►
From YouTube: Lang Team Triage Meeting 2019.10.31
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).
C
C
A
C
And
that
does
change
the
semantics,
the
current
semantics
of
array
initialization,
because
Kern
since
we
require
copy,
if
you
let's
say,
use
an
impure
function
to
initialize
an
array,
it
will
call
that
impure
function
once,
whereas,
presumably
at
runtime,
we
would
call
it
n
times
for
each
item
in
the
array
which
does
change.
Sorry.
B
C
So
well,
this
is
specifically
alright
for
there.
Yes,
so
if
you
have
something
like
like
this
currently,
this
has
one
meaning,
which
is,
it
will
call
R
and
once
we
want,
we
would
like
to
make
BEC
new
work
or
box
new
work,
but
then
the
only
real
way
to
make
those
work
is
to
call
box
new
n
times
or,
as
currently,
we
only
call
Rand
once
so.
That
is
a
bit
of
a
change
in
semantics.
There.
A
C
A
A
B
C
E
E
B
B
B
C
C
C
B
You
know
I
would
say
the
argument
has
proven
I
first,
like
that
we
discussed
this
last
week
in
the
meeting
and
when
we
left
when
I
left,
that
I
was
feeling
pretty
convinced.
We
do
the
answer
that
we
should
make
see
it
not
make
it
you
be
to
unwind,
thirsty,
calls
and
introducion
mind
after
hearing
more
the
arguments,
I'm
less
sure,
I.
One
thing,
though
I
think
I
want
to
like
take
a
little
bit
too
I
said
reorganize
that
document
and
try
to
think
you
through
what
are
the
cases
and
what
are
they
off?
B
What
are
they
appealing
choices
and
what
are
the
cases
for
and
against
them?
But
one
thing
I'm
wondering
is
whether
I
feel
like
this
is
a
place
where
the
there
are
definitely
strong
opinions
amongst
the
team
and
maybe
what
we
should
do.
I
don't
think
today
is
the
right
day,
but
is
try
to
have
that
like
present
those
cases
and
have
that
conversation,
it's
got
a
little
time
to
have
it
one
more
extended
time
than
squeezing
it
into
a
meeting.
I
know
Taylor.
B
You
had
a
lot
of
pretty
strong
opinions
on
this
point
and
I'm
wondering
if
there
are
specific
things,
both
I
wondering
if
you'd
like
to
be
in
that.
But
secondly,
if
there's
what
I
was
wanting
to
I
thought,
a
lot
of
that
had
to
do
with
the
amount
of
dead
code,
we
would
optimize
out
and
other
sort
of
performance
considerations
around
the
choice
and
I
was
wondering
if
we
can
maybe
make
measurements
on
that
in
the
future
code
base.
Do
you
think
that
makes
sense
like
what
is
it
the
simplest
one
would
be?
F
I
could
hear
you
yeah
I'd,
be
happy
to
meet
and
and
have
a
more
dick
dedicated
conversation
on
that
I
also
be
happy
to
do
some
measurements
and
work
with
other
people
to
do
measurements
on
the
future
cookies.
Here
we've
done
some
existing
measurements
around
any
news
about
financial
support
and
some
we
have
a
bunch
of
stuff
for
doing
size,
profiling
and
stuff
and
increase
or
I'd,
be
happy
off
with
that.
Oh.
B
B
C
A
A
B
B
You
find
a
good
simple
example.
You
can
write
things
that
overlap
it
in
particular.
Well,
I.
Don't
have
a
great
example,
but
if
you
have
like
a
table
for
all
T
where
T
is
not
necessarily
sized
there
here,
we
go
and
it
turns
out
so
I
win
it.
Last
time
we
talked
about
this,
we
proposed
a
solution
of.
Why?
Don't
we
look
and
see
if
there's
any
potentially
overlapping
in
bull
and
then
we'll
remove
it?
If
there
is,
we
will
not
introduce
the
synthetic
one
and
that
way,
there's
no
overlap.
Now,
there's
just
one
like.
B
Basically
the
problem
was:
there
are
two
angles:
the
one
that
compiler
introduced
in
the
one,
the
user
wrote
we're
going
to
take
the
one,
the
compiler
juiced
away.
So
there's
only
one,
and
now
we
don't
have
a
problem.
Well,
that's
true,
but
it
turns
out
that
there
is
plenty
of
code
that
unfortunately
relies
on
the
the
existence
of
both
of
those.
So
to
speak,
and
in
particular
the
any
trade
does
so
there's
this
imple
here,
which
says
for
any
type
T,
which
is
not
necessarily
sized.
B
We
can
implement
any
40
and
what
type
ID
does
is
it
calls
site
type
anything,
and
so
this
basically
implements
any
Ford
in
traits.
It
also
technically
implements
any
Ford
in
any,
and
so
this
is
exactly
the
overlap
we're
talking
about,
but
what
the
compiler
currently
does
is
just
do
to
essentially
accidental
specialization
is
what
it
kind
of
does.
B
So
if,
when
we're
actually
generating
code,
if
we
see
that
the
receiver
type
is
din
any
we
generate
a
call
through
the
V
table,
and
thus
the
type
ID
you
get
back
is
not
for
the
typed
in
any,
but
for
the
underlying
hidden
type
of
the
object
itself,
which
is
what
you
actually
wanted.
So,
unfortunately,
this
simple
exists:
it's
stable
and
what
somewhat
amusingly
when
I
did
implement.
B
My
PR
I
basically
changed
the
behavior
of
the
any
trade,
because
now
we
took
away
the
V
table
call,
and
so
you
just
always
called
this,
which
meant
it
no
matter
what
the
din
any
was
no
matter.
What
type
was
hidden
inside
the
did
anything.
You
always
got
back
the
same
type
ID,
which
was
for
that
didn't
any
type
itself,
which
was
not
the
one
you
wanted.
The
compiler
crashed
in
horrible
ways
that
took
me
a
while
to
understand
it
seems
pretty
clear.
We
cannot
do
that
change
just
like
that.
B
It
completely
breaks
the
anything.
So
what
I
realized,
though,
is
that
in
fact,
there
is
a
way
to
view
this.
That
is
fine,
but
which
is
that
we
could
think
of.
B
D
B
B
So
it's
still
like
a
backwards
incompatible
change
to
do,
but
it
would
be
sound
and
it
would
permit
the
any
trait
and
it
would
also
permit
the
other
uses
in
the
standard
library
which
I
found,
which
was
like
borrow
and
borrow
mute
and
to
string
and
I
did
find
some
people
in
the
wild
using
these
as
trade
objects.
So
they
do
exist.
So
the
only
thing
is:
there's
something
weird
to
me
about
the
solution.
First
of
all,
they
stopped.
Does
that
make
sense,
and
so
far
so
little.
A
B
D
In
regarding
specialization
just
so
that
we
understand
where
we
are,
there
is
the
concern
with
specialization
and
that
we
all
absolutely
do
want
to
have
it
and
we're
just
not
100%
certain
that
we
can
soundly
implement
it.
Or
is
the
concern
also
that
we
don't
have
general
consensus
that
we
wanted,
even
if
we
can't
do
it
I'm
just
trying
to
figure
out
if
this
is
a
technical
limitation
or
also
a
consensus
limitation.
I.
D
B
A
B
Be
useful,
yeah
I
mean
we'll
that's
right,
I
think
there's
some
question
of
timing
and
how
we
would
do
it,
but
that
right
now
this
strikes
me
as
the
most
viable
way
forward,
but
there's
still
a
concern.
I
haven't
gotten
to
yet
already,
which
is
separate
from
that,
which
is,
if
we
here's
my
tab.
If
we
do
this
approach
I'm
talking
about,
then
basically
the
compiler.
The
question
then
becomes:
when
does
the
compiler
introduce
this
new
info
of
trade
for
didn't
rate,
because
now
it
used
to
be
we
sort
of
just?
B
It
was
usable
within,
however,
that
is
a
little
weird
now,
because
now
you
start
to
get
errors
like
you,
you
have
some
trait
that
you
wrote
it.
You
don't
intend
for
it
to
begin
compatible,
and
you
write
an
info
like
this
one
and
now
you're
getting
errors,
because
we
didn't
write
default
because
the
compiler
is
putting
in
a
specializing
in
before
you
that
maybe
you
don't
actually
want.
So
that
seems
unfortunate
to
me
and
that's
what
led
me
to
ponder
an
alternative
which
is
maybe
what
we
do.
D
B
B
So
if
there's
no
overlapping
impulse,
otherwise
it
is
not
incompatible
unless
you
explicitly
opt
it
back
in
with
AD
intrud,
and
then
you
get
this
Impala,
in
which
case
you
have
to
meet
the
specialization
requirements
so
like,
but
that
would
mean
in
practice
as
if
we
just
did
this
change,
the
any
trade
would
not
be
considered
didn't
compatible.
That
would
obviously
break
a
lot
of
codes,
so
we'd
have
to
declare
it
as
dim
and
that
we
would
have
to
clear
this
as
default
and
then
what
you
could
do,
I
realized
is
or
I
was
suggesting.
B
If
we
do
have
an
addition,
you
could
plausibly
say
well
we're
moving
to
this.
Didn't
trade
model,
Ross
2021,
all
things
that
are
again
after
be
declared
in
and
we'll
both
port
your
code
forward
as
automatically
as
we
can.
We
can't
do
a
hundred
percent
perfectly,
but
the
but
but
for
a
backwards
compatibility.
We
use
this
more
complicated
mechanism
in
earlier
editions,
but
we
we
still
resolve
this
on
this
issue
and
are
there.
Yes,
both
editions
would
be
stopped
they're.
Just
it's
just
a
question
of
like
how
complex
the
model
for
twenty.
B
A
What
if
we
introduced
like
a
nightly
roast
20,
18
plus,
and
then
we
can,
we
have
ideas
that
we
wanted
to
go
inside
an
addition
at
some
point.
We
do
it
now
and
when
we
feel
ready
when
we
feel
like
we
have
collected
enough
things
in
that
flag,
we
can
like
have
an
RC
to
actually
novel
issue
so
that
we
are
sort
of
not
rushing
into
things.
Yeah
I.
D
A
H
B
B
B
B
It
just
doesn't
really
do
anything
that
opt
out
from
this
rule
so
that
I
can
get
any
and
borrow
and
stuff
compiling,
and
then
we
can
test
the
impact
on
the
ecosystem,
because
I
I
think,
though
it
should
be
compatible
with
most
uses
the
case
that
breaks
is
you
have
an
overlapping
in
both
and
you
use
the
trade
as
it
didn't
trade,
so
yeah.
My
hypothesis
is
a
lot
of
people
who
have
overlap,
but
a
lot
of
traits
are
not
in
trades
to
begin
with,
so
they're
ruled
out
and
of
those
that
are
some.
B
I
B
I
Compatible
I
see
that's
right.
It's
like
there's
two
choices,
I
think
right.
You
can
either
we
apply
those
rules
or
your
trade
is
just
not
been
compatible
right,
and
so
we
don't
want
to
break
people
who
never,
who
actually
just
want
like
an
anon
context,
a
trade,
exactly
people
who
never
cared
if
their
thing
was
ought
to
safe
or
not
shouldn't
have
to
meet
the
extra
requirements.
I
end
I'd
like
to
know
how
many
people
would
be
like
what
will
be
broken
if
we
applied
this
rule
across
the
board.
Yeah.
B
That
was
my
intended
next
step
and
I
think
there
might
be
a
version
of
this.
We
could
do
where
we
say
if
the
extended
rules
like
it's
dim
incompatible.
If
the
extended
rules
like
you,
don't
get
a
hard
error,
just
becomes
daily
incompatible.
If
you
don't
have
the
right
default,
but
I'm
kind
of
I
can
explore
that.
Also
I
should
write
that
down.
However
I'm
a
little
bit
wary,
I
would
sort
of
like
to
get
away
from
that,
because,
basically,
it's
just
increasingly
baroque
yeah.
D
B
B
B
A
A
B
D
B
Group,
it
doesn't
affect
any
documentation.
It
basically
just
affects
currently
stuff
happens
when
unwinding
occurs,
and
this
makes
that
stuff
better.
It
seems
like
an
important
implementation
step,
whatever
we
do,
if
we
ever
want
to
have
I've
done
finding
the
highlights
are
here:
I
have
planned
to
land
this
PR.
I
B
I
B
B
A
A
Drive
man
drive
makers,
meanwhile,
and
if
you
use
proper
gratitude,
you
have
to
remove
that
helpers
yourself,
which
can
be
mitigating
sometimes,
and
this
is
basically
a
feature
request
to
have
helpers
or
control
attributes
as
well,
not
just
corporate,
and
what
that
jingle
is
saying
is
that
this
seems
easy
to
support.
That's
a
convenient
feature.
A
B
Get
the
lotion
would
be,
do
we
really
want
an
RC?
You
will
be
satisfied
with
her
straight
to
unstable.
A
B
A
B
Like
our
conclusion
was
yeah,
like
I,
said:
you're
smaller
see
what
we
find,
but
a
PR,
the
links
to
a
good
write-up
with
probably
suffice.
Yes,
yes,
I,
don't
want
to
go
into
the
details
on
this
because
I
fear
I'll
take
the
whole
meeting
but
I'm.
This
is
this
wacky.
Let's
come
back
to
another,
you
may
read
my
comment
here.
So
I
want
to
get
to
tell
you,
maybe
maybe
we
should
schedule
a
meeting
for
yeah.
B
B
A
I
G
D
Well,
getting
some
small
things
out
of
the
way.
First,
for
the
record,
the
concern
that
I
have
marked
is
one
that
just
trivially
needs
a
like
a
line
or
two
added
to
the
RFC,
at
which
point
I
will
happily
lift
it.
So
this
is
purely
a
documentation
issue
and
not
something
that
should
be
a
significant
blocker
I.
Think.
A
E
I
I
F
Concern
about
I,
don't
I
mean
I,
don't
I
don't
want.
A
D
D
A
So
I
think
the
concern
to
in
certain
for
its
values
is
like
a
collection
of
different
concerns.
That's
gathered
into
one
I
think
there
is
a
question
of
limits
which
made
and
wasn't
proponent
of
I.
Think
there's
a
question
of
like
how
does
this?
Does
this
interact
with
Bing
and
like
editors?
We
can't
do
what
were
semantic
analysis
like
regular
expressions
and
stuff
I
think
is
a
concern
about,
like
some
people,
think
it's
easier
to
mentally
mentally
parse
turtle
fish
for
humans,
yeah
thanks
Robin,
there's,
maybe
there's
something
I
missed
I!
Think.
E
B
We
should
look
at
what
we
agree.
We
should
do
something
on
this
thread
right
and
there
are
basically
a
few
options.
We
can
close
it.
What
seems
bad
is
to
just
open
it
yeah
we
can
close
it,
we
can
sort
of
move
to
do
something
we
could
close
it
and
we
open.
What
are
the
space
of
things
you
think
we
should
consider
in
who
has
energy.
D
If
I
might
make
a
suggestion,
I
think
it's
safe
to
say:
none
of
us
have
especially
the
appetite
to
try
to
sort
it
out
with
each
other,
while
having
the
discussion
in
public
on
github,
with
github
dismal
mechanisms
for
doing
so
and
for
getting
snipe
there.
So
I
think.
If
we're
prepared
to
have
the
discussion
and
come
to
a
consensus,
then
we
should
come
to
that
consensus
and
then
reopen
the
thread
and
say:
okay,
we're
at
a
consensus.
We've
lived
in
our
concerns.
Here's
where
we're
at
you're.
A
A
A
F
Was
gonna
ask
what
do
you
mind?
I,
don't
I,
don't
know
how
much
we
want
to
go
into
this
panel
versus
having
a
dedicated
meeting
on
it,
but
I
I
was
a
little
bit
confused
on
there.
A
couple
of
bullets
here
about
interaction
with
them
and
other
editors
easier
for
humans
to
parse,
simpler
grammatical
complexity
and
limiting
language
for
olestra,
better
Diagnostics
performance
of
backtracking.
Those
all
seem
like
theoretical
concerns,
with
a
a
more
complex
like
linguistic
grammar.
Are
they
also
concerns
that
you
you
share
specifically
tailored
to
this
change.
A
F
If
anybody
is
mentioning
them
in
the
context
of
this
specific
change
makes
the
language
harder
for
people
to
parse
like
or
this
specific
change
is
causing
diagnostic
problems
or,
if
it's
all,
in
the
realm
of
you,
know
the
class
of
grammars
that
are
not
ll
pay
or
tend
to
be
and
have
the
ability
to
be
more
complex
and
harder
to
understand
than
LK
tremors.
So.
D
F
I
I
D
D
D
D
We
may
agree
to
be
clear,
I'm
proposing
if
we
were
to
do
that
and
I'm,
not
saying
that's
the
path
forward,
but
if
we
were
to
do
that,
it
would
be
with
a
very
specific
statement.
The
long
stabilization
period
is
to
evaluate
the
impact
on
tooling
and
ecosystem,
and
nothing
else
is
the
criteria
we're
using
for
you
know.
Do
we
decide
to
reject
this
rather
than
stabilizing
this.
F
A
it's
alright
I
am
I'm
offer
experimenting,
I'm
all
for
getting
this
innately
and
training
out
I.
Don't
think
that
the
concerns
that
have
been
raised
in
the
thread
and
the
people
who
are
concerned
about
them
will
be
satisfied
for
with
any
results
that
we
show
on
nightly,
because
I
I
don't
think
that
the
concerns
here
are
strictly
technical,
I,
don't
I,
don't
know,
I,
don't
know
what
we
want
to
do
about
that.
I!
Think
that
there's
a
real
conversation.
F
G
F
A
I
do
think
that
if,
if
we
do
this
on,
that
means-
and
some
people
will
still
like
how
how
their
concerns
will
not
go
away.
If
we
like
show
show
that
this
is
not
a
problem
for
Chile,
but
some
people
might
be
persuaded
can
give
enough
time
and
another,
and
another
point
is
that
I
remember:
we
actually
tried
this
bits
in
to
do
the
puzzle
and
it
wasn't
a
problem,
though,
in
fairness
thing
is
this:
we
christianism's
I
think
so
that
it
can
do
nothing
more
thing.
B
E
E
I
just
wanted
to
say
that,
with
my
the
actual
concern
that
I
have
on
there,
I
think
compiler
is
absolutely
free
to
go
switch
to
a
fancy
new
gll
parser,
absolutely
free
to
have
this
behind
the
feature
flag.
All
that
sort
of
thing
I
have
nope,
with
whatever
technical
choices
that
they
want
to
end
up,
making
there
in
a
way
that
makes
this
better
allows
better
diagnostics,
etc,
etc,
etc.
Those
are
all
good
things
if
they
work
out
and
that's
how
the
rusty
authors
wish
to
have
that
work.
I,
don't
think
anything.
I
Those
opinions
will
not
be
impacted
by
that,
and
so
I
think
we
have
to
decide.
We
are
overruling
them
now
if
we're
going
to
pick
up
nightly,
which
decided
those
like
we're
not
like
those
concerns,
don't
like
block
this
feature
right
so
like
and
that
feel
like
I
kind
of
fundamentally
I
I
am
not
convinced
by
those
concerns,
but
I
don't
really
want
to
be
engaged
in
the
thread
that
is
telling
people
that
we
have
decided
to
overrule
those
concern.
B
Okay,
I
think
most
of
what
I
wanted
to
say
has
been
said,
except
that
Contra
boats,
maybe
I,
don't
mind
being
I,
don't
want
to
be
engaged
but
I
feel
like.
We
can't
not
make
decisions
because,
like
we
were
not
going
to
reach
consensus
with
the
whole
internet,
you
know
that
yeah
and
we
can't
not
make
concessions
that,
were
you
can't?
Let
that
be
a
thing
that
that's
gonna
block
us
from
going
forward,
which
is
not
to
say
we
should
lightly
disregarded.
B
People
have
to
say,
I
feel
pretty
okay,
if
we
feel
like
we
can
get
and
I
think
we
have
it
pretty
exhaustive
list
of
the
no
a
list
of
the
things
people
are
talking
about,
and
we
kind
of
either
say
well.
We're
gonna
resolve
that,
specifically
by
testing
it
on
lately
and
I.
Think
as
this
wouldn't
here
like
to
make
that
effective
and
like
to
avoid
the
repeat
of
this
threat,
me
would
make
specific
conditions
and
the.
D
B
Weighed
these
two
things
and
we
feel
it
over
all
the
benefits
of
this
outweigh
the
danger
of
that,
but
like
I'm
willing
to
do
that
threat
if
we
can
come
to
an
internal
consensus,
I
guess
and
I
think
it's
important
that
we
are
not
afraid
to
do
that.
All
that
said,
you
know
if
we
decided
this
is
more
trouble
than
it's
worth
right
now.
I
could
live
with
that
too.
D
So
to
breathe
the
short
question
was
I
just
wanted
to
take
the
temperature
of
the
call
again
and
ask
it
sounds
like
we
don't
specifically
have
substantial
concerns
about
the
unknown
unknowns.
It
sounds
like
the
concern
votes
that
you
expressed
earlier
was
more.
We
don't
know
what
we
don't
know
and
we
don't
fully
know
whether
we
will
break
something
horribly.
If
we
do
this,
would
that
be
an
accurate
assessment?
And
what's
the
general
consensus
on
that
I.
B
D
B
D
To
test
that
was
what
the
REIT
that
I
asked
to.
That
question
was
that
I
did
I
feel
the
same
way.
I
do
completely
agree
that
this
is
more
a
concern
of
what
don't
we
know
rather
than
what
do
we
know,
and
that
includes
the
various
concerns
registered
on
the
thread
in
that
regard.
I
feel
like
in
reopening
this.
The
biggest
thing
we
need
to
do
collectively,
as
the
language
team
is
to
demonstrate
to
the
thread
that
we
actually
do
understand
what
we're
getting
into
that.
D
We
understand
what
lines
are
present
and
that
we
are
crossing
one
with
deliberation
and
not
using
that
as
an
excuse
to
say
it
doesn't
exist
anymore.
This
is
slightly
related
to
my
cup.
My
registered
concern
on
the
threat
of
this
doesn't
set
a
precedent,
but
I
think
it's
also
a
little
bit
of
trolling
here
and
there
aside
I
think
that
there
was
a
bit
of
a
point
in
the
thread
about
lines
and
the
language
team.
In
that
we
should
recognize
the
concept
of
a
focal
point.
D
People
call
it
a
showing
point
where
there
is
an
obvious
point
of
consensus
around
here's,
where
the
grammar
is
here's,
how
strong
the
grammar
needs
to
be.
If
we
cross
that
a
little
bit,
then
the
obvious
next
point
is
well
I.
Guess
we
just
have
a
full
gll
grammar
and
we
don't
actually
want
that.
So
if
we
want
to
reopen
this
and
say
here's
how
we're
moving
forward
I
think
it
would
be
valuable.
D
If
we
said
here
is
the
exact
scope
of
our
decision,
we
are
not
deciding
that
rust
is
now
a
gll
language
with
usage
of
arbitrary
tll
features.
We
are
deciding
that
this
one
future
is
worth
a
special
place
in
the
grammar
because
it
trips
people
up
because
this,
because
that
we
are
not
using
that
as
a
precedent
for
future
GOL,
we
will
put
the
same
amount
of
deliberation
into
any
future
grammatical
change.
That's
not
confined
to
the
previous
grammatical
subsets.
D
That
kind
of
statement
needs
to
happen
and
for
going
to
reopen
this
so
that
people
don't
have
room
to
say,
pay.
The
link
team
doesn't
know
what
they're
doing,
which
I
think
that
there
is
at
least
a
legitimate
fear.
That
is
motivating
people
to
say
things
like
that,
not
that
we
don't
actually
know
what
we're
doing,
but
more
people
have
fear
of
unknown
unknowns
and
they
want
to
know
hey.
Do
we
have
we
at
least
thought
about
the
unknown
unknown.
A
E
One
of
those
things
that
I
like
when
I
found
Russ
in
the
first
place
is
it
didn't,
have
the
like
Scott
Myers,
most
vexing
parse
thing,
because
you
don't
have
to
do
this
so
I.
Think
personally,
I've
never
found
the
turbo
fish
nearly
as
much
as
a
problem
as
a
lot
of
people
do
it's
I
think
we
also
have
this
like
right
now.
It's
really
easy
to
support
this,
because
we
only
have
like
ten
examples
of
it.
Maybe
maybe
that
I
would
get
less
worried
about
that
sort
of
thing.
E
If
we
actually
saw
people
pick
this
up
and
nightly
greatly,
but
the
other
thing
that
were
slightly
does
worry
me
about
nightly
experimentation
on
this.
Is
we
keep
seeing
more
and
more
that,
as
we've
made
things
work
on
stable,
that
people
don't
go
to
nightly
just
for
avoiding
to
:,
sometimes
so
I'm
I'm
slightly
skeptical
that
we
can
get
a
huge
enough
volume
of
stuff
on
nightly
here?
That
would
have
this
turned
on.
That
would
actually
notice
anything
going
weird,
because
most
crates
aren't
going
to
bother
I,
don't
know
if
that
answers
your
question.
B
Right,
first
of
all,
static
Morse
has
a
pretty
good
point
that
we
don't
want
to
I'm
going
to
be
careful
with
how
much
we
say
we
won't
go
forward
without
XY
and
Z
is
that
puts
potentially
gives
people
power
to
block
in
bad
faith,
though
I'm
not
sure
if
that
would
be
a
real
problem,
but
I
think
they're,
careful
phrasing
of
our
choice
needs
that
can
help.
But
the
other
thing
I
would
just
say
is
that
I
think
we
haven't.