►
From YouTube: GitHub Quick Reviews
Description
Powered by Restream https://restream.io/
A
A
C
So
just
little
yep,
okay,
so
users
want
to
be
able
to
digitize
Jason
strings
into
number
properties.
Some
particular
scenarios
include
realizing
non
infinity
and
negative
infinity,
so
I
propose
to
add
a
new
option
to
enable
this
scenario
and
also
possibly
allowed
serializing
numbers
of
strings
and
motivation
to
do.
This
is
to
better
interpret
various
API
endpoints
across
the
web
that
send
like
strings
like
this
in
these
formats.
C
Numbers
can
be
read
from
a
token
type
of
string,
and
this
doesn't
prevent
numbers
from
being
read
from
jason
token
type
number
so
write
a
string
means
numbers
will
be
written
as
g
strings
records,
not
as
just
numbers
and
last
one
is
to
allow
named
floating
point
d,
charles,
so
it
can
specifically
say
also
allow
reading
these
two
literals
industrialization
and
also
a
great
in
the
mountains
realization.
So
today
we
don't
support
that.
Also,
in
addition,
propose
a
decent
number
handy,
an
attribute
which
can
put
property
so
like
you
can
cherry-pick.
You
know
the
properties.
C
Okay,
so
just
moving
down
to
some
music
scenarios
link
examples:
okay,
so
like
a
line,
reading
numbers
from
strings
and
writing
numbers
strings.
So
you
have
a
classic
ins
number
one
and
two
specify
options
in
allowed
to
read
from
string
and
also
like
to
write
a
string,
and
then
we
do
just
that
which
is
like.
So
if
you
look
at
the
G
simple
hood
number
one
has
like
a
decent
number.
One
and
number
two
has
it:
Jason
I
mean
just
a
string
to
so
because
we
are
line
from
reading
from
string.
C
We
with
this
realizing
to
both
properties,
and
you
also
specified
to
write
the
number
I
string.
So
we
write
both
of
them.
My
string
mix
example
allowing
reading
numbers
from
some
point
constants
and
right,
including
point
constants,
so
you
have
like
your
class
with
numbers
like
int
number
and
number
two
properties,
so
without
any
option
within
14
point
constant
representations
that
don't
feel
so.
C
You
have
your
like
inch
number,
which
is
fine,
but
then
you
have
this
flute
number
any
end.
Thank
you
string,
which
is
not
fine,
so
we
throw
it
in
the
an
exception
today
and
similarly
on
serious
consideration
in
thumb,
but
just
fine,
but
your
flute
or
nano
give
you
this,
like
argument
exception,
specifically
saying
that
we
don't
support
in
that,
but
with
the
new
option
you
know,
set
your
alone
important
point
withdrawals,
and
then
you
can
know
video
in
slumber
and
also
the
onion
and
on
sterilization.
C
Ok
so
allow
reasonable
some
string
and
support.
Regina
right
includes
constants,
so
you
know
just
a
few
more
combinations
just
to
see
what
that
experience
look
like.
So
you
see
to
allow
reading
numbers
from
string
and
to
also
allow
floating
I'm
floating
point
literals.
So
the
first
instant
bar
is
a
string,
and
so
we
allow
that
and
then
it's
also
we
and
it's
a
convenient,
so
we
also
another
so
and
so
for
this
visualization
side.
C
It's
talked
about
this
a
little
bit,
but
specifically
this
is
a
little
redundant
because
you
could
have
just
specified
the
the
first
one.
But
moving
on
to
the
next
example.
C
Okay,
so
the
writes
in
number
and
that
one
should
also
be
allowed
reading
from
string
and
I
love.
Nature
also
write
our
around
where
the
cursor
is
on
the
screen
just
little
about
that
yeah.
So
this
should
be
a
lot
reading
from
scene
as
well.
Yeah
similar
that
so
because
he
said
so
so
read
from
string
would
not
write
a
string.
We
will
write
your
negative
one
and
then
also
write
your
any,
and
so
this
combination
is
helpful
for
scenarios.
C
So
another
example
is
sort
of
the
opposite,
which
is
to
write
a
string
that
allowing
him
to
floating-point
each
rose.
So
the
first
like
example.
Here
you
just
have
your
int
number
horizon?
That's
your
GSP
node!
So
you
know
that's
just
the
one
thing
so
that
that
works.
Fine
and
nothing
was
specified
for
flat
number,
so
don't
be
0
default
value.
This
husband.
C
So
I'm,
looking
at
this
irritation,
say
silicon
my
previous
button.
Okay,
so
did
this
decision
of
the
second,
you
have
a
an
int
number.
D
Look
is
that
yeah
definitely
make
something
up.
I
think
this
example
was
supposed
to
be
the
same
as
this
one,
but
without
the
quotes
around
the
integer
number
and
then
somehow
it
changed,
but
I
think
because
I
feel,
like
the
point
from
this
morning's
discussion
was
allow
named
floating-point.
Literals
here
means
that
the
read
would
succeed
for
the
nanan
quotes,
but
not
for
the
number
in
quotes,
because
you
didn't
say
numbers
real
loud
to
be
strings.
C
C
Okay,
so
some
additional
like
things
to
note
what
behavior
is
so
integers
gonna
have
this
more
places,
so
the
semantics
are
the
same
as
if
there
are
no
quotes,
except
with
that,
we
will
escape
what
I
discussed
a
point
later,
maybe
not
in
much
more
detail,
but
so
in
so
you
have
a
number
with.
You
know
it
is
more
please
so
that
would
throw
an
exception,
because
we
expect
like
an
integer
which
is
has
no
like
decimal
point.
C
Similarly,
they
becoming
any
children
or
lead
injury.
So
if
you
have
like
one
year
or
something
like
that,
then
that
also
not
work
so
leisure.
So
next
one
leading
or
trailing
whitespace
is
also
not
is
not
allowed
like.
You
don't
want
space.
So
an
additional
note
here
is
like
because
we
have
the
same
semantics
as
what
this
utilize
I
will
do.
C
Without
the
quotes,
currency
percentage
and
instead
formats
will
not
work,
so
this
is
like
tokens
in
the
payload
will
cause
an
exception
to
be
thrown
like
similar
ones,
the
ones
we
see
above
here
so
okay
in
the
future
we
can
like
based
on
user
feedback.
That's
not
something
I'm
particularly
enthusiastic
about,
but
so
nice
of
compact
conversions
from
strings
numbers
and
this
realization,
so
no
custom,
logical
setting
is
required,
but
we
will
accept
before
opting
to
be
here,
because
we
think
that
this
sort
of
behavior
should
be
updated
to
new
self.
C
C
C
A
A
So
their
behavior
that
you
have
with
respect
to
numbers,
they
be
the
in
down
habit.
Okay,.
A
C
A
A
I
mean
so
I
guess
so.
I
can
tell
you
where
I'm
coming
from
so
on
the
one
inside
I
would
be,
on
the
other
hand
like
I.
You
know
I've
spent
some
time
over
the
last
weeks
playing
with
blazer
and
in
that
context,
also,
of
course,
using
the
whole
stack,
which
include
civilization,
and
one
thing
that
I
found
was
that
what
frustrated
me
sometimes
was
that
the
way
civilization
works
in
the
infinite.
A
Do
you
know
where,
on
one
hand,
side
it's
infinity
another
hand?
It
comes
out
as
non
infinity
right
and
what
becomes
with
a
zero,
but
be
my
guess
so
I'm
just
wondering
like
how
we
think
about
that,
because
I
think
in
the
context
of
asp
net,
like
I
think
most
people
would
probably
appreciate
if
you
make
more
scenarios
to
work
but
I
understand
your
M
is
concerned
about
like
our
design
principle.
D
A
So
that's
kind
of
heading
it's
like
so
because
when
we
talk
about
defaults
by
their
scenarios
right
and
the
problem
is
like,
if
we
even
even
four
we're
right,
where
do
we
draw
the
boundary
for
breaking
changes
because,
for
example,
if
we
say
in
the
web
we
make
a
low
sorry,
we
said
write
a
string
to
true
by
default,
and
then
you
do
change
the
wire
format
right.
Attractively
speaking,
you
also
change
the
wire
format.
If
you
make
more
things
realizable,
but
I'm,
not
sure
that
we
would
care
about
that.
A
So,
for
example,
we
could
imagine
that
you
know
we
add
more
primitive
types
or
a
little
primitive
type,
so
the
default
set
of
things
we
support
like
timespan
values,
for
example
right.
But
you
know
there
would
be
a
breaking
change
in
the
sense
that
you
know
v1
produced
1x1,
JSON
payload,
and
we
too
would
produce
more
right,
but
I
think
for
developers
most
of
the
time
we
will
just
translate
into
more
things.
Just
work
about
funny
arrows.
G
A
That's
my
general
problem
is
that
I
think
to
me.
I
would
be
much
happier
for
spec
compliance.
If
you
would
say
everything
that
is
not
strictly
in
the
lines
of
this
bag
will
always
throw
because
in
at
least
you
know
when
you
when,
when
stuff
gets
dropped
right
now,
the
problem
is
sometimes
we
drop
information
and
the
only
way
to
find
out
is
by
actually
looking
at
what
you
sent
over
and
what
you
received
right
and
that's
I.
Think
the
thing
that
the
trix
beginners
is
that
you
know
you
send
an
object
to
the
server.
A
Everything
is
hunky-dory
on
the
server
like
stuff
doesn't
work
the
way
you
expect
to
until
you
actually
do
it
at
the
payload,
and
you
realize
oh
half
my
object
didn't
make
it
through,
as
if
you
would
throw
an
exception.
Then
I
asked
you
what
those
things
is.
We
can
actually
tell
you
what
to
do
what
you
could
say.
Sorry
using
a
data
type
that
we
don't
know
how
to
see
lies,
so
we
just
drop
it
from
the
payload
and
then
apologies
there's
a
fire
I'm
testing
going
on
in
my
building
right
now,.
D
Yeah
I
mean
no
matter
well,
I
guess,
unless
we
did
something
silly
like
make
the
default
be
right
as
string
but
not
allow
reading
from
string,
then
dotnet
talking
to
dotnet
will
just
work.
The
question
is
dotnet
talking
to
other
things
of
what
what
scenarios
exists
by
default,
that
that
should
work
out
of
the
box.
G
Well,
like
it
was
called
out
on
the
issue
here,
a
lot
of
a
lot
of
interchange
formats
use
strings
by
default,
because
the
number
one
thing
that
uses
Jason
is
JavaScript
because
that's
where
Jason
came
from
and
JavaScript
internally
represents.
All
numbers
is
double,
and
by
default
it
doesn't
read
as
in
32.
If
you,
if
you
want
an
at
32,
it
will
read
as
a
double
and
then
you
have
to
convert
it
to
an
int
32
which
moves
precision
loss
among
other
things,
and
so
people
use
strings
to
avoid
problems
with
interchange
across
languages.
A
Yeah
I
think
that's
something
I,
don't
know
how
you
want
to
make
forward
progress.
It's
just
a
couple.
These
two
things.
Maybe
let's
have
a
generic
conversation
about
defaults
and
how
we
make
more
things,
work
with
the
system.
Texture
Jason
stank,
but
as
far
as
the
API
is
go,
I
think
the
the
proposal
that
the
only
here
has
makes
sense
to
me.
The
naming
makes
sense
to
me
that
the
the
Flex
design
makes
sense
to
me.
A
D
D
A
D
C
Mean
the
one
thing
about
this
for
me:
like
there's
a
pattern
in
graph,
you
know
that
you
can
have
like
a
composition.
Server
between
like
various
epi
is
some
that
you
want,
and
sometimes
you
don't
so
just
this
narrow
way
you
could
get
usin
that
isn't
uniform
across
look
across
the
whole
object
graph
that
you
get
back
into
that
net.
A
Alright
but
like,
but
the
setting
would
be
like
it
doesn't
matter,
I
didn't
have
to
be
consistent
right.
It
just
means
you
you're,
basically,
if
I
say
a
low
reading
from
string
and
allowing
them
fall
and
quadratures,
it
just
means
that
if
somebody
gave
me
quotes,
I
can
read
it.
If
somebody
didn't
give
me
cool
seconds
to
read
it
right,
so
there's
still
one
conflict
that
works
for
the
whole
thing,
I
mean.
C
D
A
A
Publish
would
be
consistent
with
the
other
ones
like
if
you
like,
I,
don't
know
whether
the
other
ones
also
should
be
tab
level
or
not
so
like
if
the
other
ones
already
our
property
and
field.
Only
then
we
hope
we
should
start
with
property
and
fit
only
for
this
one,
and
then
we
can
have
a
separate
discussion
whether
it
makes
sense
to
promote
some
of
them
to
type
level,
but
I
think.
A
F
A
F
C
D
C
D
That
would
require
changing
the
Jason
reader
to
add
another
non-compliance
mode.
No
de-serialize
I
can.
D
H
H
H
A
C
G
A
A
A
C
Okay,
so
this
one
is
mostly
community,
so
if
we
can
just
go
to
the
proposed
EPA,
so
the
current
behavior
is,
we
have
just
an
ignore
condition
when
writing
default.
You
can
specify
that
globally
on
Jason
Sierra's
options
or
like
top
property,
so
this
will
mean
that
any
any
property
of
you,
that
is
a
default
value.
Let's
take
a
reference
types.
A
reference
type
of
value
would
be
ignored
on
the
revision.
C
What
we've
gotten
feedback
that
people
want
to
easily
be
able
to
just
specify
just
only
know
like
reference
that
should
be
ignored,
initialization.
So
the
way
you
could
do
that
before
was
ignore
no
values
specified
robbery
and
in
earlier
preview
we
had
Wendall
as
its
own
like
in
on
value
on
this
you
know,
but
we
move
that
and
said
we
want
to
have
like
a
simplified
concept
of
default.
Change.
C
Default
post
means
no
land
like
value
type
default,
which
I
think
is
still
a
valid
like
concept
to
have,
and
just
add,
when
writing
also
that
you
can
have
like
just
for
reference
types
and
in
the
future
like
if
people
want
this,
which
I,
don't
think
would
be
common
keys,
but
we
have
the
flexibility
to
add
a
when
writing
value
type
default
for
specific
leaders,
body
types
so
I
have
this
one
like
uses
simple
usage
examples.
You
are
here:
school
internet,
sleuthing,.
A
C
A
So
basically,
you
said
like
when
storing
Jason
for
frontale's,
then
if
operators
or
video
types
are
extremely
important,
it's
not
like
now
I
assume,
that's
because
in
Jason,
when
you
didn't
initialize
a
property,
it's
effectively
now
rather
than
zero,
because
jason
has
no
not
Jason.
Javascript
has
no
notion
of
of
like
types
with
fields
right.
It's
all
effectively
a
dictionary
right
so
versus
in.net.
When
you
didn't
set
an
int,
the
default
value
is
zero.
No,
no
right!
So
it
seems
like
I
guess
that's
where
that's.
A
Yeah
because
I
mean
for
objects
when
we
don't
send
null
values,
then
on
the
client
side
they're
effectively
equivalent
for
not
sending
them,
because
they're
undefined
or
now
all
right
versus
when
we
didn't
send
it
into
an
awesomely.
The
int
is
now
rather
than
zero.
I
guess
that
makes
sense
to
me.
Then.
C
Right
so
you
see
the
examples,
that's
what
we
have
to
just
specify
when
the
only
globally,
in
this
first
example
so
I'm,
just
to
note
that
it
does
not
come
to
civilization.
As
the
you
know,
body
suggests
it's
just
for
winning
writing
no
and
a
word
socialization
and
it
doesn't
affect
body
types.
So
I
mean
one
question
is
okay,
so
property
is
the
next
thing.
It
cause
very
proper
property
option.
C
If
so,
in
a
similar
thing
like
it's,
no
don't
don't
utilize
it!
Well.
If
so,
let's
put
this
on
the
value
type
property
I
believe
the
correct
video
v2.
True,
since
that's
not
if
I
lead
option,
because
if
a
litter
can't
be
known
is
this
something
I
want
to
be
like
I
mean
why
strict
in
the
same
way
in
other
places,
I
just
want
to
call
it
that
I've,
just
probably
the
logical
ones
right
to
get,
and
if
there's
any
opposition
to
that.
A
D
It
makes
sense
to
me
that
if
somebody
has
put
it
as
a
property
or
put
it
as
an
attribute
on
a
property
and
the
property
as
a
value
type
of
throwing
of
I,
don't
like
the
the
combination
of
things
you
told
me
to
do,
makes
no
sense
seems
fine.
It's
effectively
a
design
time
exception
that
you
catch
it.
The
first
time
you're
generating
the
serializer
for
the
type
so
I.
A
A
D
A
F
And
we
used
to
have
this:
functionality
is
deprecated
right,
we've
to
have
a
pool
as
says,
ignore
null
values
or
something
like
that,
and
then
we
removed
it
or
deprecated
it,
and
then
we
replaced
with
this
enum,
but
we
didn't
add
back
the
special
null
only
because
we
didn't
think
people
cared
about
only
know.
Somebody
cares
about
it
now
we're
kind
of
adding
it
back,
but
only
on
the
right.
The
other
bool
also
supported,
read.
A
Yet
to
me,
like
it
makes
sense
to
say
on
declined.
It
matters
basically
the
receiver
of
the
payload
make
here,
but
if
the
receiver
of
the
payload
is
dotnet,
then
I
think
it's
hard
to
understand.
Why
either
would
care
about
now
versus
default,
because
it's
kind
of
a
function
of
what
the
type
system
is
used
to
utilizing
into
right.
A
A
C
D
D
F
E
D
D
Okay
and
to
bounce
back
for
a
moment
to
the
previous
issue,
chat
raised
a
good
point.
I
guess:
I'll
go
ahead
and
call
you
out
that
Fredrik
made
a
good
point
that
Jason
number
handling
none
reads
sort
of
strange.
It
is
our
common
value
for
flag'.
So
it's
probably
fine,
but
maybe
we
should
consider
using
a
different
name
like
strict.
D
C
A
D
D
But
in
this
case
I
think
because
I
mean
the
weirdness
would
be
as
if
you
say,
strict
pipe.
Allow.
Reading
from
string
like
the
fact
that
we
named
the
strict
means
that
you,
you
may
think
that
it
means
something
other
than
none
of
these
options.
So
so
maybe
none
is
still
the
best,
because
it
it
tells
you
don't
pipe
it
with
something
whereby
pipe
I
mean
or.
E
D
A
D
A
D
A
E
F
D
Yeah,
so
the
only
feedback
that
I
would
then
have
is,
if
somebody
wants
like
strict,
if
they
were
strict,
pipe
right
is
string
and
they
think
that
means
something
different
than
just
write,
a
string
that
maybe
an
analyzer
should
warn
that
you're
combining
strict.
But
then
it's
not
related
to
this
feature.
It's
related
to
anything
where
we
have
a
zero
that
has
name
something
other
than
none
or
even
the
none
so
somewhere
on
our
list
of
potential
analyzers.
D
C
D
B
A
A
A
We
went
over
and
decided
like
here's
a
bunch
of
stuff
that
we
may
want
to
obsolete,
and
this
is
what
this
document
is
basically
capturing.
I
think
when
we
started
this
document,
we
I
think
had
a
fairly
runtime
focused
view
on
this,
which
is
kind
of
features
that
we
didn't
pour
to
core
or
reported
to
core
in
terms
of
the
API
is
to
make
more
code
compile.
A
We
didn't
actually
port
a
technology,
and
so
what
Libre,
I
think
did
is
yeah
a
pass
over
all
the
things
that
you
know
we
may
wonder
obsolete
because
they
don't
do
what
think
they
do.
People
will
do
another
pass
of
actually
library
function
that
we
don't
like
so
I
think
the
first
one
here
should
be
relatively
uncontroversial,
in
my
opinion,
because
if
you
don't
know
what
CRS
are,
then
you
should
probably
be
in
agreement
with
Russell's
leading
it,
because
it's
still
not
important
to
you.
A
A
B
B
A
A
D
A
A
A
Yeah
we
create
pdb
generator
is.
Was
that
part
of
yeah?
That's
part
of
some
call,
maybe
is
I
think
they
we
had
for
debugging
support,
which
we
no
longer
support,
because
you're
no
longer
running
out
just
on
Windows
authentication
manager,
authenticate
appeared
on
a
support
or
don't
support
it
I,
don't
even
know
what
these
are
for.
So
I
guess
that
seems
fine,
so
I
guess
how
did
you
find
this
API
as
you
harvested
them?
How
did
you
he
just
searched
the
co-pays
so.
B
Most
of
these
are
things
that
I
just
wanted
to
kill
for
a
long
time
create
pdb
generator,
actually
that
that
was.
That
was
not
my
suggestion.
I
just
need
to
spend
a
little
bit
of
time
recalling
why
that
was
added
to
the
list,
but
it
the
fact
that
it
throws
pns.
He
makes
it
a
candidate
anyway
and
that
we
don't
have
a
and
that
we
don't
have
a
task
to
ever
bring
it
back
in
a
future
version
right.
A
Well,
we
don't
mean
to
me
the
most
compelling
argument
is
why
this
is
that
API
is
because
we
already
said
we
don't
support
generation
of
binaries
with
the
runtime
right
this
week
you
should
use
a
compiler
or
metadata
reader,
writer
and
so
to
me.
Pdp
generator
is
part
of
that
right,
bifid,
safe,
doesn't
exist
either.
A
B
A
E
A
B
A
B
D
A
Yeah
cuz
I
mean
like
that's
just
doing
on
whole
diagnostic
IDs
were
quite
like
you
usually
attach
a
severity
to
an
a
diagnostic
ID.
So
it's
a
bit
weird.
If
this,
if
the
same
diagnostic
ID,
is
sometimes
a
warning
and
sometimes
an
error
in
the
same,
build
and
then
part
and
then
some
occurrences
you
can't
suppress
and
some
Yukon
droid.
That
seems
confusing
and
it.
A
D
B
I
have
an
issue
linked
to
this.
That
proposes
a
fixer
for
some
of
these
and,
depending
on
how
the
we
fix
her
can
sometimes
just
suggest
removing
the
attribute
entirely.
But
there
are
certain
cases
where
the
developer
really
needs
to
consider
what
they
were
trying
to
accomplish
and
that
the
fixer
should
not
make
recommendations
in
those
scenarios,
because
we
we
don't
know
what
you
were
trying
to
do
right.
D
So
assuming
that
I
recall
correctly,
a
fixer
is
triggered
by
a
diagnostic
ID
and
not
by
any
context.
So
if
you
want
a
fix
or
sometimes,
though,
sometimes
need
a
different
ID
that
different
ID
can
be
an
analyzer
produces
an
ID
and
it's
different
than
the
obsolete
warning.
But
if,
if
simply
the
presence
of
this
obsolete
error
means
a
fixer
should
wake
up,
you
get
the
lightening
or
you
get.
The
light
bulb
based
on
the
proper
code
was
issued,
and
then,
if,
if
it
looks
at
it
and
says,
I
have
no
fix
for
this.
D
B
A
Yeah,
so
somebody
in
shed
is
asking:
do
we
really
want
to
give
up
being
able
to
globally
mute
of
solutions
and
msbuild
in
favor
of
being
forced
to
list
each
bucket
ID
separately?
I
would
say
yes
so,
ironically,
yesterday
I
watched
a
plural,
sir
at
course
on
logging
and
my
old
friend
Nicholas
bloomer,
who
basically
talked
about
production
logging,
and
he
said
funny
enough
that
in
any
given
environment,
you
have
way
more
occurrences
of
messages
than
message
formats,
but
you
know
kinds
of
Logging's.
A
If
you
will
and
like
I
think
observations
is
very
similar
right
like
if
you
have
way
more
occurrences
or
things
you
want
to.
Obviously,
then
we
have
you
know,
categories
for
that
and
I
think
you
know
in
this
list
feels
like
there's.
Eight
you
know,
categories
being
proposed.
I
would
assert
that
by
and
large,
even
if
we
add
more
things
to
it,
will
never
exceed
like
20
or
30
of
those
right.
A
I
B
B
A
I
I
A
I'm
saying
that
we
can
make
it
two
different
features:
if
we
cared
right,
I
would
say:
I
would
not
I
would
not
say
because
an
existing
project
or
we
did
that-
that's
clearly
what
they
meant
but
I'm,
not
against
adding
a
global
opt-out
switch
I'm.
Just
saying
that
the
existing
switch
I,
don't
think,
is
the
right
way
to
do
it,
but
I
think
that's
that's
yeah
we
can.
A
We
can
talk
to
the
compiler
team
at
first
I
need
to
verify
that
that's
behavior
that
you
actually
get
because
Levi
seems
to
suggest
that
the
existing
behavior
or
the
behavior
that
people
have
right
now.
If
they
did
the
observation
for
the
for
the
previous
ID,
the
compiler
will
turn
it
off
entirely.
Well,.
D
I
A
So
let
me
just
give
the
example
at
the
Sun
screen
right
now.
If
you
look
at
eight-point
coding
security
right,
let's
say
we
missed
one
of
them
that
key
container
permission,
isolated
storage
permission.
No
one
of
those
right
I
see
no
problem
with
vetting
that
Internet
six,
because
I
mean
if
it's
just
under
the
existing
umbrella
like
yeah.
If
you
told
us
the
last
rule,
is
you
don't
care
about
caste?
Well,
then
yeah,
you
still
don't
care
about
caste
and
that's
fine
right.
A
So
I
think
that
that's
that's
reasonable,
but
if
you
turn
off
all
of
our
solutions
because
they
were
like
four
hundred
warnings
regarding
cans,
if
you
started
to
obsolete
some
HTTP
client
API
is
well
then-
and
you
probably
want
to
know
about
those
right,
so
I
would
I
would
say.
Diagnostic
IDs
are
really
four
features
right.
So
if
we
also
do
the
feature,
if
we
missed
an
API,
we
should
feel
free
to
apply
this
attribute
in
a
later
release.
If
it's
a
secondary
feature,
it
should
be
a
separate
ID.
That's
how
I
think
about
it.
A
I
A
So
that's
the
thing
right,
surely,
on
the
on
the,
on
the
one
hand,
side
YES.
On
the
other
hand,
no
right
I
mean
like
we
have
received
this
feedback
multiple
times
from
people
that
you
know
our
Docs,
that
point
out.
Issues
with
api's
are
not
very
accessible,
because
people
can
reason
about
them
or
can't
see
them,
and
they
have
asks
us
to
obsolete
stuff
right.
A
B
I
A
I
understand
that
this
pain,
but
I'm,
saying
like
the
tooling,
is
good
enough
to
suppress
that
right,
because
all
you
have
to
do.
Is
you
go
on
the
first
morning
you
control
dot,
you
say
suppressing
global
file,
it
adds
one
attribute
or
one
I,
don't
even
know
whether
they
have
to
support.
Do
it
in
the
in
the
printer
project
file,
but
there's
one
code,
fixer
generically,
then
we'll
add
a
suppression
for
the
diagnostic
ID
to
your
project
or
to
your
to
your
route,
see
a
global.
A
You
know
assembly
in
photos,
ES
file,
and
so
given
that
on
this
list
right
now,
there's
ten
category
is
clicking
ten
times
and
any
reasonable
code.
Size
I
think
is
fine,
because
it's
not
proportional
to
the
number
of
occurrences.
It's
just
proportional
to
number
of
observations
we
are
doing,
and
then
it's
not
a
large
bucket
that
will
grow
significantly
over
over
the
years
right.
What
will
grow
and
what
is
largely
the
problem
of
logical
paces.
Is
that
look
at
cast
right?
A
D
I
D
A
I
think
that's
fair
I
mean
it's
the
thing
we
go
separately.
That's
not
saying
you
need
to
think
of
these
as
features
right,
not
as
weights,
right
right
and
then
I
think
it
makes
sense.
I
mean
it's
sorted
by
its
basically
I'm
trying
to
say
it
to
me:
it's
not
a
breaking
change
to
add
a
new
API
under
Nick
under
an
existing
diagnostic
ID.
It
usually
just
means
you
up.
It
usually
just
means
you
forgot
one
time.
D
A
D
So,
like
I
know
that,
with
the
thing,
that's
not
on
this
list,
x.509
certificate
to
a
private
key,
the
that's
a
get
set
property
and
framework,
it's
defined
as
get
set
in
core
for
API
compat.
The
set
always
throws
in
core
and
no
one
should
really
even
use
the
get,
because
it
has
complicated
legacy
behaviors
and
the
we
added
replacement
methods
that
are
strongly
typed
instead
of
the
weakly
typed
property
and
one
of
the
more
common
things
that
I
see
on
Stack
Overflow
is
someone's
like
hey
I.
D
Did
this
thing
I'm
using
private
key
stuff,
isn't
working?
What
should
I
be
doing
and
it's
like?
Well,
you
should
call
like
get
RSA
private
key
or
you
know,
whatever
algorithm
you
think
you're
using
instead
and
and
so
that
one
would
definitely
be
a
whether
it's
an
analyzer
or
an
obsolete
or
whatever
of
something
that
you
write
the
code,
and
it
says
this
is
not
the
code
you
wanted
to
be
writing
and
obsolete.
Attribute
is
any
way
we
could
do
that
that
doesn't
require
building
an
analyzer
per
problem.
D
And
so
that
would
be
the
kind
of
case
where,
because
really
what
you
want
I
think
in
in
the
large
part,
if
you've
already
written
code,
you
probably
don't
want
to
be
told
something's
obsolete,
but
when
you're
writing
new
code,
you
want
to
be
told.
This
is
not
the
way
you
want
to
do
things,
but
until
someone
invents
the
notion
of
new
code
versus
old
code,
analyzers
I
think
this
is
the
best
we're
gonna
get.
A
I
From
like
trying
to
roll
out,
you
know
trying
to
pick
up
experiment
with
a
new
analyzer
pack
and
try
and
roll
that
out
say
across
dotnet
runtime
and
it's
extremely
iterative,
especially
if
you're
in
a
code
base
which
has
different,
builds,
that
build
in
different
source
files
that
aren't
always
active
at
one
time,
with
your
different
building
for
different
OSS
or
that
standard
versus
that
stain
or
one.
And
you
know
it's
very,
very
iterative,
you
suppress
one
thing
and
then
you
do
a
build
then
it
might
take.
I
However
long
your
bill
takes
before
you
can
find
out
the
next
problem
and
you
can
sit
there
for
hours
trying
to
do
this
kind
of
stuff,
and
you
know
as
much
as
we'd
like
to
believe
that
yes
is
snappy
like
with
big
projects.
You
know
running
these
analyzers
can
take
a
long
time
to
actually
even
alert
you
to
the
fact
there
isn't
one
yet.
D
The
problem
would
be
you
know,
taking
past
experience,
you
get
something
like
office
or
even
what
we
have
in
in
our
dotnet
runtime
build
system
where
it's
not
one
solution
right.
It's
the
build
system,
you're
trying
to
upgrade
to
a
new
version
of
the
SDK
as
a
as
a
build
in
central
engineering
somewhere
and
now
you
have
seven
hundred
different
projects
that
are
getting
some
level
of
warning
and-
and
you
really
do
want
to
I
mean
in
those
big
build
systems
there,
some
we're
going
to
be
your
top
level.
D
What
flags
am
I
passing
on
every
invocation
of
the
compiler
and
you
can
do
whatever,
but
and
certainly
once
you've
said,
ignore
all
obsolete
ins.
Now
everyone
is
in
trouble
forever,
but
it
is
definitely
a
problem
with
big
engineering
systems
when
you're
building
more
than
one
project
of
any
time,
any
new
warning
or
error
pops
up,
and
that's
essentially
why
we
stopped
obsoleting
things,
because
people
like
office
kept
complaining,
so
I
think
I'm
arguing
both
ways
but
I
like
doing
that.
A
Yes,
so
like,
by
the
way
the
default
code
fix
that
will
suggest
to
either
suppress
it
literally
in
stores,
which
is
surrounding
the
occurrence,
with
two
pragma,
which
you
almost
never
want,
and
then
there
is
configure
severity,
none
and
then
it
will
just
create
an
editor,
config
file
or
add
to
the
existing
and
it's
a
config
file
and
just
adds
it
globally,
which,
if
you
have
one
for
your
solution,
it
should
even
do
it
solution
wide
so
that
one
I
can
tell
for
sure.
Now.
Let
me
try
the
other
one,
which
is
there.
D
So
that
x.509
certificate
to
that
private
key,
you
essentially
never
want
it.
It
doesn't
work
for
ECDSA
it
on
windows,
gives
you
the
bad
or
the
the
bad
of
the
two
crypto
types,
because
that's
the
property
that
everyone
at
hard
caste,
RSA
crypto
service
provider
and
that
it
frequently
pops
up
on
Stack
Overflow
of
someone's,
like
hey
I,
wrote
this
code
using
private
key
and
it
doesn't
work
and
it's
like
yeah
it
cuz
use,
get
RSA
private
key.
Instead,
it
works
I.
D
Can't
make
that
one
work,
because
if
you
wrote
code
in
framework,
if
you
wrote
code
ever
using
this
property,
you
are
99%
likely
to
have
read
it
and
hard
cast
at
the
RSA
crypto
service
provider.
That's
why,
when
we
added
RSA
CNG
and
net
force
in
net
four
six,
we
said
we
cannot
make
this
ever
return,
RSA
CNG!
It
will
break
the
universe
and
we
had
to
make
another
method.
And
then
the
answer
is
now.
No
one
should
ever
read
that
property.
D
D
It's
just
probably
not
the
stuff
you
want,
and
so
I
don't
know
if
that
meets
our
current
definition
of
obsolete,
or
we
just
continue
doing,
tell
people
when
they
use
it
not
to
use
it
or
or
what
so
I'm
bringing
it
up,
maybe
as
meta
and
not
necessarily
as
something
that
we
want
to
do
in
the
five
wave.
But
if
I
can't
paint
it
in
five
I'm
happy
to.
A
D
A
Yeah
yeah
I
believe
the
compiler
doesn't
do
that.
I
think
what
I
think
the
way
the
compiler
works
is.
It
always
raises
the
diagnostic
and
then
they
filter
it
out
later
yeah,
which
is
the
problem
right,
because
they
don't
know
anymore,
that
this
came
from
an
obsolete
attribute,
but
yeah.
So
now,
I
will
talk
to
them
and
see
whether
there's
anything
we
can
do.
I
D
So
yeah
right
now,
the
stuff
that's
on
this
list,
I
think
every
single
one
of
them
is
a
PN,
SC
or
or
the
Cavs
stuff,
where
it
just
says,
return
true
and
isn't
doing
the
checks
that
it
implies.
So
the
question
is:
are
we
are
we
also
now
wanting
to
allow
obsoleting
if
you're
calling
this
you're,
probably
making
a
mistake
that
there
is
a
there's,
a
better
API?
That
is
that
does
what
you
wanted
to
do
out
of
this,
but
fixes
scenarios
that
this
one
didn't
work
in.
A
I
A
So
my
concern
vanity
suggestion
is
I
mean
I,
know
how
much
you
use
vs,
but
like
they
are
very,
very
much
the
longtail
I
mean
I.
Guess
in
oh,
you
put
a
space
here.
According
to
your
code
style,
you
shouldn't
put
a
space
here
like
it's
a
it's.
This
super
tedious,
noisy
stuff.
That
I
mean
I
pretty
much
always
turn
off,
but.
A
A
A
Maybe
maybe
different
notions
of
what
constitutes
an
IDE
suggestion
than
what
I'm
talking
about
is
to
all
the
IDE
level
stuff
like
we
move
unused,
parameter
unnecessary
assignment
all
that
stuff
shows
up
as
the
informational
category,
which
is
basically
as
errors
as
ones
in
this
informational,
an
informational
thing
in
the
error
list.
You
can
completely
turn
off
and
I
believe
you
don't
see
scribbles
in
the
IDE
either
and
that's
what
I
did.
Okay.
A
But
that
doesn't
mean
that
that's
the
only
features
we
have
I
mean
like
I,
so
I
think
there's
certainly
an
argument
to
be
made,
for
maybe
we
should
think
of
ways
to
promote
these
things
without
spamming.
Your
arrow
list
and
I
mean
I
very
much
like
the
idea
of
things
that
only
show
up
on
the
editor,
because
they
are
kind
of
way
you
develop
code
and
that's
when
you
actually
care
and
okay.
Now
is
the
time
to
not
take
a
dependency
on
obsolete
stuff
right,
because
I
will
also
say
like
to
me.
A
Like
you
know,
we
sharp
I
had
this
feature
for
years,
but
they
would
just
render.
Obviously,
today
P
is
and
the
code
completion
will
strike
out
font,
which
is
kind
of
very
strong
deterrent
to
use
the
API,
which
I
honestly
think
is
better
than
hiding
the
API,
because
it
needs
to
be
acknowledged
it's
there,
but
we
also
tell
people
like
you
really
really
shouldn't
be
using
this
anymore.
Yeah
yeah.
D
We
really
it's
just:
where
do
we
think
the
bar
is
for
what
we're
willing
to
paint
as
obsolete,
because,
certainly
you
know
I
could
propose
it
as
an
analyzer
of
in
our
bucket
set
somewhere.
If
you're
calling
this
property,
you
really
want
to
call
one
of
these
other
methods
instead,
but
the,
but
if
we
think
that,
because
I've
answered
Stack
Overflow
questions
with
oh,
it's
soft
deprecated
and
somebody
asked
once
what
does
that
mean?
It's
like
it
means
every
time
anybody
uses
it
on
Stack,
Overflow
I
tell
them
not
to.
A
Yeah,
that's
kind
of
the
reason
why
I
think
warnings
are
more
usable
because
you
copy
paste
number
from
Stack
Overflow,
and
so,
if
you
never
voted
in
intelligence,
you,
wouldn't
they
would've,
never
known
that
you
absolutely
ap
is
right,
I!
Think
to
me
it's
a
sec.
You
know
it
needs
to
be
strong
enough
as
a
warning.
All
right
leg
isn't
like
it
either
doesn't
work.
It
has
a
lot
of
like
known
problems
or
there's
an
actual,
better
feature
available
for
it
and
then
I
think.
In
that
case
the
person
would
just
search
for
that
feature.
A
D
D
A
D
A
D
A
A
I
mean
that's,
it
I
mean
that's
exactly
what
it's
a
person
does.
Actually,
if
you,
if
you
actually
suppress
it,
it
says
you
know,
configure
severity,
none.
It
actually
creates
an
editor
config
file
and
in
the
editor,
config
folder
just
says:
dotnet
Diagnostics
dot.
Whatever
diagnostic
ID,
you
said,
dot
severity
equals
none.
So.
A
A
A
D
Yeah,
all
the
new
things
in
triple
o2
need
to
be
a
new
code
and
furthermore,
if
there
are
things
that
we
think
that
we
would
have
a
fixer
for
we're
like.
Oh,
if
all
the
attributes,
for
example,
we
call
double
O,
10
and
all
of
the
things
all
of
the
types
that
are
the
attribute
type
without
the
word
attribute
on
them
are
double
O
11.
Then
we
could
make
a
fixer
for
double
O
10.
That
says
just
delete
this
attribute
altogether
I
mean.
A
D
D
We
have
here
like
publisher,
identity,
permission
and
publish
our
identity.
Permission
attribute
right.
This
type,
probably
the
the
publisher
identity
permission.
Probably
we
don't
have
an
easy
fixer
of
delete
all
your
references
to
this,
because
maybe
now
you
have
a
unbound
parameter
or
you
have
a
you
were
assigning
a
property
and
we
deleted
the
right-hand
side
of
that.
D
A
Mean
it's
fair,
I
guess
my
only
concern
with
that
is
like
because
I
mean,
strictly
speaking,
like
what
is
the
implication
of
using
the
same
ID,
because
you,
the
convicts,
are
still
called
for
each
occurrence
right
and
you
can
still
say,
I
handle
this
ID
and
then
you
can
decide
to
only
handle
a
subset
of
occurrences
that
have
that
idea,
but
Union
not
on
the
hook
to
actually
provide
a
fix.
Everything
with
that
ID.
So.
D
This
is
just
a
from
something
I
remember
of
one
of
our
analyzers
we
had
with
which
had
a
a
sometimes
works
fixer
that
Manish
said
the
light
bulb
is
triggered
by
the
ID
and
the
fact
that
there
is
no
fix
available
isn't
discovered
until
after
someone
clicks
on
the
lightbulb
and
clicks
the
fly
out,
and
then
it
says:
oh
I'm,
sorry,
there's
no
fix
for
this.
So
it's
hey,
I
have
a
suggestion
for
you.
No
no
I,
don't
and
that's
because
otherwise.
A
D
Would
have
to
run
the
fixer
to
find
out
that
it
produced
no
fix
before
rendering
the
light
bulb,
which
would
be
a
very
expensive
feature.
I
see.
So
it's
the
the
candidate
like
the
diagnostic
is
produced.
There's
a
set
of
filters
that
run
through
it
says:
oh
I
handle
that
ID
and
now
the
light
bulb
is
registered
and
then
later,
when
you
would
go
to
invoke
it,
it
finds
out.
Oh
there's,
nothing
I
can
do
for
you.
D
A
A
B
D
B
B
B
B
A
B
A
See
so
the
main
stick
the
compared
to
ms
generally
happy
with
the
people
who
own
binary
for
matter
us
are
generally
happy
with
it.
So
then
I
guess
the
dissenters
will
become
people
who
use
binary
format
or
but
like
bin
forms
and
others
right,
which
that's
kind
of
what
you
would
expect
to
happen
in
these
cases.
Yes,.
B
D
Yeah
so
I
mean
I
think
that
you
know
for
a
now
relevant
decision
its
if
we
think
that
there's
ever
a
time
that
we're
going
to
get
to
at
least
the
what's
currently
listed
as
dotnet
seven-state,
then,
if
no
one
has
issues
of
the
things
that
we
want
to
do
in
five,
then
that's
probably
the
plan
that
we
should
go
along
with
and
then
the
question
is
what
what
sort
of
public
or
what
sort
of
ref
CS
changes
are
required
for
that.
So
it's
binary
for
matters
obsolete
is
warning.
D
I'm,
sorry,
the
serialize
and
deserialize
methods
are
obsolete
is
warning.
So
now
we
go
back
to
presumably
that's
a
distinct
bucket
and
then
to
my
earlier
earlier
point
of.
Are
we
going
to
add
things
to
the
same
bucket?
Maybe
when
we
move
it
up
to
the
hole
type
that
that's
the
same
ID,
because
it's
still
binary
formatter
is
broken.
Please
don't
use
it
I
mean.
A
B
A
I
guess
what
I'm
asking
is
like
because
to
me
like
that
kind
of
gives
us
most
of
what
we
want
right.
Server-Side
code
generally
doesn't
use
it.
It's
already
documented
as
being
a
problem
and
then,
if
WinForms
continues
to
use
it,
and
so
basically
the
you
know
the
darknet
six
work
never
happens
and
then
the
dotnet
seven
also
doesn't
happen.
Then
maybe
that's
okay
right
because
we
would
say:
oh
I.
D
A
A
If
you
did
see
your
lives,
I
mean
that's,
you
know
it's
not
everything
we
wanted
but
like
it
gets
us
very
close
to
where
you
want
to
be,
and
it
seems
like
that's
also
one
of
the
short-term
things
we
can
actually
do
and
then
the
actually
expensive
work
which
is
refactoring
parts
of
when
functional
use.
It
sure
that
would
be
desirable
but
I'm,
sorry,
but
I.
Guess
from
my
point
of
view
that
did
like
you
know
if
that
takes
longer
than
net
six
or
don't
net
7,
then
that's
also
not
the
end
of
the
world.
B
But
it
at
that
point
we
now
potentially
have
relegated
ourselves
to
well.
If
this
never
gets
done
are
we
satisfied
and
my
answer
to
that
is
now,
and
the
reason
for
that
is
the
given
the
security
vulnerabilities
that
have
been
reported
over
the
past
few
months,
including
the
really
big
one
that
we
released
a
patch
for
today.
We
have
very
strong
evidence
that
people
have
not
taken
our
warnings
regarding
this
type
seriously
and
we
we
do
eventually,
in
the
long
term,
need
to
excise
this
type
entirely
from
the
ecosystem,
yeah
I.
D
D
The
one
concern
that
I
would
raise
with
our
five
plan,
not
saying
the
five
plan
that
I
want
to
remove
anything
from
it.
It's
the
absolution
that
we
add,
which
is
I,
think
the
the
biggest
thing
of
putting
it
in
front
of
someone
of
like
look.
You
have
to
either
go.
You
have
to
go
turn
off.
This
warning
or
pragma
suppress
your
invocation
of
it
and
hopefully
now
you
look
at
the
docs.
It
says
no,
we're
serious
if
you
build
anything
as
that
standard,
oh
you're,
not
going
to
see
that
warning.
A
So
I
can
speak
to
that
a
little
bit.
So
my
first
argument
to
your
statement
is
it's
it's
true,
but
I
don't
think
it's
worth
it
and
the
reason
is
if
you
look
at
new
it'll
org
binary
format
is
used
by
less
than
1%
so
like
it
is
very
mean
to
me.
Nougat
really
is
the
proxy
for
that
Center
or
like,
if
you're,
building
an
app,
and
you
would
be
composing
your
app
in
two
components:
you're
not
using
that
Center.
Oh
you're,
just
using
that
core
app
for
the
most
part
so
like
it's
really.
A
A
But
the
problem
that
we
have
there
practically
speaking
is
we
would
have
to
use
the
obsolete
attribute
as
it
is
defined
in
that
standard
we
can't
add
to
it,
which
basically
means
we
can't
give
the
diagnostic
ID,
which
no
means,
if
we
were
to
add
it.
So,
let's
say
let's
say
the
dotnet
5
plan.
Is
you
absolutely
binary
format
at
the
type?
Not
the
method,
then
I
would
not
do
that
in
air
center,
because
if
you,
if
you
attribute
a
type
and
you
don't
have
an
ID,
then
it's
very
quickly
very
noisy.
A
A
You
know
basically
explode
in
terms
of
how
many
warnings
you
get
but
I
also
wouldn't
ref
the
net
Senate
set
purely
for
obsolete
I.
Think
that
does
not.
That
is
not
worth
the
cost.
In
my
opinion
like,
if
we
already
do
it
sure
we
can
pile
on
by,
like
I
would
not
do
do
a
ref
set
update
purely
for
observations.
A
B
B
B
A
D
B
The
five
plan
is
already
in
progress:
the
we
already
have
PRS
out
to
to
disable
binary
format
or
in
plasm
all
up
to
disable
it
in
ASP
nut
by
default
with
an
opt-in
switch.
We
have
I'm
about
to
push
up
a
PR
that
obsoletes
the
deserialized
method,
so
all
of
these
are
already
in
progress,
because
there
has
already
been
approval
from
these
stakeholders
that
this
is
appropriate.
What.
A
B
We
already
have
the
guidance
document
in
progress.
Rick
and
Genevieve
are
reviewing
that
right
now,
I
already
have
the
AMS
binary
format
or
forward
link
registered
like
on
all
of
this
is
already
in
progress.
The
only
thing-
that's
not
the
only
thing
that's
going
to
slip
past.
Probably
this
week
is
the
tracing
event
source
because
I'm
still
ramping
up
on
how
to
do
that
and.
B
A
B
A
Actually
I
would
prefer
to
use
P
and
E
on
black
on
one
wasn't
to
be
honest
but
like,
but
it
has
different
reasons
because
then
our
tracing
for
these
kind
of
things
is
easier
to
reason
about,
because
not
supported
is
usually
a
state
right,
where's
the
platform.
What
support
it
is
like
a
permanent
situation,
kind
of
thing
well,.
A
Assurance
but
like
no
I'm
saying
like
look
at
if
we,
if
we
I,
mean
I'm,
just
saying
this
from
a
person
who
just
did
the
scan
of
don't
need
framework
or
sir
I've
done
it
platform,
it's
harder
for
us
to
reason
about
this.
When
some
say
it's
not
supportive
as
thoughtful,
not
supported,
because
purple
not
supported
on
the
next
is
kind
of
like
a
given
right
versus
this
one
would
be
like
yeah.
You
know
set
this
setting
or
set
this
competitive
like
and
then
it
might
work
well.
A
B
B
A
B
B
Resource
reader
doesn't
even
have
a
fallback
code
path
that
uses
binary
format
or,
for
instance,
that's
implicit
in
dotnet
8,
and
the
reason
for
this
is
again
with
the
security
patches
that
went
out
today.
We
see
that,
even
if
people
aren't
using
binary
format
or
directly
they're
using
api's
or
libraries,
it
indirectly
and
unbeknownst
to
them.
Relying
on
that
and
that's.
A
B
There
was
a
case
a
few
years
ago
where
there
was
a
partner
team
that
allowed
you
to
upload
res
X
files
to
their
stuff.
It
wasn't
SharePoint
that
it
was
a
SharePoint
like
system
and
you
could
upload
a
resurrect
file
and
they're
like
Oh
a
resurrects
files
XML.
What's
the
worst
that
could
happen
well.
Res
X
can
contain
binary
format
or
payloads,
like
that.
It
allowed
people
literally
to
RCE
their
production
service,
because
why
on
earth
would
they
ever
have
assumed
that
uploading,
a
res
X
file,
was
a
dangerous
thing
to
do
that?
B
A
D
A
D
B
D
D
A
D
A
B
B
Of.Net
will
not
introduce
serializable
attributes
on
existing
types
or
on
new
types,
master,
disk
and
yeah
fastest,
and
this
also
means
that
we
would
go
through
all
of
the
open
issues
in
the
in
the
runtime
repo,
where
people
of
us
can
you
make
this
type
serializable
point
them
to
this
document
and
then
close
all
of
them
as
well.
Folks,
right.
B
B
D
D
Disagreement
with
that
one,
we
control
the
first
bullet
in
six.
Well,
I
mean
we
don't
really
control
the
last
two
points,
though
the
things
like
winforms
WPF,
WCF,
anything
else
in
the
dotnet
org.
We
can
give
them
suggestions,
but
we
don't
write
their
code
and
we
don't
read
their
code.
So
we
don't
we
oh
no,
making
a
statement.
We
don't
actually
own
doing
the
work.
I
mean
we
can
jump
in
and
do
the
work
for
them.
D
Yeah,
but
so
I
think
that
all
the
parts
that
we
can
act
on
from
libraries
were
probably
all
in
agreement
with
and
that
we
would,
as
the
things
that
we
need
partner
teams
to
cooperate
on
as
they
get
completed.
Then
we
start
ratcheting
to
further
steps
in
the
plan
and
if
the
time
line
moves
we'll
be
sad,
but
this
all
seems
reasonable
to
me.
I,
don't
know
that.
There's
anything
else
that
in
this
meeting
we
would
talk
about,
especially
because
now
there's
only
four
of
us.
B
D
And
then
emo
there
was
a
question
in
the
chat.
If
you
didn't
notice
of,
would
we
make
an
analyzer
for
the
net
standard,
T
FM's
of
warning
we've
things
that
we've
made
obsolete
in
future
versions
so
that
you
can
go
back
and
get
that
warning
in
a
four
things
like
the
calves,
attributes
of
hey
you're
building
that
standard?
You
know
that
net
core
doesn't
respect
these
right.
I
think
that's
something
we
can
consider,
but
it's
definitely
for
you
to
think
of
because
it's
the
question
of
value
versus
noise.
E
A
B
There
is
actually
one
open
question
here.
One
thing:
that's
not
resolved
if
you
have
a
sucker,
the
the
original
draft
of
this
was
ambiguous
as
to
whether
as
to
what
methods
exactly
would
be
Marcus
obsolete
in
dawn
at
5:00,
the
current
text
says
only
did
only
the
deserialized
method,
not
serialize
and
deserialize.
Are
we
okay
without
it?
Would
we
prefer
to
do
battle
for
me,
I.
D
Mean
deserialize
is
where
the
danger
lives,
but
if
you're
doing
serialized,
presumably
you're
wanting
someone
somewhere
else
to
do
deserialize
so
yeah
telling
you
hey,
you're,
writing
this
thing
and
at
some
point
we
expect
your
reader
won't
read.
It
anymore
seems
it's
slightly
noisy.
But
to
me
it
feels
like
it's
worth
it
so.
B
D
Doing
doing
it
if
the
type
means
it's
not
just
the
places
that
you're
using
it,
but
if
you
wrote
methods
that
we're
passing
an
object
around
to
configure
it
that
it
Flags
every
single
signature
as
do
something
giving
you
a
release
of
like
here
are
the
four
places
that
are
actually
a
problem
deal
with
that
and
you
probably
deleted
the
rest
of
the
code.
Before
we
made
there
be
more
warnings,
seems
like
a
better
signal-to-noise
ratio
to
me.
B
Yeah
so,
for
instance,
with
principal
permission
attribute,
we
didn't
obsolete
the
type
we
obsoleted
the
constructor
as
error,
and
the
reason
for
that
is
that
allows
you
to
query
to
see
whether
there's
a
principle
permission
attribute
on
a
method
in
your
code.
Using
the
reflection
stack,
we
just
don't
want
you
to
apply
it
to
anything,
but
if
you
want
to
do
type
of
that
sure
have
at
it
eventually,
once
we
obsolete
the
type
itself,
then
you'll
start
getting
compiler
warnings.