►
From YouTube: 2021-03-11 RFC 2229
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).
B
It's
going
well,
we
got
a
couple
pr's
out
last
week.
Let
me
share
my
screen
so
that
it's
just
easier
to
go
over
the
project
board.
C
B
B
All
right,
let's
see,
okay,
I
think
that
looks
good,
so
we
have
a
pr
for
the
handling
rubber
back
structures.
I
think
there
was
one
comment
on
adding
another
a
comment,
so
I
have
done
that
there
is
or
this
should
be
in
review.
B
So
we
have
a
pr
on
getting
rid
of
our
old
data
structures
and
using
the
plugin
captures
everywhere
and
roxanne
has
a
pr
going
for
patterns
and,
yes,
everything
is
sort
of
doo-doo
in
progress,
but
the
general
highlights
are,
I
guess
we
have
three
things
main
left
other
than
the
capture
macro
itself
and
we
have
diagnostics.
B
B
A
B
Keep
going
then
we
have
the
closures
not
being
cloned
or
autographs,
not
being
satisfied
issues
and
then
insignificant,
destructors
I'll
be
okay.
A
B
Yeah,
I
think
it's
going
to
be
just
like
we
discussed
about
it
earlier,
so
it
shouldn't
be
too
bad.
I
think
implementing
insulin
instructors
won't
be
that
bad
figuring
out,
the
list
of
which
ones
to
mark
as
insignificant
might
be
a
little
bit
more
work
to
some
degree.
So
we
have.
B
Yeah,
it's
just
it's.
I
mean.
B
B
All
right,
I
have
our
instagram
yeah,
so
I'll,
let
miles
I'll
put
polish
complete
and
let's
see
how
far
we
get
and.
B
Yeah-
and
I
think
that's
most
work-
that's
left
for
is
like
three
main
sort
of
things
that
are
left
for
feature
complete.
I
think
we
should
go
and
discuss
patterns
because
there's
an
edge
case,
that's
sort
of
showed
up
around
diagnostics,
and
I
don't
have
a
really
nice
solution
for
it.
So,
let's
see,
if
I
can.
B
A
Just
thought
of
another
case,
I
wonder
I
wonder
if
what
would
happen
if
you
have
never
types
worry
about
it
later,
but
never
types
can
make
some
variants
not
possible.
A
It's
not
it's,
not
something.
We
have
to
solve
right
now,
especially
because
network
types
aren't
stable.
But
if
we
I'm
just
thinking
about
how
plausibly
you
could
have
like
results
with
never
as
the
okay
type,
in
which
case
it's
sort
of
a
univarianty
num,
because
it
can
only
have
an
error,
are
you
familiar
with?
Never.
B
A
So
in
the
case
of
the
my
example
of
result,
normally
you
have
two
types
like
the
okay
type
in
the
error
type.
But
if
you
made
one
of
those,
never
you're,
basically
saying
that
it
can't
that
that
variant
can't
happen
because
so
like,
if
you
imagine
you
created
error
of
e,
what
okay
type
should
you
have?
You
could
sort
of
have
any
okay
type,
because
it's
not
the
okay
variant,
so
it
doesn't
matter
what
type
you
put
there.
A
So
one
type
you
might
put
is
never,
and
that
then
means
it
could
never
ever
be
the
okay
type.
It
can
only
be
the
error
variant
so
plausibly
we
could
say,
oh
if
you're
matching
and
it
does
require
a
discriminant
read
in
but
not
like.
We
could
say
that
it's
a
univariant
enum,
because
although
the
enum
has
many
variants,
we
know
that
some
of
those
variants
are
impossible.
I'm
not
going
to
worry
about
this
right
now.
B
So
your
comment
was
that
the
only
time
we'll
see
a
down
cost
rejection
is,
if
there's
a
single
variant
enum
and
therefore
we
should
be
able
to
test
for
index
0..
So
I
don't
think
that
is
true,
so
that
is
true.
If
we
were
looking
at
the
hair
place
and
trying
to
build
a
mirror
place
out
of
it,
because
that
is
when
we
won't
have
any.
B
That
is
when
we
would
capture
our
enums
completely,
but
when
we
are
going
from
here
to
here,
the
workplace
can
actually
contain
that
downcast
and
we'll
try.
We
want
what
we
really
want
to
do
is
truncate
for
it
so
like.
If
we
have
something
like
this,
which
is
like
some
foo
it
could
have
for
now.
We
know
it's
just
sum
up
bar
x
and
on
that
we're
building
down
cost
zero
to
access
our
sum
and
then
field
zero
and
then
field
one
to
x
actually
end
up
accessing
this
wire.
B
B
But
again
in
this
particular
scenario,
it's
a
little
hard
to
do
because
the
logic
becomes
a
little
cyclic
in
the
sense
that,
to
figure
out
the
type
of
this
place
before
downcast
rejection,
we
need
to
figure
out
which
up
part
does
this
resolve
to
which
may
or
may
not
easily
work
out,
and
it
could
be
a
little
hard
to
follow.
A
D
A
D
B
So
I
did
say
about
truncating,
so
the
reason
we
might
want
to
truncate
is
just
to
a
thing
like
the
easiest
answer
would
be
to
exit
early,
because
since
we
are
only
doing
this
conversion
for
characters,
we
know
if
you
don't
have
a
particular
projection.
Why
build
a
longer
list?
Why
will
the
longer
set
of
projections.
A
So
one
thing
I'm
wondering
is:
if
you
have
your
example,
had
a
downcast,
oh
okay,
never
mind
good.
I
I
was
I
was
missing
line
115..
I
was
wondering
when
the
variant
got
reset
back
to
none.
B
Yeah
yeah,
it
gets
reset,
got
it
yeah.
So
I'll
leave
a
comment.
I
think
these
two
need
to
be
removed
now
what
needs
to
be
removed?
Oh.
A
B
Oh
yes,
I'll
just
do
that:
okay
and
second
yeah.
So
the
reason
this
sort
of
works
out
is
because
the
way
we
defined
our
her
place
says
field
index
within
index,
so
we
can
easily
represent
what's
actually
happening
here
and
also
checking
for
zero
is
not
an
if
and
only
of
check
so
because
multivariate
genomes
can
also
have
a
zero
variant.
B
So
that
is
also,
I
guess,
might
not
be
the
best
way
to
think
about
it
and
yeah.
Let's
just
go
to
the
diagnostics
issue,
so
we
basically
saw
this.
So
we
have
this
simple:
okay,
I'll,
probably
open
the
test
file.
One
second.
B
So
we
basically
have
a
contradicting
read
here,
because
tuple
gets
moved
in
c1
and
also
in
c
do
and
that
ends
up
basically
causing
a
bar
checker
or
error
or
conflict.
So
what
we
saw
was
we
were
missing.
We
were
missing
certain
parts
of
the
error
message
which
described
what
was
the
problem
was,
I
guess
the
short
answer
why
we
are
missing
them
right
now
is
earlier.
The
borrow
checker
conflict
was
between
a
closure
capture
and
another
place
that
was
previously
moved
now.
B
Breed,
I
do
have
the
mirror
for
this
code,
and
so
the
first
problem
was
like.
B
B
B
Yeah,
I
think
you're
right
qui.
I
feel
like
that
in
this
particular
scenario
that
will
still
not
work,
because
because
tuple,
okay,
so
okay,
this
is
not
my
comment.
B
Yeah
exactly
exactly
so
yeah
that
will
still
because
I
partially
moved
place
consumer
issue.
Okay,
I'm
gonna
scroll
down.
B
Yeah,
I
have
to
touch.
A
It
doesn't
appear
to
be,
I
think
it
will
help
the
mirror
has
written.
It
would
help
because
they're
not.
B
B
A
So,
okay,
the
other,
so
there's
two
options,
the
other
one
of
them
is
we
can
we
can
move
the
fake
reads
instead
of
having
them
be
separate
instructions
they
could
plausibly
be
part
of
the
closure,
r
value
or
something,
but
that's
kind
of
not
sure
I
like
that
change,
but
it's
possible
to
imagine.
Do
you
understand
what
I
mean.
A
B
A
F
D
A
F
A
A
fake
read
before
a
closure,
and
then
the
borrow
checker
knows
to
skip
ahead
to
find
the
closure
instruction
for
the
diagnostic
purposes,
but
that
just
seems
like
a
pain
in
the
neck
is
the
fake
weeds
are
always
no
they're,
not
always
even
gonna
come
oh,
they
might
always
be
in
the
same
basic
block
as
the
closure.
B
B
A
A
E
B
I
had
a
another
weird
hack,
which
is
so
basically
the
way
we
sell
the
where
to
point
for
this
is
we
point,
give
it
a
span
instead
of
a
span,
if
we
give
it
a
multi-span
that
just
also
points
to
the
closure
and
just
is
annotated
with
used
here.
A
A
B
It
all
it
needs
is
the
closure
def
id
so
that
it
can
just
point
to
the
closure
or
close
your
hair
id.
So
I
can
get
the
spam
and
that's
about
it.
B
A
B
B
F
A
So
it's
a
fake
read,
I
see
yeah,
I
somehow
I
thought
it
was
more
complicated
than
I
didn't
understand.
I
think
that's
the
right
way
to
go.
I
don't
have
a
strong
opinion
about
whether
it's
in
this
pr
or
in
a
follow-up
pure,
especially
if
yeah
I
would
like
it
to
land
before
this
thing
hits
beta
or
something
but,
like
I
don't
know
when
that's
happening
but
like
a
temporary
diagnostic
regression
seems
okay.
A
A
B
A
B
Yeah,
so
I
think
this
is
yeah
yeah.
I
think
they
changed
a
lot
of
how
they
made
the
hair
and
mirror
arena
allocators
of
boxes.
So
this
might
be.
I
don't
know
how
much
this
gets
impacted.
A
B
All
right,
yeah.
A
A
I'll
do
an
app
boris,
delegate
too,
so
that
you
can.
You
can
tell
boris
to
land
it
whenever
you're
ready
and
not
have
to
wait
for
me
to
miss
your
ping
for
a
few
days
or
something.
C
B
A
B
Yeah
all
right
yeah.
This
is
all
I
have
unless
you
guys
have
anything
else.
C
I
guess
you
want
to
guys
sing
it.
I
don't
know
if
you
want
to
just
discuss
it,
but,
like
I
feel
like
april
1st,
to
get
everything
done
might
be
like
not
possible
anymore,
just
based
on,
like
the
delay
we've
had
on
all
the
other
ones,
on
all.
A
B
I
I
looked
at
it
before
this.
It
really
will
come
down
to
finding
time
for
a
couple
small
issues,
but
yeah,
let's
go
through
it.
Actually,
let's
go
through
the
project
where
I
think
it
looks
much
more
comforting
there,
so
these
all
ideally
get
merged
within
the
next
week
or
so.
Assuming
we
get
time
for
rebase
and
the
pattern,
one
doesn't
turn
out
to
be
a
complete
like
difficult
process
yeah
this
one,
I
don't
think,
will
stay
here
long.
This
is
same
as
these
two
are
going
together.
So
that's
good.
B
This
one,
hopefully,
is
not
that
bad.
What's
left
for
feature
complete
on
this
side,
if
we
go
bottom
up
is
like
I
guess,
somewhat
easier,
which
is
we
need
to
write
a
blog
post,
which
is,
I
guess,
somewhat
non.
I
think
we
can.
B
We
discussed,
we
probably
want
the
word
group
should
do
it,
so
I
can
discuss
and
we
can
start
drafting
something
yeah
then
we
have.
This
is
again
going
to
be.
A
simple
change
is
just
changing
our
if
statement.
So
this
is
not
that
much
of
a
work.
B
So
actually,
I
think
maybe
we
should
discuss
this
next
meeting.
I
paint
decompiler
a
while
back
to
figure
out-
and
we
had
talked
about
this
briefly
in
the
past-
is
what
this
should
look
like,
captured
so
right
now.
It's
for.
I
think
this
is
consti
val.
B
The
way
they
describe
closures
is
just
capture
variable
and
it's
just
the
name
of
the
symbol
that's
captured,
but
I
didn't
see
it
being
used
in
too
much
detail
and
I
did
not
know
what
is
the
best
way
to
express
the
way
it's
done
today,
using
so
that
it's
still
copy
and
everything,
and
I
think
I
might.
B
A
Yeah,
so
you
need
to
store
possibly
not
just
a
symbol.
That's
the
problem
right!
Yes,
yeah!
Okay,
you
can
ping
them.
I
mean
one
easy
thing
to
do
is
gonna
be
relatively
easy,
which
you
should
ask
them
how
they
feel
about
it,
but
parameterizing
path,
alum
by
tcx.
Does
this
code
have
like
a
tcx
lifetime
anywhere?
I
don't
see
any
lifetimes.
A
B
Yeah
they,
I
remember,
I
think
they
just
moved
into
decompower
at
one
point
and
it
just
got
unanswered.
I
should
I'll
ping
it.
I
get
pimping
again
and
see
if
there's
something
so
for
now,
there's
just
successfully
somewhere
yeah,
there's
just
a
fix
me,
which
is
saying
make
it
more
descriptive,
so
that.
B
Yes,
I
think
that's
the
more
like
the
unknown
sort
of
there
is
straits
with.
I
think
jenny
and
logan
will
end
up
working
instead
of
I
guess
me
and
roxanne
I'll
work
on
insignificant,
destructors
with
proof
and
some
no
go
back.
Sorry,
sorry
and
yeah
diagnostics
is
left
I'll,
I'll
I'll
ask
chris
if
he
can
get
it
done
somewhere.
A
F
B
Yes,
I
think
the
what
we
decided
was:
okay,
actually
one
question
around
this
are:
do
we
want
a
separate
length
for
this?
I'm
assuming
or
do
we
and
then
we
just
create
a
length
group
which
is
like
rc229
land
group.
It
just
runs
all
because,
even
though
it's
the
logically
it's
similar
just
finding
root
where,
like
places
that
don't
specify
some
attribute
of
their
sorry
places
that
don't
satisfy
the
attribute
of
some
attribute
of
the
root
variable,
I
feel,
like
the
actual
lint
name,
seems
different
than
captured
vr
and
that.
B
A
Don't
really
have
a,
I
hadn't
thought
about
it,
but
you're
right
that
it's
somewhat
different.
D
A
B
B
Yeah
and
yeah,
it's
like
we
have
like
basically
three
chunks
of
issues.
Hopefully
they
don't
show
up
turn
like
to
be
like
have
much
of
their
own,
like
edge
cases
on
top
of
themselves
being
edge
cases.
So.
A
E
B
At
some
point,
I
also
want
to
discuss
something
from
like
our
administrative
purposes,
because
in
the
sense
that
we'll
be
doing
a
presentation
for
this
around
somewhere
around
mid-april
ourselves,
so
what
one
thing
that
came
up
with
discussions
with
our
like
the
prof
who
runs
this
course
is
that
maybe
you
should
like
be
the
person
who
go
like
goes
through
the
presentation
as
well.
Maybe
I
think
you,
you
also
get
you
have
to
grade
it
as
well.