►
From YouTube: Triage Meeting 2020.04.09
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
B
So
hi
everyone
so
moving
in
first
to
the
PMI
issues,
we
first
found
a
compiled
requesting
count
info
appropriate
life
sign
for
lack
of
ground
juice.
The
issue
is
seven
zero,
nine
one
seven
and
Eddie
be.
Perhaps
you
would
like
to
discuss
this
one
or
any
Co.
You
also
that
to
come
into
place
just
a
minute
scope.
A
A
If
I'm
wrong,
it
seems
like
we,
so
you
can
sort
of
think
of
this
type,
as
almost
as
though
and
I
was
basically
like,
like
a
kind
of
struct
where
it
says
now
captured
lifetime
today
with
the
type
parameters
from
the
function,
and
then
this
call
requires
that
that
be
static,
and
that
doesn't
work
because
a
it's
not
static,
and
if
this
were
a
struct,
we
would
definitely
give
an
error
in
this
situation.
We
don't,
it
seems,
like
we
didn't
report
an
error
before
I
think.
A
A
A
I'm,
not
I,
don't
actually
think
can
I
say
this.
I,
don't
think
it's
unsound
means
to
have
the
old,
behavior
necessarily
but
I'm
sure,
but
there's
the
primary
this
kind
of
intersects,
like
some
some
rules
from
that
we
adopted
in
RFC
1214,
which
basically
say
if
you
have
a
struct,
let's
say
or
just
any
type,
if
you
have
a
type,
if
you
have
a
type
like
foo
of
X,
and
it
has
two
outlets,
some
region
take
Y,
then
X
has
doubt
lifted
ye
for
all
components.
A
X
right-
and
there
are
reasons
for
this,
but
it's
actually
kind
of
simpler
than
you
might
need
like
and
I,
don't
like
the
behavior
in
some
cases
and
I've
actually
been
wondering.
If
we
can
change
it,
though
it
would
be
interesting.
The
most
obvious
example
would
be
something
like
this.
Basically
the
outlook.
Well,
the
question
is:
does
the
out
the
out
live
to
be
to
the
outlet
needs
to
at
minimum
guarantee
that
whatever
data
is
in,
there
lives
longer
than
tick
Y?
It
also
has
another
rule,
which
is
what
RFC
1214
talks
about.
A
However,
you
can't
do
that
with
these
types
because
they're
anonymous,
so
you
can't
write
like
an
info
click
today,
some
trait,
for
you
know,
capture
lifetime
with
decay,
at
least
not
today.
Now,
if
we
added
some
feature
like
type
of
or
something,
then
that
might
change
the
story
here,
then
we
would
have
to
be
careful
what
the
problem
would
be.
Then
you,
you
wouldn't
be
able
to
use
this
to
get
inside
the
value
of
an
Associated
site.
Something
seems.
A
I
think
we
could
do
either
like.
If
we
wanted
to
do
a
warning
period,
it
would
be
possible
I,
it's
a
little
bit
annoying
because
it's
region
inference
but
I.
Imagine
we
could
do
it.
I
would
have
to
do.
Is
you
know,
walk
the
this
particular
case
of
this
kind
of
type?
We
would
walk
regions
and
I
said,
but
this
is
another
little
knot
and
we
we
didn't
check
in
the
type
checker
right.
It's
sort
of
before
that.
B
C
B
A
B
B
C
B
B
C
A
D
C
Before
it's
now
like
there's
an
easy
method
that
takes
for
it
like
I,
get
most
of
the
muscle
checking
game,
it
was
the
core
concept.
It's
not
like.
The
a
personal
invitation
does
not
remove
a
lot
of
special
cases
only
only
when
the
ADP
implementation
is
like
clearly
for
a
and
does
the
same
thing,
Norman
Chekov.
C
D
C
B
E
C
B
B
A
B
So
moving
on
updates
discriminants,
so
this
is
I.
Think
mostly
I
wanted
to
highlight
this.
There's
probably
not
a
lot
to
discuss
here,
but
what
this
changes
is
that,
instead
of
always
having
you
64
as
the
discriminants
for
Ninon
or
instruct
us
works,
he
uses
something
that
fits
better
for
the
type
by
projecting
to
an
Associated
type
internally,
which
is
which
is
not
exposed
on
the
trait
isn't
exposed
either.
So
it's
mostly
any
communication
attempt,
but
it's
I
thought
it
would
be
good
to
know
about
this.
B
B
B
G
Yeah,
so
basically
it's
important
that
all
code
continues
to
all
code.
We
can't
start
making
programs
fail
to
compile
because
we've
tried
to
promote
something
and
that
promotes
something
that
we
weren't
promoting
before
and
then
promotion
fails.
So
that's
what
most
of
Ralph's
concerns
are
and
that's
the
main
concern
of
this
PR.
There
was
sort
of
we
discussed
this
earlier,
like
in
a
meeting
several
months
ago
and
the
concern
there
was.
We
don't
want.
B
That's
Holly
Saylor
also
makes
the
point
that
I
mean
computer
type,
can
already
change
significant
behaviors,
and
you
can
make
compilation
on
a
latent
issue
just
to
make
a
point
as
well.
So
what
this
allows
now
is
that
you
can,
instead
of
having
first
defining
a
constant
and
writing
back
back
new
in
equals,
affecting
you
in
that
one.
You
can
just
inline
the
constant
into
the
the
elements
expression
in
the
expression.
B
G
A
Yeah,
this
is
probably
like
I
might
be
happier
starting
with
that
which
is
sort
of
not
controversial,
addresses
the
use
case
in
a
uniform
way
that
will
work
for
all
types
and
then,
if
we're
like,
but
it's
so
annoying,
you
know
this
common
case.
It's
still
too
annoying.
Maybe
we
should
sugar
it.
That
would
feel.
A
B
B
B
B
B
H
Saying
if
we're
going
with
this
compromise
of
entry,
suggesting
that
people
use
an
explicit
console
in
this
context,
then
we
should
have
some
checkpoint.
Where
we
actually
go
back
and
say,
is
this
enough?
Do
we
need
to
take
a
step
further?
I
would
like
to
avoid
the
case
where
we
just
forget
about
this
yeah.
A
Well,
the
the
author
of
the
Emmaline
d
would
create
wrote
and
they're
using
what
seems
to
me
to
be
sort
of
an
embedded
like
use
case,
some
kind
of
camera,
that
they
need
to
copy
data
from
and
they
needed
specific
alignments.
And
they
found
that
this
structure
was
useful
because
it
lets
trucks,
be
generic
over
their
alignment
by
having
a
bunch
of
types
of
different
sizes
that
you
can
use,
which
is
a
nice
hack
actually.
But
they
didn't
seem
particularly
perturbed
by
the
idea
of
having
a
limit
like
4k.
B
A
B
D
B
H
H
A
A
H
H
H
Right
exactly
when
you're
doing
a
wrapper
align
that
big,
it's
not
going
to
be.
Oh
I,
think
we
need
to
do
this
for
performance.
It's
going
to
be.
If
this
doesn't
work.
The
program
is
wrong
so
that
much
I
can
understand.
If
we
can't
enforce
it,
we
should
say
so,
but
on
the
other
hand,
what
is
somebody's
recourse
if
they
are
actually
able
to
enforce
it
on
their
platform
and
rusty
is
just
not
as
smart
as
their
platform
is
for.
B
D
B
H
C
D
A
D
A
B
A
D
A
A
C
J
D
A
C
And
also
the
way
I
realize
the
monetization
aspect
of
the
size
of
a
flow
chain
is
that
it's
almost
like
a
resource
issue.
Like
the
same
way,
you
couldn't
compute
some
value
in
memory.
If
people
don't
see
it
yet,
but
if
it
imagine
generating
your
constant,
wasn't
very
with
me
that
just
doesn't
remember,
you
can't
give
it
we're
being
computed.
So.
D
A
A
A
B
A
A
A
I
think
we're
not
I
feel
like
we're
talking
a
bit
past
each
other
all
I
was
saying
is
Felix
suggested
that
were
maybe
we're
only
giving
errors
to
crates
that
weren't
going
to
get
the
alignment
they
expected
anyway.
So
it's
sort
of
not
it's
like
we're,
giving
them
an
error.
They
should
have
gotten
but
I'm
saying
that's
not
the
case.
I.
F
Was
at
least
trying
to
say
that
some
of
the
crater
breakage
could
be
cases
that
fall
into
that
category.
I
don't
know
them,
but
so
I
wasn't
trying
to
say
that
all
of
them
were
blanket.
No
failing
at
runtime.
I
just
meant
that
it
might
be
a
misnomer
to
say
that
all
these
cases
of
breakage
represented
things
that
weren't
already
broken
at
the
OS
level.
B
A
A
H
A
H
A
You
are
that
they
have
what
you
are,
but
if
you
don't
do
that,
you're
not
actually
fixing
the
bug,
and
if
you
have
workarounds
they
can
mow
skate.
I
mean
the
assumption
here
is
that
few
people
actually
want
to
do
this
and
that
the
majority
of
the
folks
who
do
either
have
another
decent
option
or
are
on
some
embedded
platform
where
it
doesn't
matter.
I.
Think.
A
H
H
B
B
B
H
B
B
C
A
B
B
C
Well,
actually,
no,
we
can,
which
is
collagen
ethics
like
CP
the
square
bullet
points
you
you'd
have
to
be
able
to
know
that
the
discriminant
don't
overlap.
So
the
only
thing
that
we
could
allow
right
now
is
when
the
only
one
this
like
only
one
very
big
baby,
and
we
could
like
at
some
special
cases,
but
unless
there's
tons.
B
C
C
D
E
B
B
C
C
C
C
C
A
B
Just
a
quick
just
a
bit
quick
notes:
resolving
okay
roughing
for
try
box
Josh's
posts
that
we
would
agree
to
use.
Okay,
wrapping
for
try
box
and
I
separated
out
an
issue
for
for
thoughts,
consensus
building
and
well
meet
the
text.
Read
the
proposal
and
check
it
out
and
check
your
boxes.
If
you
agree
or
do
something
else,
I.
B
B
C
C
Right,
so
if
you
go
to
lay
down
for
that
and
then
actually
run
it,
you
can
see
that
the
yeah,
the
penny
messages
is
not
what
it's
nothing
like.
What
you'd
expect
just
actually
works,
then
it
was
before
:
location,
it's
inside
the
the
macro
micro,
tiny,
and
that
because
we
don't
track
the
actual
macro
factories,
we
don't
like
anything
like
IG
note
or
Mac
related
information
of
crossfades,
so
it
just
happens
that
the
family
tip
is
really
useless.
One
and.
D
A
F
A
B
But
it
also
feels
like
we're.
We
seen
something
of
subpar
quality
I'm
very
like
saying,
but
him
yeah,
but
we
were
like
promoting
it
even
more
and
saying
that.
Okay,
you
users
start
using
this
feature
in
your
this.
Such
beauty,
your
own
stuff
and
and
it
might
have
like
serious,
used,
lose
bugs
it's
like
some,
but
that's,
maybe
not
the
best
experience.