
editable
by
the
community
that
you
could.
You
know,
do
the
whole
experience,
including
the
ui.
I
write
in
the
web
browser
and
I
really
want
to
say
thank
you
so
much
for
all
the
hard
work
you
put
into
this.
I
know
you
looked
at.
You
know
a
bunch
of
different
platforms
and
I
really
did
a
lot
of
work
behind
the
scenes
here.
A
I
think
this
is
going
to
be
a
huge
accelerant
that
we
can
build
out,
take
our
demos
and
examples
and
things
like
the
pet
clinic
and
really
give
people
a
whole
experience.
That's
completely
frictionless
to
help
them
learn
about
our
community
and
what
we're
doing
so.
This
is
an
incredible
effort
and
thank
you
so
much
for
all
the
hard
work.
Jordan.
Absolutely
that's
great.
Any
questions
for
jordan.
A
Super
steve:
are
you
ready
for
your
demo?
Do
you
want
me
to
go
through
a
couple
of
announcements
and
things
like
that
before
we
go
on.
E
I'm
good
I'm
good.
We
can
start
let's.
A
Hand
it
off
to
you
for
your
demo.
We
love
starting
the
meetings
with
demos.
I
think
they
keep
everybody
engaged
and
interested,
and
people
always
love
to
see
what's
going
on
now.
This
is
a
bunch
of
work,
that's
probably
three
or
four
threads
that
are
all
being
brought
together.
You
know
for
kind
of
one
presentation
here,
so
I'm
excited
about
this
one
go
ahead.
That's
great!
I'm
excited.
A
E
Aren't
you
which
steve
I'm
lowercase
steve
who's.
A
A
E
All
right,
let's
see,
let
me
oops
sorry,
let
me
disappear
this.
E
Okay,
first
thing
I'll
talk
about
is
over
here
on
the
right,
so
we
have
a
new
blob
store,
s3
capability
provider,
so
there's
a
new
interface
in
the
interfaces,
repo
that
describes
a
blob
store.
E
So
most
people
are
familiar
with
the
key
value
store
which
our
main
implementation
of
that
is
backed
by
redis
and
keys
and
values
are
strings
or
serialized
structure.
Blobs
are
sort
of
intended
to
be
a
little
bigger.
They
have
a
concept
of
a
container
which,
for
s3,
is
a
bucket,
and
so
there
the
interface
has
operations
like
you
can
list
all
the
buckets
you
can
create
and
delete
buckets.
E
You
can
create
and
delete
objects
within
them,
and
we
also
have
a
contribution
pending
from
the
community
of
a
different
implementation
of
blob
store
that
uses
a
file
system,
and
that's
that's
a
pending
pr.
Now
this
one
here,
it's
not
in
published
to
our
oci,
the
azure
cr.
So
today,
if
you
want
to
grab
this,
you
have
to
build
from
source,
but
within
a
day
or
two
we'll
get
that
up
on
azure.
E
There
are
a
lot
of
ways
to
configure
your
aws
credentials.
You
can
do
it
through
environment
or,
if
you're
running.
If
you
happen
to
be
running
in
kubernetes,
then
it
can
pick
up
the
ec2.
E
Sorry,
if
you're
running
in
ec2,
you
can
pick
up
the
ec2
credentials.
So
that's
regardless
of
kubernetes
or
not
so
a
lot
of
ways
to
configure
that
and
then
there's
an
app
actor.
That
is
in
the
examples,
repo
pr
it's
about
to
merge
that
uses
this.
It
uses
the
blob
store
as
the
back
end
and
the
http
server
capability
provider.
E
So
when
you
send
it
a
image
over
http,
it
will
upload
it
to
the
blob
store.
So
I'll
give
you
a
quick
demo
of
that.
I
have
a
couple
of
images
here
and
I
have
the
wasm
cloud
running
on
this
machine.
E
Images
is
a
command
that
it
responds
to
that
lists.
All
of
the
current
images
in
the
bucket
that
we've
designated
for
images
and
you
can
change
which
bucket
it
goes
to
through
your
link
definition.
E
E
It
also
has
a
little
bit
of
code
to
try
to
determine
the
image
type
and
sets
sets
the
mime
type
and
normalizes
the
extension
for
that.
So
it
also
avoids
uploading
duplicate
files,
since
sha256
is
effectively
collision
free.
So
I'll
just
show
you
an
example
of.
E
To
upload
an
image,
let's
say
upload
this
large
bird,
which
is
a.
E
And
after
it
does
the
post,
it
returns
a
little
bit
of
json
information,
here's
the
shot
of
the
file
it
uploaded,
the
original
name
and
the
new
name
where
it's
being
stored
in
the
bucket.
E
If
you
do
that
again,
it's
returned
faster,
but
I
got
the
same
result
because
it
it
computed
the
sha
and
realized
it
had
already
been
uploaded.
So
it
doesn't
bother
uploading
it
to
s3
again.
E
So
this
is
a
kind
of
a
cool
demo
that
you
can
take
to
use
for
all
sorts
of
image
processing.
We
may
do
some
variations
of
this
for
some
machine
learning
demos
as
well.
So
I
wanted
to
talk
about
this
file
size
here.
You
might
remember
that
the
default
mask
configuration
has
a
limit
of
one
megabyte
for
a
nat's
message
and
although
you
can
configure
that
to
be
a
little
bit
larger,
that
nat's
team
recommends
that
you,
you
don't
exceed
two
megabytes.
E
You
don't
exceed
very
many
megabytes,
because
the
performance
can
really
degrade.
E
We
have
implemented
a
big
effort
from
kevin
and
the
team,
transparent
chunking
of
messages,
so
in
any
api
now
in
wasmcloud,
if
you
send
a
message,
that's
larger
than
a
threshold,
we
will
break
it
up
and
into
chunks
and
send
it
over
nats.
This
uses
a
really
cool
feature
of
matt's
jet
stream.
That
has
some
built-in
object
stores
and
it
can
resort
the
packets
to
to
be
in
order,
and
then
we
reassemble
them
on
the
other
side.
E
And
then
the
receiver
receives
the
message,
and
so
that
should
allow
you
to
send
messages,
at
least
in
this
several
hundred
megabytes
now,
and
so
we,
we
don't
believe
that
wasn't
cloud
users
should
have
to
worry
about
the
nat's
message
limitation
and
it
was
really
the
this
implementation
of
the
blob
store.
That
made
us
realize
that
this
is
a
a
really
important
problem
we
need
to
address
now.
So
that's
a
that's.
A
real,
exciting
capability.
A
All
right,
steve,
so,
to
recap,
there's
a
lot
that
we
just
demoed
today:
the
new
blob
store,
s3
provider,
transparent,
chunking,
a
new
example
in
the
repo
that
ties
the
together.
Oh
and
the
the
ability
to
use
the
I
am
sds,
assumed
role,
functionality
that
is
on
ec2
and
transparently
available
in
kubernetes
through
stuff
like
cube2im.
E
Big
summary
checklist
here.
Yes,
yes
thanks,
I
forgot
to
mention
the
ias,
assume
assume
role
if
you're
using
sts,
which
some
enterprises
do.
A
Yeah,
that
is,
that
is
a
ton
of
stuff.
You
know
that
is,
you
know,
really
enterprise.
You
know
level
functionality
and
you
know
not
too
transparently.
You
know
this
was
driven
by.
You
know
real
users
out
in
the
field
that
we
have
on
the
enterprise
side
that
you
know
needed
to
check
some
corporate
security
boxes
in
order
to
get
things
into
production.
So
it's
amazing
to
see
all
this
come
together
and
I'd
love
people
to
test
this
out
and
give
us
some
feedback
on
this
demo.
A
I
think
you
really
went
the
extra
mile
if
you
also
did
the
mime
checking
on
the
files
and
the
shot
256
is
cool
and
that
and
to
point
out
that
is
that
calculated
in
a
streaming
fashion.
So
you
get.
You
know
a
little
bit
of
the
file.
You
calculate
a
little
bit
of
the
hash.
F
A
E
Does
it
does
do
the
hash
within
wasm
cloud
before
it
sends
to
s3?
So
you
avoid
the
uploading
the
bits
to
aws.
There
is
one
more
capability,
new
capability
provider
that
I
don't
know
if
you
announced
on
last
week's
call,
but
there's
another
implementation
of
the
key
value
which
is
backed
by
hashicorp
vault,
and
you
can
use
it
for
storing
secrets.
A
A
But
that's
another
one
that
we
should
get
a
demo
checked
in,
for
this
is
phenomenal,
open
discussion.
Anybody
have
questions
on
this.
A
All
right:
well,
I
guess
that
there's
no
question
steve!
Thank
you
so
much
for
taking
the
time
to
walk
us
through
all
that
I'm
excited
to
see
where
all
that's
moving.
I've
just
got
a
handful
of
small
things
to
share
this
morning.
A
A
If
you
go
right
to
redbadger.com,
you
can
find
find
a
link
to
it
and
then
second,
the
wasm
cloud
crew
will
already
be
in
europe
for
kubecon
in
may,
but
we
may
also
be
joining
our
friends
in
london
for
q
con
next
month,
so
we
got
kind
of
a
last-minute
ask
to
see
if
we
could
surge
in
and
do
a
presentation
I'll,
let
you
guys
know
as
soon
as
we
get
anything
confirmed
there
just
open
floor.
Are
there
any
other
community
announcements
that
people
want
to
make
or
share
today
about
stuff?
A
That's
going
on
around
around
the
webassembly
world.
D
I
do
and
I
was
gonna
share,
but
you
all
know
how
that
goes.
D
Who
knows
well
yeah,
so
I
just
wanted
to
share,
and
maybe
taylor
was
going
to
talk
about
the
same
thing,
but
not
only
will
we
be
in
europe
but
taylor
and
I
will
be
doing
a
talk
at
cubecon.
We
got
our
proposal
accepted,
which
is
very,
very
exciting,
we'll
be
talking
about.
Let's
see
it's
on
wednesday,
we'll
be
talking
about
running
webassembly
and
wasmcloud
during
like
the
main
kubecon
conference.
So
we're
pretty
excited
to
do
this.
A
I
like
how,
on
your
on
your
background,
like
it,
doesn't
get
just
the
space
between
your
headphones
and
your
head,
so
it
kind
of
looks
like
a
halo.
A
A
So
what
I
was
gonna
say
is:
there's
something
we
had
neglected
to
share
the
past
two
or
three
weeks
of
some
changes
that
landed
in
the
last
bug
fix
release
zero,
fifty
two
four:
it's
not
the
right
number
and
what
we
discovered
is
it's
just
an
import
performance
enhancement
and
I
did
not
come
prepared
to
demo
this,
and
it's
really
not
that
impressive.
A
So
when
you
ran
on
constrained
devices
with
maybe
half
a
gig
or
a
gig
of
memory
that
had
just
a
little
bit
less
than
your
normal,
like
servers,
it
would
spike
and
you
would
completely
get
killed
or
die
because
you'd
run
out
of
a
memory
when
you're
starting
a
provider,
and
so
we
did
some
digging
and
stuff
and
and
some
memory
profiling
and
found
out
that
some
things
are
getting
held
onto.
So
we
actually
introduced
some
changes
that
made
it
so
that
those
things
never
get
written
into
memory.
A
You
ended
up
with
like
a
double
or
triple
copy
of
the
provider
in
memory
which
was
super
weird
and
so
now
that's
fixed.
So
if
you're
running
the
latest
version,
you'll
notice
that
when
you
start
a
provider,
your
memory
usage
goes
up
by
like
it
was
like
a
megabyte.
I
think
when
you
start
a
provider,
it's
pretty
nuts.
A
So
in
addition
to
the
actual
memory
that
the
the
operating
system
allocates
to
the
provider,
the
only
thing
you're
using
to
actually
run
it
is
like
another
like
less
than
a
meg
of
about
a
mag
or
less
than
a
meg
of
memory.
So
that
was
just
something
we
forgot
to
point
out
that
I'm
not
sure
many
people
were
running
into
yet
we
ran
into
it
cosmonic
because
we
are
running
at
scale
and
on
constrained
devices,
and
so
we
ran
into
that.
A
A
We
spun
up
like
four
or
five
different
actors
or
four
or
five
different
providers
and
like
a
thousand
actors,
and
we
had
only
used
like
100
mags
of
memory,
200
megs
of
memory-
it
was
pretty
pretty
insane.
So
that
was
just
something
that
we
forgot
to
share.
That
happened
and
some
performance
tweaks
that
occurred.
A
That's
awesome,
taylor.
That's
another
great
one!
You
know
we
should.
You
know,
try
to
do
a
better
job
of
curating
these
together
in
a
way
that
they're
more
accessible.
But
I
do
really
appreciate
the
core
group
of
folks
that
comes
here
as
well,
as
the
you
know,
50
to
100
folks
that
usually
watch
us
week
to
week
on
youtube
for
you
staying
up
with
our
progress
and
what
we're
working
on
and
what's
important
to
us.
So
we
already
did
some
community
announcements
open
floor.
F
But
there's
a
draft
vr
on
the
interfaces,
repo
that,
if
you
pulled
up,
maybe
we
could
chat
about.
A
F
Yeah,
so
forgive
the
background
noise,
but
my
question
is:
can
we
have
situations
like
this
pr,
where
different
implementations
of
a
provider
may.
A
Okay,
kevin
steve,
I
don't
know
it
would
invite
one
of
the
two
of
you
to
maybe
comment
about.
You
know
the
different
perspectives
on
dialing
in
an
interface.
E
Yeah,
I
I
think
I
I
think
I
understand
the
question.
So
the
question
is:
how
do
we,
how
do
we
evolve
interfaces
and
either
in
non-breaking
ways
or
breaking
ways?
Is
that
what
you're
asking
about.
E
There
are
some,
so
that's
that's
kind
of
an
involved
question
and
we
need
to
spend
some
time
to
document
ways
that
an
interface
can
involve
there
are,
there
are
non-breaking
ways:
interface
can
evolve
and
and
breaking
ways
and
right
now,
there's
not
a
a
formal
interface
version
number
that's
codified
in
and
so,
which
version
the
interface
you
get
would
if
it
changed
significantly,
it
would
depend
on
when
you
actually
compiled
it
and
that's
not
a
not
a
great
thing.
E
So
what
we
so
this
is
probably
this
is
probably
on
us
to
document
what
to
do
if
there's
a
breaking
change.
Well,
first
of
all,
if
there's
a
new
function
added
the
because
of
the
way
we're
using
traits
in
rust,
it
would
require
adding
that
new
function,
even
if
it's
a
stub
that
throws
an
error
like
a
not
implemented
error.
E
If
it's
not
implemented,
that
would
still
require
changes
to
everybody.
Who's
implemented
that
interface,
so
that's
kind
of
a
hurdle
for
adding
a
method
right
now,
and
that
again
is
something
we'd
like
to
be
able
to
make
non
non-breaking.
Probably,
what's
going
to
happen
is
we'll
have
to
change
the
name
of
the
interface,
so
instead
of
wasm
cloud
colon
key
value,
it
might
be
wasmcloud
colon,
key
value,
colon
v2,
something
that
that
effectively
makes
it
another
interface,
because
you
never
want
to
accidentally
link
an
actor
to
a
provider.
E
If
it
wasn't
signed
to
say
it
knows
that
version
of
the
interface
so,
but
we
do
need
to
so
that
we
don't
get
an
explosion
of
incompatible
versions,
and
you
know
we
certainly
don't
want
to
get
to
like
the
dll
hell
of
the
the
old
windows
days.
So
we
want
to
figure
out
a
way
to
do
it,
where
it
doesn't
create
a
infinite
number
of
versions
to
manage.
F
Yeah,
that's
actually,
I
had
exactly
that
question
too
about
when
you
know
one
implementation
of
a
provider
may
not
actually
support,
or
should
it
be
able
to
not
actually
support
every
operation
and
brooks
to
answer
your
question
in
the
chat.
Yes,
I
had
that
question
too,
which
is
that
you
know.
Is
there
any
approach
in
the
specific
scenario
where
dynamo
and
redis
and
provider
xyz
for
key
value
may
slightly
differ
in
their
apis
right
and
we
want
to
provide
a
single
api
from
the
perspective
of
the
interface
itself,
yeah.
E
I
know
kevin
has
his
hand
up,
and
I
know
he's
got
a
lot
to
add
on
this
on
this
topic.
I
just
want.
C
E
That
the
vault
key
value
provider
does
not
implement
all
of
the
methods
in
the
key
value
interface,
and
you
can
see
in
the
readme
for
that
provider
that
some
of
them
are
just
going
to
return
an
error
because
it
doesn't
make
it
there's
no
list
operation.
That
makes
sense
for
a
hashicorp
fault.
So
so
that's
that
that's
one
way
to
handle
it,
but
there
are
others
but
go
ahead.
Kevin.
C
When
you
attempt
to
call
that
right,
I
think
the
other
thing
is
there's
a
philosophical
debate
that
I
don't
think
we've
had
yet
that
revolves
around
whether
or
not
a
new
version
of
an
interface
should
constitute
a
new
interface,
because
you
know
put
another
way,
is
v2
of
the
key
value
interface,
something
that
needs
to
be
signed
on
on
an
actor
separately
than
the
one
of
the
interface
and
there
are.
C
C
You
know
how
how
the
interface
has
evolved,
and
I
think
part
of
the
real
problem
is
there's
no
universal
answer
in
terms
of
you
know,
because
somebody
could
actually
create
a
new
version
of
an
interface
that
is
such
a
big
deal,
that
it
does
need
to
be
a
separate
signature
or
a
separate
claim,
but
then
something
like
adding
an
optional
feature
like
the
scan
operation
yeah.
Does
that
really
need
another
interface
name?
Or
is
that
really,
just
like
you
know,
an
incremental.
C
I
mean
it
is:
it
is
technically
non-breaking
right
I
mean
you
could
add
this
operation
to
an
interface
and
you
could
have
multiple
actors
floating
around
in
a
lattice,
some
of
whom
are
aware
of
this
operation,
some
of
whom
aren't
and
all
would
be
well
in
the
world.
There
would
be
no
conflict
so
yeah,
like
I
said
it's
it's
nuanced
and
there's
probably
a
lot
of
discussion
that
I
think
we
need
to
have
over
the
the
provider.
Revisioning
story.
F
Cool
well,
I
hope
it
was
a
good
if
this
is
the
right
form
to
bring
up
these
types
of
questions
despite
the
screaming
child.
In
the
background,
yeah.
A
No
matt
look,
and
I
think
that
this
is
one
of
the
real
strengths
of
wasm
cloud
and
one
of
the
things
that
I
really
appreciate
about
our
design.
I
really
think
that
all
truly
scalable
software
is
at
its
core,
very
pluggable,
and
I
think
what
I
like
about
what
we're
doing
here
as
a
feature
is,
is
ultimately
users
can
make
their
own
decisions
around
adding
their
own
capabilities
or
how
they
choose
we're
compatible
with
a
number
of
different
sort
of
perspectives.
A
On
this
and
to
kevin's
point,
we
probably
need
to
ship
everybody
on
this
call
some
bourbon
and
sit
down
and
sort
of
come
up
with
a
what
is
what
is
our
kind
of
community
stance
on
this
before
the
spiral
out
of
control?
I
do
love
the
velocity
of
new
capability
providers
that
we've
had
coming
online
this
year
and
the
diversities,
those
that
are
public
and
on
track
to
becoming
first
class
providers
like
blob,
store
and
vault,
as
well
as
on
the
as
well
as
on
the
on
the
private
side.
A
So
it's
just
neat
to
see
folks
doing
that
I
did
just
host
a
bourbon
tasting
for
70
people
on
monday
or
60
people
or
so,
and
I'm
happy
to
get
one
going
for
the
for
the
community
call
folks.
I
think
it
would
be
a
lot
of
fun
and
it's
probably
overdue.
A
Now
we
talked
about
doing
the
happy
hour
in
dc
anyway
for
folks,
like
matt,
that
are
local,
so
watch
maybe
watch
for
some
messages
on
slack
about
that.
I've
already
got,
I
think
another
30
or
so
port,
so
we
can
do
one
for
maybe
30
people
do
we
want
to.
Is
there
more
to
discuss
on
this
kevin?
Let
me
ask
you
to
sort
of
like
decide
when,
when
we
draw
the
discussion
to
a
closer,
I'm
sorry
stuart
were
there
other
people
that
wanted
to
chime
in,
on
their
perspective,
on
this.
G
Sorry,
I
just
wanted
to
say
a
big
thank
you
to
brooks
and
taylor
for
talking
last
week,
because
I
was
absolutely
amazing.
Thank
you
so
much
for
doing
that.
It
was
really
great.
I'm
sorry,
you
didn't
get
to
see
the
room.
Unfortunately
we
said
we
were
going
to
do
that
and
we
didn't
so
it's
bad
on
our
bud
but
yeah.
Thank
you.
So
much,
that's
really
cool!
No!
Not
on
this
subject.
I
just
wanted
to
go
in.
A
C
Yeah,
I
didn't
really
have
any
parting
comments.
I
think
it's
you
know
one
of
those
things
where
we
definitely
would
love
to
hear
from
everybody
and
continue
the
discussion.
You
know
people
want
to
get
into
the
weeds
on
it
in
the
slack
channel
or
in
a
slack
channel.
Then
I'm
more
than
happy
to
do
that,
but
yeah.
To
answer
the
other
question.
I
think
this
community
meeting
is
one
of
the
right
forums
to
bring
up
questions
like
this.
F
A
C
It
could
be,
but
you
know
we
could
also
reach
the
conclusion
that
we
aren't
going
to
dictate
how
people
version
their
stuff
and
we're
just
going
to
have
some
semi-prescriptive
guidance
on
you
know.
A
Okay,
all
right,
that's
great
kevin,
we'll
see
how
this
discussion
goes,
but
we
should
probably
be
making
something
very
intentional,
though
around
just
some
guidance,
because
I
think
that
matt
really
brought
up
a
great
topic
today.
Thank
you
for
that
matt
other
discussions.
Does
anyone
want
to
anyone
else
have
anything
on
their
mind
or
community
driven
or
anything
that
they
want
to
discuss
today.
D
It
I
got
it,
let's
see,
okay,
so
I
can
do
it.
So
I
have
two
community
callouts.
Both
of
them
are
in
the
wasm
cloud
host
runtime,
which
I
really
enjoy
doing
for
a
community
call
out,
because
I
know
that
you
know
if
you
haven't
touched
elixir
before
it
can
be
a
little
intimidating
and
I
really
want
to
bring
people
into
messing
with
it.
I
personally
really,
like
you,
know
the
functional
side
of
elixir,
but
anyways
I'm
getting
off
on
a
side
topic.
D
So
we
we
had
a
bug
report
submitted
from
somebody
in
the
community
that
got
an
error
that
we
realized.
Two
different
code
paths
can
kind
of
lead
to
getting
this
error.
So
when
you
try
to
invoke
an
actor,
you
can
get
this
log
in
the
host
log
saying
that
it's
missing
a
claim
for
a
contract
id.
This
totally
makes
sense
and
should
be
denied
when
an
actor
isn't
signed
with
the
con
the
capability
contract,
but
we
also
get
the
same
error
even
if
the
actor
is
signed
with
the
capability
contract.
D
If
the
link
definition
is
misconfigured
in
the
host.
So
in
both
of
these
situations
we
should
be
denying
the
invocation,
but
we
can
actually
disambiguate
this
a
little
bit
better,
and
so
this
is
a
great
first
issue,
because
you
get
to
actually
poke
at
some
of
the
code
that
we're
looking
or
the
some
of
the
invocation
code,
and
it
should
be
a
fairly
quick
fix
with
a
high
impact
and
then
the
second
one
is
a
fairly
simple
one,
but
also
has
a
high
impact.
D
D
I
was
trying
to
test
like
doing
a
really
long,
rpc
timeout
and
noticed
that
it
was
not
quite
picking
up
the
the
value
that
I
was
giving
it.
So
two,
like
I
said
two
two
fairly
quick
punts.
The
the
goal
of
these
community
call-outs
is
for
anybody
to
be
able
to
jump
in
and
and
kind
of
fix
this
and
they're.
You
know
everybody's
kind
of
short
on
free
time,
so
so
get
in
and
contribute
easily.
So
I
saw
the
chat,
but
I
don't
know
if
it's
related
it
is
not.
A
All
right,
two
community
call
outs,
two
great
demos.
Today,
three
capability
providers.
We
got
a
regular
12
days
of
christmas
going
on
up
in
here
this
week.
Anyone
else
on
the
call.