►
From YouTube: GitHub Quick Reviews
Description
Powered by Restream https://restream.io/
A
Alrighty,
so
now
we
are
live.
This
is
the
second
round
of
our
api
review,
of
the
I
guess,
dom
apis
for
the
release.json
jack.
Do
you
want
to
take
it
away
and
give
us
a
quick
summary
of
what
changed
since
last
time,.
B
Mostly
a
lot
of
renames
going
like
from
all
the
other
releases
at
the
release
index
stuff
to
like
using
the
groupings.
We
talked
about
last
time
like
the
overall
concept
of
like
a
release
which
previously
or
a
channel
that
would
have
been
something
like
you,
know.net,
five
or
nine
core
3.1.
B
B
I
got
rid
of
all
the
I
numerals
everything's,
backed
down
by
collections
underneath
for
the
product
grouping
for
the
releases
jason.
I
just
extended
the
product
collection,
the
the
reader
collection,
so
I
created
the
prior
collection.
B
I
get
rid
of
all
the
sync
methods,
so,
given
that
the
time
when
there's
actual
I
o
happening,
which
is
either
to
get
one
of
the
json
files
and
parse
it
or
potentially,
if
you
have
a
release-
and
you
want
to
like
download
something
all
these
files
are
typically
hundreds
of
megs
for
some
of
the
installer
files.
So
everything
is
async.
B
B
So
if
we,
if
we
hit
one
of
those
unknowns,
we'll
set
the
support
phase
just
to
an
unknown
for
the
time
being,
and
we
support,
I
don't
think
they
would
change
too
much,
because
we
cover
preview
and
rc
now
lts
won't
change
kind
of
life
won't
change,
so
I
don't
see
unless
we
end
up
doing
something
weird
like
a
beta
or
something
that's
tagged
as
an
alpha
release
or
something
like
that
that
will
just
go
to
unknown
the
I
release.
Interface
is
gone.
B
That's
now,
backed
by
an
abstract
class
called
release
component
from
which
all
the
sdk
and
runtime
releases
derive
from,
and
I
think
some
of
the
ambiguity
has
also
been
removed
in
the
process
where
we
said,
like
you
have
a
release,
but
it
doesn't,
it
didn't
implement
I
release,
it
was
actually
something
like
asp.net,
quarantime
or
the
windows
desktop
runtime.
That
implemented
the
I
release,
and
that
was
confusing,
so
that's
all
been
removed.
A
A
C
So
one
question
I
had
about
the
product
is:
there's
the
latest
runtime
version,
and
so
it
kind
of
comes
back
to
that
kind
of
ambiguity
between
what
we're
calling
the
run
time
and
and
differentiating
that
from
the
the
generic
run
times,
and
so
the
latest
runtime
version
was
a
little
bit
confusing
on
this
particular
class,
and
that
was
something
that
still
is
a
little
bit.
I
find
con
confusing.
I
understand
the
underlying
problem,
but
I
wanted
to
raise
that.
B
B
Yeah,
the
the
latest
runtime
version-
I
mean
it
should
always
match
like.
If
you
look
at
the
three
runtimes,
they
should
all
within
a
release
and
you're.
Looking
at
the
latest
release,
all
all
three
of
the
runtime
components,
if
they
are
present,
should
be
the
same
version
as
what
the
latest
run.
Time
is
the
thing
that
typically
has
a
completely
different
version,
I
think,
is
the
sdk
and
that's
likely
why
why
those
were
introduced
originally.
A
B
A
B
Yeah
yeah,
then
you'll
see
there's
the
latest
runtime
version
and
the
latest
release
version.
The
the
michael
is
that
the
the
point
that's
confusing
for
you
to
the
latest
release
and
latest
runtime
version
or.
C
D
B
A
product
releases,
the
is
the
full
so
so
think
of
the
product
as
what
we
previously
had
like
as
the
channel
so
dotnet
5
is
your
product
within
that
you
have
a
specific
release
of
that
product
which
says
like
five
zero
preview
one
within
that
release.
It
is
made
up
of
a
number
of
components.
It
has
potentially
three
runtime
components
for
core
runtime
desktop
asp.net,
and
then
it
has
one
or
more
sdk
release
associated
with
it
as
well
or
sdk
component.
Rather,
that
makes
up
the
whole
release.
B
D
D
D
All
right
so
the
to
I'm
just
trying
to
think
through
michael's
question,
so
the
I'm
trying
to
find
where
that
latest
run
time
is,
is
that
on
product
release.
D
D
And
also
it
doesn't
have
a
latest.
Oh
it
does
have
a
latest
release
version.
Okay,
I'm
sorry!
I
just
overlooked
that.
So
if
we
have
the
latest
release
version,
what's
the
value
of
having
the
properties
for
the
latest
runtime
version,
the
latest
sdk
version
and
michael
I'm
sorry,
if
I
just
stomped
on
your
question,
I
kind
of
had
the
same
question
and
I'm
trying
to
phrase
it
in
a
way
that
makes
it
concrete.
So
I'm
sorry
for
that,
so
yeah.
B
So
the
the
sdk
version
definitely
would
only
give
you
the
highest
sdk
version
within
the
release
if
there
are
multiple
sdks
and
so
think
of
something
like
what's
a
good
example
2-1,
where
212
ships,
the
800,
600
and
500
sdks,
the
latest
sdk
would
always
be
set
to
the
highest
feature
band,
even
if
there
are
multiple
ones.
E
B
D
All
right
so
now
I'm
confused
about
using
the
word
release,
and
I
have
two
questions,
so
let
me
ask
them
separately
and
shortly
and
then
we
can
go
there.
One
is
that
I
think
you
touched
on
a
really
important
point,
which
is
to
what
degree
should
this
api
reflect
the
underlying
json
file,
which
is
definitely
where
we
started
and
what
degree
shall
it
be?
Should
it
reflect
what
the
user
might
think
about
it?
D
A
second
second
question
is
tightly
related
because,
as
a
human,
when
you
say
release,
I
think
of
something
I
can
go
and
download
something
that
has
a
uri
and
if
you're
saying
that
multiple
sdks
could
exist
in
one
release,
then
in
terms
of
the
files
we
create,
those
are
separate
files
to
download,
and
so
I'm
not
sure
for
me
if
the
name
release
feels
right
for
basically
the
major
minor
grouping,
whatever
the
right
name
for
that
is,
it
seems
like
that's
what
you're
and
I
think
you
might
want
to
mute
it
for
a
second.
B
D
But
three
one:
two:
how
does
three
one
two
have
an
sdk,
that's
what
I
meant.
That's.
How
does
three
one
two
have
a
group
of
sdks
from,
and
maybe
I'm
thinking
about
this
problem.
That's
why
I'm
asking
the
question.
I
am
thinking
about
the
problem
from
the
perspective
of
the
files
we
create,
and
is
that
a
good
way
to
think
about
the
problem,
or
is
there
a
better
way
to
think
about.
B
D
Well,
not
necessarily
because,
if
I
think
of
the
sdk
as
something
I
want,
the
highest
version
of
the
sdk
that's
available
and
what
target
runtime
it
runs
against
is
not
important
to
me,
which
is
the
mindset
we
would
like
our
customers
to
be
in.
You
know
they're
not
right
now,
but
we
want
them
there,
then
the
making
a
strong
connection
between
the
underlying
run
time,
the
sdk
happens,
to
run
on
and
the
sdk
itself
is
not
the
way.
I
think
about
the
problem.
D
D
B
A
pain
in
jack,
it's
okay,
I've
been
so
needy,
been
this
for,
like
the
last
few
months
in
the
json
and
stuff.
So
I
mean
the
the
release
so
far.
Everything
how
we've
approached
this
publicly,
even
like
when
we
announced
like
a
new
release,
it's
it's
runtimes
and
sdks.
B
They
always
go
hand
in
hand
like
there's
a
it's.
The
okay
unit
of
this
is
everything
that
we're
making
available
today,
like
these
things
all
go
together
and
there's
different
parts
of
them.
That's
where
the
release
components
come
in
and
yes,
there
happens
to
be
a
runtime
version
associated
with
it
and
an
sdk
version
or
multiple
sdk
versions,
but
it's
one
cohesive
unit.
D
So
it's
like
our
year,
identifier,
it's
like
our
our
except
if
we
have
a
minor.
So
I
understand
that
part
all
right.
B
D
All
right
all
right,
like
I
said
I
think
I'll,
let
the
meeting
progress
and
then.
D
B
I
don't
know
if
somebody
would
potentially
go
look
at
the
json
first
and
then
go
look
at
the
api
and
say
like
well.
I
want
that
value
in
the
json
like
now
it
was
called
like
a
channel
that
was
difficult
for
me
to
do
just
to
make
that
switch
over.
Like
I
look
and
see
like
oh
there's,
something
called
channel,
I
don't
see
channel
doesn't
appear
anywhere
in
api
there's
now
something
called
product
and
stuff.
So
balancing
that
is
is
proving
to
be
more
difficult
like
in
terms
of
what
you
see
in
the
json.
B
A
I
mean
that's
the
general
problem
when
you
make
your
basically,
your
your
internal
schema
like
the
public
api
right,
but
I
feel
like
that's
probably
not
in
the
long
run
necessary
because
you,
if,
once
you
ship
this
thing,
you
would
tell
everybody,
use
this
thing
and
then
maybe
two
years
down
the
road
you
get
to
change
the
file
format
again
right,
because
now
you
can
say
we
have
an
abstraction
over
the
file
format.
And
if
you
read
the
file
format
directly,
then
you
know
you
have
two
years
time
to
move
to
the
api.
A
Basically,
but
yeah
I
mean
I
don't
know
like,
I
would
generally
say
like.
If
you
want
to
do
the
world
of
service,
I
would
try
to
make
apis
that
make
sense.
Rather
than
sticking
to
the
names
you
already
have.
B
A
I
mean
I'm
not
sure
that
product
is
the
best
thing.
I
think
that's
what
we
said
last
time,
but
clearly
kathleen
does
have
a
hard
time
understanding
what's
going
on
here.
So
maybe
that's
not
the
best
name.
Then.
The
other
question
I
have
is
like
you
were
along
the
lines
of
what
kathleen
was
asking.
So
the
latest
release
version
latest
runtime
version
latest
sdk
version.
Presumably
they
are
part
of
the
same
structure
that
gives
you
the
product
already
right.
So
they're,
all,
oh
one,
you
don't
have
to
fetch
any
data
for
that
right.
A
B
Yeah
yeah,
they
they
go
to
the
the
uri
or
if
you
decide
you
know,
for
closed
off
air
gap
environments.
If
somebody
wants
to
needs
to
host
it
inside,
you
can
change
and
overwrite
the
url.
So
that's,
basically
why
they're
there
and
the
last
one
is
just
to
use
a
support,
any
sort
of
caching
mechanism
where
you
want
to
download
the
file.
You
want
to
use
the
one
that's
on
disk,
but
you
might
want
to
check
online
if
there's
like
a
newer
version
available
and
stuff
like
that,.
B
So
yeah
that
one
can
probably
be
renamed
that
one
specifically
relates
to
the
latest
released,
whether
it's
a
servicing
release
or
if
the
last
release
contended
security
updates.
A
B
A
D
Yeah
that
yeah,
that
seems
it's
actually,
because
would
it
that
what
would
you
say
then,
if
there
was
no
runtime
security
issue,
but
there
was
an
sdk
security
issue
with
the
latest.
D
So
there
is
a
security
issue
inside
that
release,
but
we
don't
know
for
sure
that
it's
the
latest
sdk
that
has
the
security
issue
and
how
are
people
going
to
use
this
flag
if
they're
going
to
think
this
means?
Is
there
anything
in
this
release?
That
has
a
security
issue
and
we
have
a
501
where
the
sd?
D
Let's
say
we
get
things
split
well,
it'd
be
six
okay,
so
we
get
things
fully
split,
so
we
can
do
separate
sdk
releases
from
the
runtime
releases,
which
is
the
world
we
expect
to
live
in
in
a
year,
and
that,
though,
so
we're
bundling
the
the
sdk
and
the
runtimes
yeah
man
and
with
some
of
the
stuff
we're
talking
about
on
on
the
runtime
support
for
the
sdk,
with
the
sdk
running
on
versions.
D
This
I'm
afraid
this
is
I'm
afraid,
our
releases
that
jason
may
have
a
problem
and
the
sdk
and
then
I'm
sorry.
The
api,
reflecting
problems
in
the
underlying
file
seems
like
a
terrible
direction
to
go.
So
you
know
if
we
get
to
a
point
where
the
the
runtime
and
the
sdk
are
fully
separated
and
the
run,
then
the
sdk
can
run
on
more
than
one
runtime,
which
is
the
world
we
would
like
to
live
in.
D
So
we
can
better
support
lts,
then
then,
at
that
point,
then
we
could
have
an
sdk
that
can
roll
forward
across
certain
runtimes.
Now
we
may
not
get
there,
but
that
is
the
general
thing
where
you
know
we're
talking
about
things
like
that
that
better
support
lts
and
I
don't
see
how
that
fits
into
the
releases.json.
A
A
On
your
mental
model,
I
suppose
right
because
you
could
argue
that
the
way
the
release
do
json
is
is
modeled
it's
effectively.
You
know
I
think
of
this
as
like
heading
in
like
logical
grouping
right,
so
you
would
have
a
50
header,
another
5
or
header.
You
have
the
releases.
You
know
five
one,
five,
two,
five,
three
five.
Four!
A
I
mean
we
don't
do
point
releases,
but
you
know
501,
maybe
or
502,
whatever
the
version
ever
since
that
being
and
then
within
those
you
either
have
you
know
a
runtime
release
or
an
sdk
release
right
now,
whether
that
sdk
can
run
on
400
or
300.
It
is
kind
of
orthogonal
to
this
right.
I
think
the
problem
you're
pointing
out,
though,
is
very
real,
which
is
the
we
use
these
version
bands
to
to
where
basically,
100
and
400
could
both
be
considered
latest
right.
A
D
A
A
Fair
enough
I
mean
the
I
mean
yeah.
If
you
rename
is
security.
Update
to
latest
is
security
after
I
guess
latest
release
is
security
update.
Then
that
would
only
be
true
if
the
400
version
would
have
a
security
update
right.
If
the
100
version
has
one,
then
it
would
return
false
right.
Is
that
what
you
would
expect.
D
Yeah
yeah:
that's
where
I
get
confused
on
how
we
use
the
security
update
when
we
have
multiple
components
in
there
and
some
of
those
some
of
those
components
will
be
grouped
under
either
ace.net,
potentially
or
certainly
right
now
under
sdk,
and
we
could
have
multiple
security
statuses
across
those
those
latest
sdks.
I.
A
Yeah,
I
think
at
this
point
I
would
probably
just
say
like
what
do
you
need
this
property
for
at
all?
Like
I
mean
you
could
just
say:
okay,
we
remove
that
and
say
if
you,
if
you
want
to
have
a
if
you
want
to
actually
answer
the
question,
go,
do
await
get
releases,
async
and
then
issue
your
link,
query
against
that
result,
and
then
you
can
make
sure
you
get
the
result
that
you
want.
B
D
Yeah,
I
I
think
that
that
one
remembering
that
the
reason,
one
of
the
reasons
we
built
releases.json-
is
to
drive
the
current
layout
of
a
website
and
that
in
that
current
layout,
the
first
column
says.
Is
it
a
security
release
and
in
that
whole
layout
we
do
not
reflect
the
the
isolation
we
want
to
make
between
the
run
time
and
the
sdk,
and
so
eddie.
D
But
the
history
going,
but
the
direction
we
want
to
go
forward.
I
think
our
safest
bet
is
to
remove
that
that
field
entirely
and
look
at
possibly
some
helper
methods
to
help
people
get
the
answers
to
the
questions
they
want,
which
is
a
separate
question
from
this
part
of
the
api
review.
So
yeah
yeah.
B
I
mean
we
can.
We
can
certainly
take
it
because
there
is
another
way.
Like
emma
said,
you
can
go
and
query
it
plus,
even
if
you,
even
if
you
hit
a
release,
give
them
to
answer
your
other
question.
The
any
vulnerabilities
that
were
exposed
are
listed
on
the
release
itself:
they're,
not
children,
from
the
specific
release
component,
so
the
cve
doesn't
hang
off
the
asp.net
core
runtime
or
the
sdk.
It's
the
release
that
says
this
release
addresses
five
cves
and
that's
all
it
tells
you
that
says:
here's
the
sdk,
here's
the
runtimes.
B
Actually
had
the
vulnerability,
all
I
know
is
there
was
a
vulnerability
and
was
addressed
in
this
overall
release.
So
if
I
get
everything
I
I
I
patch
the
vulnerability
essentially
but
there's
no
way
to
differentiate
and
say
like
it's
in
this
specific
component
and
the
rest.
Maybe
I
can
ignore
because.
D
D
D
Let's
say
we
have
a
cva
against
nuget,
okay,
so
people
that
we
have
you
know
a
million
machines
in
the
world
that
have
a
run
time,
and
that
is
perfectly
fine
and
100
000
machines
that
have
an
sdk
which
is
vulnerable
and
we're
going
to
patch
a
million
machines
that
won't
actually
be
a
patch,
because
we
haven't
even
made
a
change
to
the
runtime.
It's
not
even
a
different
run
time,
so
how
you
know
with
the
clear
possibility
that
the
sdk
can
have
cdes
that
do
not
affect
the
runtime
in
any
way.
D
G
I
don't
know
that
is
security.
Update
here
could
even
be
used
correctly.
I
can
imagine
a
scenario
where
I'm
checking
this
to
see.
If
I
need
to
update
the
version
of.net
on
my
server
or
something
and
somehow
I
skip
a
security
update
and
the
next
release
is
not
a
security
update.
G
B
Accurate
or
is
there
they
are,
they
are
cumulative.
So
if
you
do
a
security
update
and
then
two
servicing
releases,
the
last
servicing
release
would
have
any
previous
security
updates
rolled
into
it.
B
It
wouldn't
be
deemed
a
security
update,
but
it
will
be
deemed
to
have
included
the
previous
fixes.
So
say
we
we
ship
like
a
3.3
runtime
and
we
we
fix
the
security
issue
and
then
we
release
314,
and
that
is
a
servicing
release,
because
this
you
know
a
null
pointer
reference
or
something
like
that.
If
you
have
314
installed,
you
will
get
the
security
fix.
That
was
included
from
313.
G
G
A
B
B
A
I
mean
you,
you
could,
instead
of
having
the
property
right,
you
could
exactly
do
the
the
basically
the
helper
method,
but
you
could
say,
there's
a
method.
You
know
an
async
method
on
product
that
says,
you
know,
has
security,
update
and
you're
passing
what
you're
currently
on
right,
and
this
thing
gives
you
back
a
bool
that
says:
yep
you,
you
probably
want,
or
maybe
it
even
gives
you
back
the
release
that
that
corresponds
to
you
know
the
the
first
update
that
you
should
take.
C
A
Yeah,
so
I
have
one
more
question
I
mean,
let's
just
start
from
the
inner
out
first,
so
we
don't
jump
all
over
the
place.
So
that's
finished
product.
Maybe
so
the
only
question
I
have
for
product
is
create
async
the
right
verb,
because
it
seems
a
bit
to
me.
Create
kind
of
means
you
fabricate
out
of
thin
air
versus
this.
One
would
really
be
more
like
load,
because
you're
basically
going
to
a
particular
url,
and
you
fetch
your
data
effectively.
A
B
F
A
B
G
B
Have
used
address,
I
mean
the
reason
why
I
picked.
It
logically
was
just
a
different
differentiate
between
like
you
need
to
point
this
to
the
thing
that
shows
the
releases
index
json
specifically
or
the
releases
json
and
differentiate
between
those
two,
but
that's
the
only
reason
but
yeah,
I'm
I'm
fine
with
with
address
and
the
dot
comments
can
just
clarify
that.
I
think.
G
C
John
cool,
you
explain
the
get
supported
phases,
member
on
product
collection.
What's
what's
the
intended
use
of
that?
What
is.
B
The
intended
use
of
just
you
know
for
all
the
products
that
I
have
I
mean
you
can
write
it.
It
doesn't
distinct,
basically
in
the
back,
but
it's
just
to
get
a
list
of
a
summary
of
all
the
support
phases
that
you
have
the
typical
place
where
I
use
it
is
I
like
in
my
tool,
I
create
all
the
available
support
phases
and
then
the
user
has
filters
on
the
command
and
they
can
specify
and
say,
like.
B
B
B
B
Products,
but
I
mean
I
can
just
query
it
and
just
set
up
like
a
static
property
for
it
as
well.
I
don't
know
how
much
convenience
we
should
provide.
G
It's
it's
easy
to
add
apis
later.
If
we
find
out
they're
going
to
be
used
by
everyone.
G
G
For
create
async,
I
notice
we
have
some
string
parameters.
I'm
guessing
these
load
from
file
is
that
a
file
path.
G
B
G
Yeah,
we
should
add
a
string
overload
and
rename
this
one
to
create
from
file
or
something.
B
And
then
on
the
under
products,
while
we're
on
the
subject
should
the
get
releases
should
that
stay,
or
should
that
also
become
like
load
releases,
to
people's
symmetrical
with
the
the
product
collection?
One.
A
G
I
mean
it
would
be
good
to
be
consistent
here.
I
don't
know
why
we
would
have
load
on
one
class
and
get
on
the
other.
G
A
G
B
Yeah
yeah
because
it
uses
the
implied
releases,
json,
uri.
A
A
B
G
This
static,
read-only
uri
that
we
have
on
product
collection,
probably
changed
that
to
a
read-only
property.
In
case
we
want
to
change
the
uri
in
the
future.
G
A
So
one
thing
I
wanted
to
ask
earlier
about
the
support
phases
is
so.
What
is
this
eno
member
thing?
Is
this
a
json.net
thing.
B
No,
no,
that's,
I
think,
from
the
data
annotation
stuff,
the
the
the
way
that
the
the
the
support
phases
are
all
abbreviated
inside
the
json
is
lts
and
eol,
but
I
figured
it
might
make
more
sense
to
have
a
more
worthy
neighbor
for
the
name
for
the
actual
member.
So
it
didn't
say
eol.
Would
it
actually
say
that's
life.
B
B
G
A
Yeah,
that's
kind
of
I
mean
this
is
the
problem.
When
you
see
all
those
public
types
right
I
mean
you,
can
I
don't
think
it's
terribly
bad
to
have.
You
know
attributes
on
public
apis
because
they're
not
really
in
your
way.
So
I
think
that's
fine,
but
the
question
is
now:
if
you
want
to
change
it
later,
the
problem
is
now,
when
you,
let's
say
add,
or
remove
attributes
from
those
members,
do
you
expect
somebody
else
to
be
broken
by
this
right?
A
So
if
you
put
effectively
civilization
attributes
on
inner
members,
you
basically
other
people
that
it's
okay
to
serialize
these
guys,
and
you
probably
don't
want
to
support
that.
That's
why
I
would
generally
agree
with
corey
that
you
generally
want
to
split
your
civilization
data
from
your
public
api,
so
you
can
version
them
independently
and
you
don't
own
somebody
else's
civilization.
A
A
B
B
C
F
F
I
just
I
just
hopped
in
so,
but
I
think
I
I
caught
the
question
so
so
maintenance
is
is
used
to
denote,
essentially
the
grace
period
from
when
a
subsequent
release
comes
out.
So
in
the
case
of
an
lts
you
know
2
1
and
3
1
as
soon
as
6
is
released.
Both
of
those
will
go
into
maintenance
and
they'll
they'll,
be
in
maintenance
mode
for
six
months
before
they
go
eol.
F
7-0
will
be
a
current,
so
that
would
only
impact
other
currently
other
current
releases.
That
are,
you
know
in
market.
So
if
we
did
a
say
a
six
one
somewhere,
so
current
is.
C
B
F
Yeah,
what's
interesting
is
the
the
way
our
roadmap
lays
out
something
being
called
well.
Okay,
so
so
5-0
will
be
current,
so
that'll
be
the
first
one
since
2-2
that
that
will
be
labeled
current.
B
F
F
Yeah,
so
so
going
back
to,
I
guess
it's
two
yeah
in
1011
we
had
this
debate
a
lot.
By
the
time
we
got
to
2-0.
We
had
kind
of
sort
of
mostly
settled
on
if
it's
not
ga,
it's
a
preview,
and
so
that's
where
the
whole
preview
thing
came
in.
But
then
there
was,
you
know,
other
agitation
to
call
something
in
rc
so
that
we
could
make
them
go,
live
and
things
like
that.
F
A
Yeah
I
mean
I
I
think
originally
I
was
one
of
those
people
who
said:
let's
drop
all
these
preview
labels
because
we
used
to
have
alpha
beta
and
all
of
that
and
that
to
me
was
not
very
useful.
I
think
between
rc
and
preview.
That
makes
sense
right
because
rc
communicates
clearly,
you
know
you
already
have
a
go
live
license
so
that
separation
makes
sense.
I
agree
with
lilo
that
I
hope
that
we
don't
reintroduce
some
of
the
other
quality
markers
that
we
have.
A
F
A
So
but
then
to
just
evade
the
right,
so
basically
current
is
missing.
We
should
add
it.
I
still
don't
know
when
maintenance
shows
up.
F
Okay,
so
so
the
the
let's
use
the
historical
example
so
2-1
shipped
and
we
had
declared
lts
so
that
one
would
have
an
lts
marking
we
shipped
2-2,
which
was
not
lts.
So
when
that
one
was
released,
it
was
released
as
current.
F
No,
not
necessarily
it's
it's
the
it's.
It's
a
it's
an
in-market
release
for
which
the
its
successor
has
also
been
released.
D
It's
it's
one
that
we
consider
the
only
reason
it
is
out
there
is
because
we
think
you
should
have
a
little
bit
of
time
to
get
off
of
it
exactly.
It's
definitely
a
get
off
of
this
one
message,
and
you
know
it's
yeah
to
keep
things
simple.
We
don't
do
the
same
thing
for
the
last
six
months
or
three
months
of
the
lts
cycle
just
say
get
off
of
it
and
that's
really
the
intention
of
the
maintenance.
D
Right
so
in
december
of
2021
we
will
see
5.0
in
maintenance
mode.
C
F
D
F
A
F
No,
so
so!
Well
so
we
remember,
we
had
this
really
weird
state
and-
and
this
is
we're
still
trying
to
figure
out
how
to
talk
about
what
we're
doing
where
two
one
and
three
one
were
were
both
listed
as
lts
and
so
they've,
just
kind
of
lived
along
side
by
side
is:
what's
happened.
E
B
F
Yeah,
as
we
talked
more
about
it,
the
the
fts
term
is
what
became
current,
because
the
fts
was
just
a
little
too
confusing.
Yeah.
H
D
Maintenance,
we
say
that
we
give
people
a
year
to
move,
and
so
two
one
is
less
than
a
year
now.
So
we
can
talk
about
this
offline
has
nothing
to
do
with
jacques
api.
F
D
D
I
I
could
make
a
case
for
that.
So
let's
do
that
offline,
and
I
want
to
talk
to
you
anyway.
I
don't
think
you
had
joined
the
call
yet
I
have
a
lot
of
concerns
about
the
basic
releases
json
structure
when
we
completely
separate
the
sdk
and
runtime.
So
I
think
you
and
I
should
do
some
forward
thinking
on
what
those
scenarios
might
look
like.
A
A
B
So
I
I
should
maybe
I've
mentioned
it
like
previously
like
on
the
old
one,
because
we
had
like
the
the
single
entry
of
an
sdk
and
potentially
another
block.
That
was
an
array
of
sdks.
I
just
got
rid
of
that.
There's
now
just
one
thing
that
says
sdks
and
if
there's
one
sdk,
you
get
one
back
in
the
collection.
If
there's
three
sdks
you
get
three
back,
but
the
sdk
and
sdks
properties,
there's
just
an
sdk's
property.
Now.
A
D
A
Ahead,
yeah
so
like
on
the
product
release,
we
have
is
preview.
Presumably
this
returns
true
for
both
rc
and
preview
bits.
Should
we
also
have
a
is
go
live
or
should
we
just
expose
the
support
phase
on
the
release
itself.
B
So
that
is
again
a
direct
mapping
of
a
an
attribute
in
the
in
the
actual
json
that
we
literally
look
at
which
it
doesn't
try
to
interpret
any
specific
values
like
if
the
valiant
adjacent
says
it's
true,
then
I
return
true,
which
means
if
the
json
is
incorrect,
it
will
also
return
an
incorrect
value.
B
Potentially,
the
one
thing
that
would
concern
me
if
we
did
any
form
of
interpretation
is
if
the
interpretation
changes
over
time,
people
a
might
become
reliant
on
how
certain
values
were
interpreted
and
then,
if
we
make
changes
that
would
potentially
result
in
incorrect
behavior.
A
So,
for
example,
if
I
do
my
link
query
against
you
know
all
the
releases
and
I'm
trying
to
find
you
know
the
latest
preview
then
sure
I
can
just
order
by
release
and
then
just
say:
where
is
preview
equals
true,
but
then
I'm
wondering
like
does
it
like?
It
seems
you're,
not
really
exposing
whether
something
is
rc
or
not
right
so
like?
How
would
I
do
that
at
all.
B
B
B
If
is
a
preview,
so
it
has
either
like
a
build
label
or
a
pre-release
label
on
it,
and
it's
not
a
a
non-preview
version.
According
to
december,
then
it
will
return.
True.
A
A
A
F
B
D
Well,
I
do
think
that
we
have
a
set
of
functional
questions
that
we're
going
to
want
to
ask
about
these
things,
and
that
would
be
one
of
them
where
that
effort
to
find
out
whether
or
not
our
policy
is
that
it
would
be
go
live
and
I
think
people
care
about
go
live,
not
whether
it's,
I
don't
think
or
see
that
those
letters
mean
anything
to
people,
but
everything
is
go,
live
and
for
now
our
policy
is
go.
Live
means
that
rc
is
the
start
of
the
of
the
name
leaf.
D
I
A
F
F
D
D
F
Yeah,
so
I
mean
if,
if
we
wanted,
if
we
wanted
somebody
to
to
know
that
you
know
by
asking
questions
of
the
jason,
then
when
then
we
need
another
property
in
the
json
in
essence,
because
that's
not
represented
at
all.
A
F
Yeah,
it's
either
it's
either.
You
know
one
of
the
preview
designations
that
we
just
talked
about,
or
it's
one
of
the
you
know
in
support
or
out
of
support
phases.
A
A
A
So
whenever
you
have
a
case
where
something
is
answered
by
the
product,
but
not
the
release,
I
would
say:
that's
that's
an
inconvenient
bug
in
your
data.
F
F
F
That'd
be
a
pretty
short
time
window.
You
know
when
we,
when
we
update
any
of
these
files
in
the
github
repo
there's
a
task
that
runs
that
that
flushes
it
out
to
the
production,
endpoint
so
definitely
possible.
But
it's
a
pretty
short
window.
F
I
don't
know
it's
whatever
the
copy
task
up
to
the
azure.
Blob
storage
is
so
it's
it's.
B
Although
if
the
data
gets
missed,
then
you
can't
have
it
because
I
did
for
a
short
period.
That's
why
the
the
there's
a
method
called
ease
out
of
support,
because
there
was
one
point
in
time
where
I
noticed
that
the
support
phase
wasn't
flipped
over
and
updated
yet,
but
because,
when
the
product
shipped
originally
the
end
of
end
of
life
date
was
published.
So
it
actually
checks
both
the
out
of
support,
phase,
the
support
phase,
four
for
end
of
life
and
the
end
of
life
date
and
if
either
of
those
meet.
B
Because
if
there's
a
discrepancy
between
the
data
or
a
delay,
then
you
might
get
the
wrong
information
back.
And
I.
F
F
Yeah,
I
think,
when
you
hit
that
we
didn't
have
the
automation
in
place
yet
to
keep
the
the
two
locations
in
sync,
okay,
and
so
that
that
got
turned
on.
I
don't
know
two
three
weeks
ago.
A
So
that
means
you
know
to
corey's
point
is
like
yeah,
you
basically
there's
no
way
to
avoid
race
conditions
right,
because
there's
no
database
or
anything
that
would
ensure
any
sort
of
consistency,
but
I
mean
I
personally
don't
believe.
That's
a
huge
issue,
because
you
know
we're
not
like
this
is
not
one
of
those
files
that
will
that
was,
there
will
be
version
super
aggressively
right.
It's
like
every
time
we
release,
which
is
like
you
know,
less
than
a
dozen
times
a
year
right,
which
seems
reasonable.
F
I
suppose
one
thing
we
could
do
with
our
with
our
syncing
automation
is
we
could
we
could
take
care
of
the
flush
up
to
the
azure
storage
into
kind
of
a
stage
location
and
then
you
know,
move
everything
to
the
production
side
with
with
a
single
call,
just
in
a
blob
storage
copy.
The
same
situation
applies,
but
I
think
the
the
latency
would
be
even
shorter.
B
B
G
It
could
be
solved
fairly
trivially
through
some
versioning
of
the
records
right,
like
maybe
the
static
methods
shouldn't
care
about
versioning,
because
they're
kind
of
independent.
But
if
I
call
get
releases
on
a
product
and
it's
inconsistent
with
the
the
product
model,
that's
kind
of
weird
right
when.
G
The
the
file
format,
with
kathleen,
maybe
think
of
adding
some
sort
of
you
know
versioning
on
the
records
so
that
calling
get
products
or
sorry
get
releases
could
be
consistent.
D
The
master
of
this
of
this
file,
we
we
can,
we
can
chat,
but.
D
A
E
F
Entirely
possible,
you
know
we
should
you
know
kathleen
and
you
and
I
and
some
others
should
probably
just
stick
our
heads
together
about
some
things
that
we're
discovering
our
weak
spots
and
and
figure
out
how
they'll
play
out.
D
Yeah,
I
think
that
makes
sense.
I
think
probably
yeah
michael
said
me.
You're
interested
and
report
should
probably
include
jacques
in
that
so
I'll
try
to
set
something
up
here,
while
we're
continuing
and
let
us
sit
down
and.
D
F
D
Do
and
if
I
can
ask
a
quick
time
zone
question:
michael:
are
you
on
the
are
you
in
pacific
coast
fargo?
Where
are
you
at
central.
C
D
Central
time,
okay,
I
don't
want
to
set
this,
and
do
you
work
until
five
local
time,
six,
six
to
your
local
time,
okay!
Well!
Thank
you
for
your
time.
Thank
you.
Thank.
D
D
H
D
B
A
B
So
the
way
it's
structured
in
the
data
like
you
would
see
an
entry
like
in
two
one.
You
would
actually
see
entries
for
the
windows
desktop
runtime
like
the
it
would
be
there,
but
it
would
be
set
to
null
and
then
in
some
of
them
it's
completely
like
it's
not
even
there,
and
some
of
the
other
quirkinesses
that
you
will
find,
especially
in
3-0.
B
B
A
A
Mean
it
seems
so
when
you
say
it's
set
to
null
so
when
I
know
access
to
one's
windows
desktop
runtime,
the
property
itself
is
not
now,
but
now
I
go
to
you
know
basically
what
it
wants
to
release
component.
So
that
means
no
display
version
is
now
files
is
now
name
is
now
releases.
Now
version
is
now.
F
So
the
display
version
is
kind
of
the
prettified
generally
only
with
preview
pre-releases.
Since
we
have
you
know
the
the
preview
dash
build
number
hanging
off
the
end
of
the
actual
version.
The
display
version
just
cleans
that
up
and
does
you
know,
preview
dot,
n.
B
F
F
Was
well
the
the
the
purpose
of
that
thing
was
simply
to
give
the
the
marketing
site
a
clean
thing
to
display
that
right,
that's
work.
A
Yeah
sorry
yeah,
wouldn't
it
make
sense
again,
just
a
display
version
as
a
string,
it
seems
a
bit
weird
to
return
an
actual
parsed
entity
that,
but
you
generally
would
not
expect
people
to
do
comparisons
on
parts
of
the
display
version
right.
Presumably
you
would
just
take
it
as
a
as
a
string
and
print
it
right.
B
A
A
B
It
does
I
just
disagree
with
like
a
union
of
all
the
of
everything.
That's
a
runtime
release
component,
so
it
looks
for
a
runtime.
When
does
this
stop
and
aspirate
runtimes
of
all
the
components
that
are
part
of
the
release?
It
basically
excludes
there's,
basically,
all
the
components
except
sdks,
essentially
that
it
does
so
that
link
acquired.
It
just
spits
out
an
innumerable,
but
it's
backed
by
collection
underneath
which
in
turn
is
backed
by
like
a
list
or
an
array
or
something
so
I
don't
think
anything
gets
exposed
there.
A
A
K
A
Yeah,
so
I
guess
you
guys
can
see
my
actual
screen
right,
not
just
the
virtual
desktop.
Sorry,
it
wasn't
a
depression.
You
only
see
one
screen,
but
so
it
might
be
slightly
disorienting.
When
I
keep
talking
back
and
forth
so
yeah
the
idea,
I
think
to
me
it's
just
pick
one
I
mean
today.
You
basically
have
the
interfaces
exposed
like.
I
would
only
collection
everywhere
that
works
for
me,
but
I
would
not
mix
the
two.
A
So
if
you
go
for,
for
example,
the
runtimes,
I
would
probably
be
shown
as
I
would
only
collection,
rather
than
as
with
only
collection
at
that
point,
but
I
mean
it's
a
good
point,
because
I
think
you
also
have
some
things
explosive
with
only
collection,
right
yeah.
Here
we
go
so
yeah.
I
would
pick
a
strategy.
F
A
I
mean,
I
think
again,
backing
them
by
the
only
collection
is
generally
what
you
want,
because
I
guess
in
get
releases.
It
probably
doesn't
matter
because
you
probably
you
know,
get
an
array
and
return
that
that's
also
fine
but
like
the
fundamental
problem
is
when
you
actually
expose
data
structures
that
you
keep
around
yourself
and
then
people
can
cast
away
the
read,
loneliness
and
just
mess
with
the
data
structure.
A
That's
a
bit
of
a
problem
right.
In
this
case
I
mean
you
probably
compute,
an
array
return
the
array
and
if
the
person
messes
with
the
array,
no
harm
is
being
done
in
the
sense
next
time
they
call
they
get
back
a
new
array
anyway,
but
like
it's
really
problematic
for
persistent
things
where
now,
your
object
model
gets
inconsistent,
because
somebody
somewhere
did
something
wrong
right,
but
yeah
in
terms
of
types
just
pick
a
strategy,
we
generally
prefer
classes,
but
in
your
case
it
probably
doesn't
matter
too
much.
A
B
So,
while
we're
banging
with
like
the
release
components
at
this
point,
there
is
a
files
property
which
gets
your
files.
B
After
release-
and
this
is
maybe
a
question
more
for
lee-
I
was
wondering-
should
I
do
some
magic
so
internally
we
have
a
number
of
files
and
it's
mostly
under
windows,
and
I
think
on
the
mac
os
side
as
well,
where
we
have
installable
files
that
look
like
the
actual
installers,
but
they
have
special
names.
They
either
have
a
gis
or
a
dash
nj,
suffix
or
prefix,
and
those,
I
believe,
are
used
for
marketing
purposes.
F
They
they
were
used
for
marketing
purposes.
They
no
longer
are
so
wherever
those
things
stopped.
They're
they're
not
going
to
come
back
ever.
B
F
A
So
we
talked
about
product
release.
I
guess
release
component
now.
B
A
B
Yes,
I
have
yeah
that
was
one
of
the
the
benefits
of
going
with
the
abstract
class
and
actually
parsing
the
json
directly,
because
I
have
found
myself
a
lot
of
time
going
all
the
way
down
to
like
an
sdk
release,
and
then
it's
like.
I
need
something
back
from
the
top
again
to
find
something
related.
So.
B
So
in,
like
writing
some
good
with,
like
the
the
the
the
new
revision
of
the
api,
I
I
ended
up.
Writing
like
a
little
wrapper
for
myself
called.
I
don't
remember
what
I
called
it.
I
called
it
like
the
release
manager
or
like
a
product
suite,
or
something
like
that
where
I
added
some
extra
wrappers
and
that
was
pretty
simple
to
do.
B
A
Yeah
I
mean
yeah,
I
don't
have.
I
haven't
played
with
the
api,
but
I
think
the
question
is
still
like.
If
you,
if
you
for
each
over
the
thing
and
then
you're
you
know
clearly
in
the
inner
part
of
the
body,
you
still
have
access
to
the
variable
in
the
outer
side.
But
as
soon
as
you
pass
these
guys
around
and
do
something
like
you
print
something
somewhere
I
can
see
the
need.
Is
it's
convenient
if
you
can
just
go
up
by
one,
but
I
don't
think.
A
Yeah,
it's
it's
just
usually
with
object
models.
Whenever
I
didn't
do
it
eventually.
Somebody
said
why
can't
I
do
that
and
then,
if
your
design
wasn't
really
meant
for
that,
it
might
be
very
hard
to
retrofit
later.
But
that's
that's
up
to
you
because
I
mean
you
ever
whenever
you
have
these,
like
you
know
back
pointers,
it
depends
on
how
you
construct
these
guys
where
it
gets
tricky
at
times.
B
B
C
B
So
if
you
want
something
like
that,
you
have
to
construct
like
for
my
convenience,
because
I
always
deal
with
everything
at
once.
I
have
a
sort
of
like
a
helper
class
where
it
adds
everything
and
from
that
helper
class
I
can
head
into
like
either
the
products
or
the
specific
releases
of
a
product
I'm
just
like,
but
but
they're
separated.
Now
in
the
new
design,
I
don't
think
there's
an
easy
way
to
go
back
from
a.
B
B
A
A
A
I
have
a
general
question
like
do.
You
actually
have
constructors
on
most
of
the
stuff
you
do
sort
of
right,
not
for
all
of
it,
but
for
some
of
it.
What's
the.
B
Don't
so
for
the
index
file
like
for
the
product
where
the
actual
product
itself
that
I
don't
think
I
actually
did
a
construct,
because
I
just
have
like
a
private
constructor,
because
it
can
just
be
like
serialized
straight
up.
The
ones
that
do
have
constructors
are
more
like
the
the
actual
release
and
the
release
components
where
I
actually
sit
and
parse
jason
moore
by
myself.
I
take
a
blob
of
the
jason
and
I
I
pick
it
apart
to
get
the
information
that
I
want.
A
So
I'm
not
sure
I
heard
so
you
correctly,
but
basically
you're
saying
for
things
we
use
the
the
the
type
itself
as
the
serialization
as
a
serialized
representation.
You
have
a
constructor
or
because
it
seems
to
me
weird
that
it
can
construct
a
product
in
the
cve.
But,
for
example,
not
a
not
a
component
like
it
seems
like
part
of
it
is
just
instantiated.
Well,
the
other
part
is
not.
B
A
A
B
B
C
Kind
of
on
a
related
subject,
you
know
the
types
appear
to
be
immutable,
except
for
version
the
release
version.
You
can
change
the
major
minor
patch
in
pre-release.
B
Oh,
you
know
why
that
is
yeah.
This
is
the
leftover
from
my
tool.
There
are
places
where
when,
but
this
gets
like
a
lot
more
like
the
nitty-gritty
stuff
that
I
don't
think
we
want
to
expose,
but
there
are
some
cases
where,
when
I
look
at
something
that's
installed
on
a
machine
I
need
to
like.
I
look
at
like
a
raw
like
windows,
installer
file
right.
So
I
look
at
an
msi
and
stuff
and
I
try
to
figure
out
like
this
thing.
B
I
have
a
partial
version
like
in
the
actual
entry
that
goes
into
ad
move
programs
that
sits
in
a
registry
key
and
I've
got
some
version
information,
but
it's
not
it's
mangled
and
compressed,
and
the
bits
are
manipulated
and
from
that
I
try
to
artificially
construct
something
that
would
look
like
a
release
version
to
then
potentially
pinpoint
where
it
fits
in
the
whole
releases,
especially
for
things
that
are
not
officially
publicly
released
so
think
of
like
daily
internal
bulls
and
things.
Those
are.
B
B
Because
at
that
point
I'm
manipulating
bits
on
like
instant
stuff
that
I
just
want
to
quickly
shove
in
and
change
the
version.
I
don't
want
to
construct
a
string
and
then.
K
K
B
B
Yeah,
because
this
was
a
version
yeah,
if
it's,
if
it's
fully
immutable,
then
I
guess
not.
If
the
status
gets
away,
the
carnival
can
go
because
yeah
it
had
the.
When
I
originally
liked
the
implementation
I
sort
of
looked
at
what
version
implemented.
B
Precedence
and
president's
equals.
That's
that's
similar
to
comparisons
because
they
are
technically
similar
to,
except
for
like
the
additional
concepts
that
we
have
like
feature,
bands
and
stuff
like
that.
But
you
can
december
presidents
where
it
like
looks
at
the
it
ignores
the
it
takes
the
pre-release
level
into
account,
but
ignores
the
it's
called
the.
K
B
That's
december
president's
like
if
the
pre-release
is
the
same,
but
the
bolt
metadata
is
ignored.
A
A
A
And
then
I
guess
the
comparisons
that
you
have.
They
are
all
consistent
with
what
compare2
does
by
default
right,
yeah.
B
There's
nothing
interesting
there,
except
there's
one
download
where
it
gives
you
the
option
to
verify
that
because
the
file
hash
is
recorded.
So
if
you.
A
B
So
I
was
trying
to
find
out
like
what
I
should
do
if
the
hash
fails
the
moment.
The
best
thing
that
I
could
find
was
to
throw
like
just
a
cryptographic
exception
with
a
message
that
says
like
the
file
hash
doesn't
match,
it
might
give
the
file
name
and
the
hash
value
that
was
expected
versus
the
hash
that
it
found
or
computed.
A
B
A
F
B
K
I
mean
usually
a
cryptographic
exception
is
thrown
inside
a
cryptographic
operation
that
has
failed,
but
so,
where
specifically,
are
we
looking.
B
In
the
download
async
on
the
release
file,
if
the
the
hash
verification
fails
that
it
throws
a
cryptographic
exception,
I
couldn't
figure,
if
there's
a
better
exception
type
that
I
should
throw,
because
it
is
technically
using
the
cryptographic
apis
to
do
the
verification.
But
it's
not.
K
Yeah,
I
can't
think
of
a
better
thing.
I
guess.
B
The
only
reason
would
be
potential
cost
in
terms
of
like
computation
time,
which
shouldn't
be
that
much,
but
the
files
are
quite
big,
especially
if
you're
doing
a
full
release,
although
yeah.
That
would
be
the
only
reason
why
I
put
it
in
this,
for
somebody
like
saying:
don't
want
to
do
it
for
some
reason,
but
practically
it
doesn't
make
sense
like
if
you're
downloading
a
full.net
core
release
of
all.
B
B
The
truth,
a
little
bit
of
over
engineering
on
my
part
of
like
somebody's
gonna,
ask
for
this
at
some
point.
A
I
mean
yeah,
that's
kind
of
my
question
because
I
mean,
if
I
mean
it
seems
to
me
like
this-
is
not
an
operation.
You
do
really
in
a
hot
loop
right
like
you,
but
as
soon
as
you
fetch
a
file
from
somewhere.
That
is
running
bits.
It
seems
like
you
can't
really
trust
that
without
jumping
through
a
lot
of
hoops,
so
you
probably
should
verify
it
in
all
cases.
A
A
C
A
C
Yeah
one
other
question
about
the
release
file
in
the
download
api
is:
is
the
file
name
parameter
correct?
Should
that
be
that's
the
destination?
Should
that
be
it's
a.
J
C
B
I
think
more
accurate
because
yeah,
because
it
should
it
can
be
full
path.
Yeah.
K
K
A
B
Yeah,
it's
unlikely
that
you're
going
to
be
doing
anything
with
the
the
the
only
places
where
that
might
matter
is,
if
you
grab
something
like
one
of
the
non-windows
files.
That's
like
zip
or
like
one
of
the
tar
balls,
and
you
want
to
do
some
extra
manipulation,
maybe
extract
like
if
you're,
if
you're,
writing
a
tool.
On
top
of
that
like
pull
down
the
file
and
then
you
know,
grab
a
tool
or
you
know
a
targets
file
or
something
like
out
of
like
one
of
the
archive
type
things.
B
Yeah,
because
I
mostly
use
that
for
places
where,
like
I
mentioned
earlier,
like
some
in
some
of
the
release
components,
there
are
some
files
that
get
duplicated
in
multiple
places.
So
when
you
look
at
the
the
overall
product
release
itself
and
you
ask
for
the
files
to
get
all
the
files
back,
I
I
do
a
distinct
on
those.
I.
C
B
A
So
let's
go
look,
look
at
release
component
first,
because
that's
the
base
type.
That
probably
makes
the
most
sense,
which
I
can't
find
here
right
here
in
the
middle.
I
already
did
that
right.
So
that's
fine
and
then.
E
C
B
I've
wanted
a
strongly
typed
version,
but
the
information
there
is
supposed
to
be
human
readable,
so
the
visual
studio
version
and
the
visual
studio
mac
version,
so
the
non-support
ones,
are
the
version
of
visual
studio.
Where
that
specific
release
component
was
inserted
for
in
the
first
time,
which
means
it's
automatically
supported
the
support
properties
denote
it's
a
string,
that's
very
freeform
takes
and
it
denotes
all
the
versions
that
would
support
that
component.
B
B
B
B
B
F
Yeah,
I
think
those
I
think
those
are
something
where
we
could
probably
safely
go
through
and
clean
up
any
data
that
we
think
is
interesting
because
again
it's
it's
for
the
marketing
site,
and
I
can't
imagine
anybody
relying
on
those
things
right
now,
because
they're
such
a
mess.
D
So
yeah
do
you
think
yeah
they
are
a
mess
going
forward,
we're
going
to
simplify
visual
basic
by
never
changing
the
number
again.
So
that's
where
we've
had
most
of
the
problems
in
the
past.
So
at
any
rate
the
should
we
have
that
be
a
string,
though
in
case
in
the
future.
Some
goofy
things
winds
up
in
there
or
should
we
have
it
be
a
formal
version
there.
D
Because
it
kind
of
depends
on
given
the
the
pipeline
we're
using
you're
sort
of
responsible
for
kind
of
the
data
hygiene.
Yes,
so
I
think
it's
a
call
on
your
end.
D
F
D
Actually
be
nine
and
ten,
and
that
the
only
thing
that
would
lea
is
that
if
we
do
it
as
a
string,
are
we
going
to
have
problems
sorting
it
are
we
gonna
want
because
we're
gonna
be
at
ten
really
soon,
and
are
we
gonna
wind
up
sorting
that,
like
back
with
c
sharp
one.
F
D
I'll
ask
you
differently:
we
won't
find
10.
If
we
ask
for
the
highest
version
from
a
string,
we
will
get
nine,
even
if
10
is
available.
B
A
Okay,
that's
fair
and
which
case
yeah.
At
that
point,
you
want
something
you
can
compare.
I
mean
I
guess
that
goes
generally
back
to
like
what
is
what
is
the
purpose
of
these
things
right,
because
if
I
mean
strings
to
me,
make
sense
when
you
display
data
to
the
end
user
or
if
you
want
to
find
an
exact
match,
any
other
policy.
You
want
something
strongly
typed.
F
D
Yeah
well-
and
I
I
sent
you
all
an
invite
to
the
meeting
and
one
of
the
questions
on
there
is
how
we
mesh
together
this
as
a
source
of
truth
for
the
world
and
a
source
of
marketing
site.
How
do
you
think
about
that?
So
yeah.
B
Yeah
I
mean
for
my
tool:
I've
never
needed
like
the
f,
sharp
and
c
sharp
stuff,
but
the
visual
studio
version.
The
visual
steer
version,
visual
stereo
support
well,
but
it
will
become
interesting
for
my
tool.
Eventually,
when
I
I
find
it's
installed
and
I'm
like.
Can
I
install-
and
I
I
determine
like
this
newer
sdk
out-
and
I
want
to
say-
is
this
sdk
compatible
with
this
visual
studio?
Can
I
actually
patch
the
machine
with
this,
and
would
I
break
it
and
I
have
no
way
of
making
that
determination
at
this
point.
D
And
I'm
thinking
of
the
global
tool
that
will
rely
on
this.
That
is
going
to
be
the
the
viewport
for
users
to
ask
questions
of
releases.json
and
a
question
you
know
now.
Maybe
you'd
argue
that
the
the
best
way
to
approach
the
question
of
you
know
what's
the
highest:
what's
the
sdk
with
the
highest
version
of
c
sharp
is
to
say
well
is
the
latest
sdk?
Why
are
we
asking
the
question?
So
I
think
that
that's
one
of
those
things
we
should
maybe
dive
a
little
deeper
on.
D
We
meet
because
we're
out
of
time
anyway,
and
I'm
afraid
I
do
have
to
go
at
time.
I
do
have
another
meeting,
and
so
I'm
gonna
say
goodbye,
but
the
best
meeting
I
could
come
up
with
was
next
tuesday.
So
it's
a
week
before
we
are
gonna
meet
again
for
that
small
group
of
four.
If
you
all
feel
like
we
have
to
do
it
sooner,
then
let
me
know
how
you
feel
about
lunch
meetings
and
such
things,
because
you
guys
are
a
busy
group
to
try
to
schedule
all.