►
Description
Join members from the .NET teams for our community standup covering great community contributions for Framework, .NET Core, Languages, CLI, MSBuild, and more.
A
That
audio
is
not
on
or
something
like
that,
so
welcome
to
the
second
installation
of
our
language
and
one
time
let
me
to
stand
up,
maybe
hopefully
talk
about
community
stuff
as
well
as
stuff.
You
want
to
share
earlier
that
we
haven't
shared
a
blog
post
or
any
other
venue.
Quite
yet
so
let
start
with
introductions
which
one
you
start
yeah.
B
A
So
then,
maybe
that's
just
jumpy
and
now
I
have
to
find
a
button
to
press
it.
Hopefully
I
get
this
right.
Oh
yeah
look
at
this,
so
we
all
be
a
smaller
the
window
and
we
see
my
screen
big.
So
that's
exactly
what
I
wanted
so
I
don't
know
whether
people
saw
that
like
what
about
people
on
the
table.
Did
you
guys
see
the
tweet
from
banana
yeah
yeah.
A
D
C
A
So
if
these
issues
for
and
so
you
wrote
a
very
in-depth
blog
post,
that
is
that
you
should
definitely
check
out
so
it's
you
know
the
picture
kind
of
like
speaks
for
itself
already
like
he.
He
ran
the
the
technical
benchmark
because
you
have
not
heard
of
tech
and
power,
it's
kind
of
an
industry
benchmarking.
A
The
second
kid
you
have
server
handle
before
it
falls
over
so
and
as
with
any
benchmark,
it's
kind
of
a
little
bit
like
you
know,
microscopic
in
the
sense
that
it's
not
a
representation,
usually
for
what
you
actually
do,
but
it
gives
you
an
idea
of,
like
you
know
how
much
bare
metal
you
can
get
out
of
the
system,
and
so
he
has
a
chart
here
and
he
compares
the
things
that
he
cares
about
and
there's
more
stuff
in
there.
But
the
blog
post
goes
into
a
lot
of
detail
on.
A
You
know
how
ASP
networks,
right
now
you
know
where
the
performance,
throughput
improvement
came
from.
I.
Think
two
one
was
our
release
where
we
had
the
most
gains.
I
think
right,
like
maybe
two
to
one,
was
pretty
radical,
yeah
yep,
and
then
you
just
see
like
what
what
ends
up
happening
and
I
get
super
detailed
and
he
could
looks
at
all
the
other
things
as
well,
like
you
know
the
my
just
my
sequel
benchmark
as
well
howdy-ho.
Well,
that
does-
and
you
know
it's
pretty
much
something
you
should
check
out
if
you've
never
seen
this.
A
He
also
talks
about
community
and
donate
corn
asp
net
core,
and
the
funny
thing
is
that
somebody
I
think
on
his
tweet
complained
that
you
know
it's
Microsoft.
So
there
must
be
some
sort
of
negative
aspect
to
that.
Ben
just
said
look,
this
is
this
is
an
open
source
platform.
It's
not
accords
money
on
a
Linux
box
like
how
much
money
you
want
it
and
I
think
it's
a
it's
pretty
much
to
like
I
mean,
of
course,
I
reapply.
A
C
Yeah,
this
was
as
far
back
as
yesterday.
C
Yeah,
so
there
was
this
developer
and
Natalia
Portillo,
who
I
think
is
from
Spain
girly
and
geek
handle,
and
so
the
question
was
with
all
the
hype
of
memory,
memory
of
T
and
span
of
T.
There's
one
thing:
one
unclear
thing:
if
I
want
to
access
memory
of
T
was
with
Rey
indexers
I
want
to
use,
I
must
use
dot
span,
but
is
that
doing
a
memory
copy
that
would
defeat
the
purpose?
The
very
good
observation
that
would
indeed
defeat
the
purpose
of
that
was
the
case
and
so
yeah.
C
So
we
there
is
no
copy
at
all,
which
I
later
said
somewhere.
It's
actually
a
very
cheap
operation
to
get
a
span
from
a
memory
of
T,
but
we
should
back
up
just
that
tiny
tiny
bit,
which
is
so
there
are
these
two
types
span
of
T
and
memory
of
T
and
they
kind
of
are
intended
for
different
reasons.
So
actually
just
start
building
this
app,
it's
actually
a
Sudoku
or
f5.
C
We've
been
saying
that
right,
Sudoku,
solver
and
so
I'm
using
memory
of
T
and
span
of
T
in
them
and
it
I
have
to
say
this
is
actually
the
first
app
where
I've
really
tried
to
use
them,
and
my
experience
is
of
using
span
of
T
and
memory
of
T
feels
very
similar
to
learning
how
to
use
task
and
an
async
weight
correctly,
and
so
you
eventually
kind
of
figure
out
what
the
pattern
is.
But
it
feels
like
quite
strange
at
first.
The
other
thing
is
super
funny
about,
and
I
was
gonna
have
Steven
tobbe.
C
This
cuz
he's
the
one
who
gave
me
this
little
code
sample
this
memory.
Memory
pin
thing
feels
very
similar
to
me
too,
like
if
you
want
to
wrap
some
kind
of
async
work,
particularly
if
it's
using
the
old,
a
sink
system
use
this
thing
like
completion
source
yeah.
The
memory
pin
thing
this
memory
handle
memory.
Pin
thing
feels
similar
to
me
so
that
I
don't
know.
If
that's
a
very
good
analogy,
maybe
it's
probably
horrible
analogy
but
anyway.
A
A
It
allows
you
to
pass
this
around
as
one
unit
and
then
there's
no
api's
that
allow
you
in
a
safe
way
to
get
a
pointer
right
because
in
c-sharp,
if,
as
soon
as
you
want
to
talk
about
anything
with
an
asterisk,
you
have
to
be
in
unsafe
context,
right,
yeah,
and
so
the
interesting
idea
of
span
is
that
at
the
very
bottom
of
your
stack,
let's
say
a
web
server,
for
example,
and
the
u.s.
gives
you
a
buffer
where,
basically,
your
network
output
goes
in
and
then
what
you
do
is
you
you?
A
You
have
your
pipeline
right,
Yahoo,
HTML,
parsing
whatever,
and
then
you
just
pass
the
span
up,
stack
right,
and
so
everybody
else
who
just
wants
to
read
or
write
stuff
they're,
just
basically
get
a
span
and
span
as
an
index
around
the
length
and
you're
doing
whatever
you
want.
If
you
have
had
some
unsafe
code
in
the
middle
that
needs
to
access
this
thing
as
a
pointer
then,
yes,
all
you
have
to
do
is
basically
get
get
the
memory
address
and
do
whatever
you
want
to
do.
And
then
you
just
go
about
your
business.
A
C
I
think
in
this
particular
case,
so
everything
you
said
made
sense
in
the
case
that
you
said
the
web
server
and
your
your
web
app
would
be
intuitive
assemblies,
and
so
one
would
be
compiled
unsafe,
yes,
the
first
one
and
the
app
one
would
have
the
pleasure
of
not
being
compiled
and
luxury
of
not
being
compiled
unsafe.
In
this
particular
case,
you
could
imagine
that
maybe
this
person's
building
a
library
but
maybe
they're
just
doing
writing
this
code
as
part
of
their
application,
at
which
point
compiling
unsafe,
is
less
pleasant
right.
A
C
And
it
also
depends
completely
on
what
the
workflow
slushed
usual
each
pattern
of
an
API
is
like
the
thing
I
was:
writing
I
wanted
to
be
able
to
new
up
this
class
and
so
that
it
has
to
hold
on
to
a
memory
of
T.
We
can't
just
deal
completely
it
gives
out
spans,
but
it
has
to
hold
on
to
this
to
this
and
data
that
it
actually
gets
from
somewhere
else
right
in
the
case
of
this.
A
It
yeah
why
not
the
interesting
thing
with
span
is
that
span
is
a
type
it's
a
new
kind
of
type
right
as
to
what
we
call
ref,
only
structs
or
if
only
times
so,
basically,
what
this
does
is.
It
doesn't
allow
the
colleague
to
stash
it
away
right
and
that's
the
memory
of
teachers.
I
was
just
talking
about.
A
This
device,
IO
control,
would
actually
take
a
spared.
What
you
would
do
is
you
would
pin
the
memory
constructor
span
over
that
over
that
address
base.
It's
a
new
span
passing
in
a
point
in
the
length,
and
you
have
a
span
over
that
native
memory
and
then
I
can
call
the
device
I
all
method
and
then
Wednesday
method
returns.
I
can
I
know,
I
can
unpin
the
memory,
because
this
guy
could
not
have
half
stashed
away
the
address,
so
that's
safe
right.
Otherwise
it
wouldn't
be
safe
because
the
guy
might
refer
to
memory.
A
That
now
is
gone
right.
What
is
elsewhere
right?
So
this
is
the
other
nice
thing
when
you
think
of
a
network
stack.
Is
that
at
the
very
bottom?
Is
somebody
allocates
memory
reps?
It
nurse
pan
passes
it
out
and
then,
as
the
stack
on
lines,
you
know
you
can
know,
do
something
interesting
with
that.
The
only
downside
is,
it
doesn't
work
well
with
async
right
as
soon
as
you
do
acing
operations
you
have
to
let
the
stack
unwind,
then
you
get
a
completion
later
and
that's
a
memory
of
T
comes
in
yeah.
C
A
C
B
C
Got
this
method
I've
got
this
for
each
inside
of
it,
and
it's
like
oh
well.
Clearly,
I
just
want
to
have
an
eye
sinking
in
my
bowl
of
span
of
T.
Wouldn't
that
be
awesome,
yes,
but
you
can't
do
that.
You
can't
hurt
for
the
exact
reason
that
you,
just
by
the
way,
I
tried
anyway.
This
is
a
good
conversation.
I
actually
so
I
had
an
email,
third
owner,
Stephen,
Tobin
and
Yann,
and
they
were
actually
basically
just
I
was
their
sneakernet
gopher.
For
most
of
this,
these
responses,
it's.
C
C
A
C
D
A
C
A
C
A
B
B
He
has
a
really
deep
perspective
on
Visual
Basic,
what
it
means
and
the
differences
between
Visual,
Basic
and
c-sharp,
and
he
put
together
this
blog
post,
which
printed
is
like
58
pages.
So
it's
an
investment.
What
I
really
think
folks
should
definitely
do
is
take
a
look
at
it.
Can
you
scroll
down
to
where
the
rules
are,
because
it's
just
gives
you
an
idea
of
what
yes.
A
B
B
It's
really
big
anyway,
absolutely
it's
absolutely
fantastic,
blog
post
and
if
you
touch
Phoebe
at
all
its
in,
if
it's
in
your,
if
you
have
any
VP
assets,
if
you
program
baby
every
day,
what,
if
it's,
if
it's
anywhere
there
or
if
you
just
want
to
see
an
absolutely
amazing,
in-depth
exploration
of
a
language
bookmark
this
thing
and
take
a
look
at
it
when
you,
when
you
have
a
question
or
you
have
some
time
to
look
over
it,
because
it
is
an
amazing
piece
of
work.
Oh
yeah.
D
Especially
if
you're
just
into
language
design
as
a
whole,
like
if
you're,
not
a
VB
developer
or
even
a
c-sharp
developer,
like
just
sort
of
the
depth
that
he
goes
into
in
terms
of
discussing
trade-offs
and
like,
for
example,
like
how
the
rules
for
null
comparison,
work
for
V,
B
versus
C
sharp
and
how
they're
slightly
different
for
certain
things
and
how
that
enables
certain
things
and
disables.
Other
things
is
like
it's.
B
Really
great,
you
guys
talked
a
lot
about
why
some
things
happen.
So
some
things
are
just
because
it
came
from
be
sick.
Some
things
are
because
a
particular
decision
was
made
and
a
little
bit
of
context
on
that,
and
so
it
gives
you
a
little
bit
of
insight
on
the
fact
that
languages
on
he's
so
inner
Dave
and
so
long-term
and
I'm.
Just
waiting
till
you
do
this
for
four
F
sharp,
hey.
B
This
is
not
a
spec,
though
this
is
really
it's.
It's
and
plus
it's
much
more,
it's
much
more
entertaining
than
a
spec.
It's
not
a
formal.
It's
like
here's,
the
story
of
VB,
and
so
anyway,
it's
an
absolutely
fantastic
post
and
I
would
encourage
that
people
go
look
at
it
just
to
be
in
awe
if
nothing
else,
because
it's
one
of
the
most
protein
Oh
big
deal
kind
of
post
I've,
seen
it
a
long
time
so
and.
D
What
I
find
is
also
even
more
impressive
about
this
is
I,
think
I
counted
only
one
place
where
in
the
Y
section
he
was
like
I
have
no
idea
every
other
case.
He
like
knew,
even
though
he
didn't
always
implement
all
that
stuff.
So,
like
you
can
tell
it,
he
actually
went
back
and
he
actually
asked
all
of
these
people
who
had
worked
on
this
in
the
past
and
said
like.
Why
is
this
the
way
that
it
is.
B
He
didn't
have
the
runtime
and
I
was
just
in
a
meeting
yesterday
about
the
steps
we're
going
to
take
in
terms
of
when
things
are
going
to
come
out
in
preview.
Three
in
preview,
a
four
in
preview,
five
across
the
the
VB
getting
that
the
VB
runtime
in
the
hands
of
people.
So
they
can
do
nominate
core.
That
looks
like
VB,
instead
of
not
looking
like
VB,
so.
A
B
A
Obvious
features
that
are
missing
right
and
it's
always
hard
to
explain
to
people
like
very
much
why
language
design
is
hard
I
mean
if
you
never
implemented
a
compiler.
If
you
never
thought
about
how
expressions
compose
it's
very
subtle,
to
talk
about
things
like
well,
what
rules
do
you
have
for
conversions
and
like
if
you
apply
them
and
you
select
the
method?
What
would
happen
which
one
would
you
end
up?
A
Picking
and
interesting
thing
when
I
think
with
this
post
is
that
he
walks
to
all
these
corner
cases
where
we
be
made
different
decisions
from
C,
sharp
right
and
that's
I
think
part
of
the
reason
why
the
languages
can
feel
and
practice
fairly
different.
You
know
modulo
the
fact
that
there
are
different
keywords
right
and
I
think
that,
if
you're
interested
in
language
design
in
general,
even
if
you
dislike
VB
I,
think
it's
a
very
good
post
to
read
and
small
things
like
this
I
didn't
even
know
about
this.
A
You
think
this
is
something
anybody
should
know.
I
am
I
had
no
idea
that
Mimi
does
it.
Smart
quotes,
it's
cool,
so
another
thing
I
wanted
to
a
little
bit
pitch.
Is
that
I
don't
know
about
you,
but
like
I
have
a
hard
time
keeping
up
with?
What's
going
on
in
the
wild
and
there's
a
few
interesting
projects
that
are
going
on
so
at
Thompson,
I
think
he's
back
at
Microsoft
he
used
to
work
on
visual
studio,
/
t
FS,
/
git
integration.
A
A
But
he
always
was
working
I
think
in
something
ready
to
version
control,
so
so,
but
what
he
did
is
yes,
the
developer
tools,
weekly
newsletter
you
can
just
set.
You
know
some
submitting
your
own
email
address
and
I've
done
this
and
there's
no
spam
right.
It's
easy.
It's
just
hand
curating
things
that
are
interesting.
So
let
me
just
look
at
this
one
here
he's
talking
about
here's
some
interesting
articles.
D
Best
thing
you
should
talk
about
this
yeah,
so
this
is
a
related
thing,
is
called
F
sharp
weekly
run
by
Sergey
T
on
I,
think
he's
from
Belarus,
so
I
don't
know
the
correct
pronunciation,
but
he's
been
doing
a
weekly
mailing
list
but
like
he
takes
like
maybe
like
one
week
off
a
year
or
something,
and
he
just
he
a
grenade,
they're
interesting
blog
posts,
things
that
are
happening.
The
F,
sharp
community
things
that
are
happening
like
on
our
open
source,
repos
and
things
like
that
and
just
list
them
all
out
tweets
it
out.
D
He
gets
thousands
of
readers
each
month,
it'sit's
fantastic,
f-sharp
weekly.
If
you
subscribe
to
it,
you'll
basically
just
get
something
each
week
of
like
anywhere
from
ten
to
like
fifty
links,
just
curating,
absolutely
everything
that's
going
on
in
the
f-sharp
community,
and
so
you
can
get
a
lot
of
really
excellent
blog
posts,
not
interesting
ways
to
do
things.
D
There's
like
really
good
pull
requests
that
are
happening
not
necessarily
for
the
F
sharp
codebase,
but
also
just
an
F
sharp
community
projects
will
get
linked
to
here
so
like,
if
there's
a
new
feature
coming
in
to
say
iodide,
which
is
a
cross-platform
IDE
for,
while
ID
plugin
for
visual
studio
code.
If
there's
like
a
really
interesting
feature
going
in
there,
he'll
link
it
there,
so
you
could
see.
Oh
there's
this
new
feature
coming
into
tools
that
I
use,
for
example,
and
so
like
it's
really
great
from
a
discoverability
standpoint.
D
He
also
actually
helps
hosts
an
advent
of
basically
an
advent
of
code
sort
of
thing
for
F
sharp
during
Christmastime.
So
basically
we'll
start
for
the
month
of
December,
not
necessarily
Christmastime
and
basically
what
he
does
is.
He
does
a
call
for
blog
posts
about
a
couple
weeks
in
advance
and
basically
just
wants
people
from
the
community
to
write
a
blog
post
corresponding
to
a
particular
day
in
December,
and
then
he
publishes
it
that
way
and
then
it
by
the
end
of
December.
D
You
have
just
this
utterly
massive
list
of
posts
of
just
cataloging,
just
kind
of
interesting
things.
You
can
do
with
the
language,
so
this
year,
I
think
he
had
like
three
posts
per
day
like
three
like
individual
separate
posts
per
day
coming
from
people.
He
kept
running
out
of
space
and
having
to
allocate
more
more
space.
So
there
were
a
lot
of
people
that
were
really
interested
in
this,
and
so
there's
an
incredible
backlog
of
just
really
great
posts
as
well.
So
I
recommend
checking
it
out.
So.
D
B
D
D
C
There
are
some
there
are
a
bunch
of
mailing
lists
or
it's
really
bla
daily
blogs
like
there's
man,
I,
can't
even
remember
them.
How
I
used
to
follow
them
super
closely,
but
there's
like
some
morning
money,
a
cup
of
something
there's
a
bunch
that
we
get
pulled
into,
but
they're
more
general
than
done
that
they
just
happen
to
get,
and
we
just.
C
D
C
D
C
D
So
he's
he's
one
of
those
really
cool
people
where
we
will
file
an
issue
would
like,
for
example,
we
did
some
investigative
work
on
memory
allocations
in
the
tooling,
and
he
immediately
just
jumped
in
and
said.
Oh
I
think
I
know
how
I
could
fix
this
problem
and
within,
like
a
day,
he's
submitted
a
pull
request,
and
so
there's
actually
three
pull
requests
that
are
going
to
be
going
into
Visual,
Studio
2019
that
are
all
his
work.
One
of
them
is
this
one.
This
was
pretty
cool,
I
think
called
remove
yeah.
D
So
this
is
this
one's
actually
more
of
a
cleanup
effort.
This
is
him
basically
looking
at
stuff
and
going
oh
there's
just
a
bunch
of
kind
of
dumb
stuff
going
on
in
this
code.
Let
me
just
clean
it
all
up,
and
so
did
there
were
a
few
others
that
are
rather
impactful
the
way
that
we
do
format
strings
the
way
that
the
compiler
reads:
F
sharp
format
strings.
D
He
completely
changed
so
that
it
doesn't
like
it
had
this
problem,
where
it
would
reallocate
a
whole
bunch
of
very
large
strings,
and
that
would
throw
that
on
a
large
object
heap
and
so
then,
in
a
long-running
process.
The
large
object
he
would
get
a
lot
bigger.
Eventually,
the
GC
has
to
pause
all
threads.
You
get
a
big
UI
delay
in
Visual
Studio
that
kind
of
sucks.
He
just
completely
made
that
problem
go
away
and
there
was
another.
D
We
have
an
algorithm
inside
of
our
compiler
that
computes
a
string
distance
so
like
if
you
mistyped
something
and
you
get
a
compile
error,
saying
I,
don't
know
what
this
identifier
is.
The
compiler
can
actually
generate
suggestions,
it'll
that'll
say:
oh
did
you
mean
any
one
of
these
like
five
things
that
are
sort
of
generated
by
a
score,
that's
calculated
by
a
string,
distance
algorithm,
and
so
he
cleaned
up
that
string,
distance
algorithm
so
that
it
does
just
less
work
as
a
whole.
So
you.
A
C
D
C
D
I
think
this
is
the
one
that
I
that
I've
mentioned
about
the
foremast
rings,
and
it
was
actually
it
was.
It
was
just
a
part
of
the
codebase
that
was
kind
of
just
silly,
like
in
retrospect,
you're
like.
Why
would
it
ever
be
implemented
the
way
that
it
was?
But
you
know
you
only
get
that
sort
of
thing
in
retrospect,
like
you
know
hindsight,
20/20
and
so
he's
the
one
that
allows
us
to
have
hindsight
of
20/20.
There's
a
comment.
This.
C
B
A
D
B
D
Of
the
mastermind
behind
this,
the
idea
being
that
there's
a
whole
lot
of
really
really
smart,
f,
sharp
programmers
out
there
and
sort
of
a
wide
variety
of
domains,
and
increasingly
we've
been
seeing
like
in
the
past.
There
was
a
lot
of
sort
of
financial
scientific
stuff
associated
with
F
sharp
we're
now
seeing
a
ton
of
stuff
associated
with
the
web
and
with
cloud
computing.
D
If
we
host
a
challenge
to
basically
have
two
forms
of
submissions
that
people
could
have
where
one
is
like,
you
submit
like
a
paper
or
a
technical,
blog
post,
or
something
just
talking
about
cool
and
interesting
stuff,
that
you
can
use
the
language
for
and
like
sort
of
demonstrating
a
unique
thing
that
the
language
offers
for
the
domain
or
like
an
interesting
pull
request.
That
sort
of
shows
things
that
you
can
do
in
code
that
are
really
unique
in
that
you
won't
be
able
to
find
in
other
languages.
D
You
can
just
submit
that
through
the
F,
sharp
challenge,
and
then
it
gets
voted
on
by
some
of
the
experts
in
the
community
is
sort
of
like
you
is
this
something
that
we
think
is
really
cool
and
there's
prizes
like
everybody
who
submits
gets
a
prize
like
like
you
get
stickers,
and
you
know
your
name
gets
listed
on.
The
thing
is
like
somebody
submitted
and
then
the
people
who
win
get
like
sort
of
a
community
title
like
official
f-sharp,
recognized
that
sharp
expert
by
the
f-sharp
Software
Foundation,
so
yeah.
D
A
What
you're
saying
and
actually
there's
another
thing,
I
just
thought
about
this,
which
we
also
haven't
plucked
today
before
I
think
under
the
red
foundation.
There
is
this
thing
here:
the
announcing
the
foundation
of
membership,
so
so
much
of
the
F
sharp
there's
a
net
foundation
which
hopefully
have
heard
of
at
this
point,
and
otherwise
we
need
to
make
more
for
it
there.
A
Although
money
is
not
the
issue,
if
you
don't
have
the
money
view
of
less
money
or
you
don't
have
any
money
that
you
can
still
join
the
foundation,
but
the
idea
is
you
want
to
use
the
membership
to
make
the
information
more
independent
of
Microsoft
in
terms
of
money,
but
also
in
terms
of
people
in
terms
of
influence.
So
the
idea
is
that
this
membership-
you
also
get
the
right
to
vote,
for
example,
to
become
a
board
member
to
vote
for
other
board
members.
A
So
this
is
like
our
continued
quest
of
like
making
the
dotnet
Foundation
more
meaningful,
more
impactful
to
people
outside
of
Microsoft,
and
so,
if
you
have
not
done
this,
go
here,
check
it
out,
read
it
and
then
apply
and
I
think
the
first
round
of
submissions
I
think
is
reviewed
since
I
think
it's
a
Friday
and
I
think
there's
another
wrong
coming
so
by
all
means,
go
there
check
it
out.
B
In
Mac
said:
oh
my
gosh,
so
it's
like
yeah,
you
know
what
who'd
ever
need
this
path
longer
than
256
characters.
Who'd
ever
need
that.
So
this
has
been
huge.
You
know
a
huge
journey
for
for
dotnet
and
visual
studio
and
we
keep
in
Windows
and
Windows
and
Windows.
So
it's
it's
a
stepwise
journey,
so
it
gets
fixed
in
one
place
which
exposes
the
next
place
that
it
needs
to
get
fixed,
but
there's
action
going
on
here
and
that's
really
exciting,
so
you've
got
max
that
up.
B
So
that's
great,
so
msbuild
is
now
supports
max
path
and
Rainer
wrote
a
fantastic
blog
post,
not
a
blog
post,
I'm.
Sorry,
it's
an
issue,
and
so
we'll
have
the
link
in
the
get
this
link
up.
That's
a
good
question:
I
guess
yeah!
So
anyway,
this
wound
up
this
is
this
is
a
comment
on
an
issue
okay,
so
we
need
to
get
it
probably
up
a
little
bit
more
than
that,
but
there's
a
few
things
you
have
to
do
to
make
this
work
and
one
of
those
is
to
enable
Windows
to
handle
max
path.
B
Obviously,
it
wouldn't
work
without
that
and
right
now
this
is
works
on
the
command
line,
we're
good
to
kind
of
gold
on
that,
but
visual
Studios
does
some
more
steps
to
take.
The
cool
thing
was
when
this
got
kicked
off.
One
of
the
things
Rainer
did
and
I
just
want
to
point
this
out,
because
it's
part
of
how
like
we're
all
working
together
and
more
of
a
collaborative
open-source
manner
that
is
really
helping
everybody.
So
Ryan
are
finish.
B
B
There's
a
PR
and
a
review
right
now
to
get
the
next
step
done
and
so
I
don't
know
whether
this
process
is
I,
don't
know
when
we
say
it
works
everywhere
and
I
realized
that
it
kind
of
doesn't
work
until
it
works
everywhere,
but
it's
been
really
exciting
to
take
one
more
step
and
kind
of
a
big
step
in
this
process
that,
at
one
point,
looked
like
you
know
this
mountain
that
can
never
be
climbed
and
one
little
piece
at
a
time
we're
actually
getting
actually
getting
there.
Yeah.
A
My
dancin
there
on
the
vclt,
I'm,
like.he
I,
think
started
about
I,
would
say
a
year
year
and
a
half
ago
reaching
out
to
the
shelter
member
knows
the
longer
yeah.
Well,
I
think
Jeremy
started
about
a
year
ago.
Minute
feels
quicker
might
be
long
thing
is
that
he
I
think
he
talked
to
various
people.
You
talk
to
the
shell
team
to
the
people
that
own
CMD
right
the
con
host
people.
He
talked
to
you
know
the
way
32
folks
and
every
team
you
talk
to
everybody
said
man
max
5.
A
A
But
it's
and
so
like
he
completed
the
circle
and
then
the
end
he
concluded
like
well.
Everybody
thinks
that
they
are
fine,
but
everybody
else
thinks
everybody
else
is
screwed.
So
maybe
there
is
something
we
can
do
here
and
then
that
a
massive
meeting
a
few
times
and
they
concluded
yep.
We
can
fix
this
problem.
It
just
becomes
very
iterative
right
and
I.
A
Think
people
still
think
that
max
per
issue
is
easy
to
fix
and
if
you
believe
that,
like
I
encourage
you
to
read
this
fairly
old
blog
post
from
2007
from
Kim
Hamilton,
she
did
it
kind
of
an
exhaustive
approach
similar
to
Anthony
on
terms
of
max
path.
So
it's
a
it's
a
three
part
blog
that
talks
about
all
the
issues
with
max
path
and
like
how
it's
meant
realizing
itself
in
API
interactions
in
application
behavior
and
what
we
can
do
in
dotnet
to
fix
this
and
her
conclusion
is.
We
can't
do
anything
right.
A
A
If
they
blow
up
well,
then,
if
we
don't
ever
fixes
that
you
know
it,
both
it
will
fix
itself
and
for
the
most
part,
most
applications
actually
have
switched
to
a
different
kind
of
API
is
where
they,
where
they,
you
know
the
max
path,
constant,
isn't
implied
by
the
API.
So
things
will
just
generally
work
and
I.
A
Think
that's
where
we
are
right
now
is
that
you
know
we
basically
get
to
the
point
where
every
single
component
works
with
longer
paths
and
then,
if
you
look
at
the
build
process
where
you
skate
the
processes
or
other
things
that
eventually
more
and
more
stuff,
just
work
right
and
I
think
what
what
would
take
us
another
five
years,
probably
until
literally
everything
will
work,
then
you
care
about,
but
you
know
I
think
vs
is
not
on
the
path
yet,
but
msbuild
is
many
of
the
two
of
them
is
vulgar.
A
B
Was
really
up
to
I
felt
like
you
did
it's
gonna
be
another
five
years
and
then
how
fast
that
we
got
attention
on
that
issue
was
like.
Maybe
it
won't
be
five
years,
maybe
it'll
be
faster,
but
it's
still
it's
still
journey.
It's
still
not,
but
but
something's
work
I
mean
it
depends
on
what
you're
doing
and
so
some
things
work.
And
you
know
it's
it's
a
journey.
We
keep
keep
going
down
that
one
so
yeah
I
was
I
was
super
happy
that
we've
we've
got
that
in
and
in
and
moving
forward
so
yeah
another.
A
C
Context
on
this
and
why
we
didn't
fix
this
issue
in
2007,
for
example,
is
we
had
this
concern
that
people
would
generate
files
as
part
of
a
dotnet
application
and
then
not
be
able
to
view
them
and
explore
right,
which
felt
a
bit
untenable
at
the
time?
More
recently,
we've
rejected
that
as
a
as
a
concern
right.
A
Well,
there's
also
I
think
2007.
We
still
care
about
sandboxing
right
and
this
is
whole
problem
of
like
path
when
you
go
with
longer
paths.
There's
ways
in
Windows,
where
you
can
basis
say
when
you
have
longer
path,
but
you
bypass
some
of
the
windows
path.
Normalization
which
I
know
is
on
our
end
to
fix
and
that
can
result
in
multiple
strings
designating
the
same
file.
A
So
if
you
have
a
sandbox
that
says,
certain
paths
should
be
inaccessible
to
you
app,
you
might
be
able
to
construct
strings
that
no
longer
are
caught
by
the
you
know
the
denialist
effectively.
If
there
was
another
concern
that
we
had,
which
since
then
we
basically
said
dotnet
llama,
is
the
business
of
sandboxing.
So
if
it
still,
if
it's
disallowed
than
the
US
will
tell
us,
otherwise
we
don't
care
right.
So
yeah
the
world
has
changed
and
I
think
we're.
A
B
B
A
lot
of
the
work
that
that
the
teams
are
doing
right
now
are
getting
ready
for
the
wind
forms
and
WPF
work
as
well,
so
there's
stuff
going
on
there
and
improving
our
improving
our
Stewart,
we're
working
on
improving
the
story
on
how
many
SDKs
you
have
on
your
machine
if
you
install
dot
net
core
and
so
that
that
that's
a
that's
a
really
that's
a
really
nice
big,
we
we
will
make
it.
We
will
make
a
big
step
in
preview.
I.
B
I
actually
I
actually
test
everything,
so
I
should
keep
mine
super
clean
so
that
I
know
what
the
experience
is
for
new.
When
somebody
has
a
like,
they
have
a
fairly
clean
machine
and
they
put
on
preview
three
and
so
I
keep
actually
fairly
clean.
She
might
actually
found
a
bug
that
way.
It
wasn't
a
bug.
It's
it's
it's
a
reality.
B
It's
not
a
bug
and
that's
the
global
tools
target
generally
2.1,
which
is
great
because
that's
kind
of
your
lowest
common
denominator,
but
since
our
roll
forward
policy
doesn't
roll
forward
under
three,
if
all
you
have
on
the
machine
is
preview
three
and
you
install
a
global
tool.
You
get
a
message
and
I'm
like
oh
I.
Don't
like
that
message.
That
message
could
be
so
much
nicer,
so
we're
working
on
that
whole
problem
is
all
up,
but
it's
for
me.
I
keep
a
clean
machine,
but
I
bet
rich
has
more
than
I.
Do.
A
A
B
So
in
the
change
I
can
talk
about
the
change
that
we're
making,
which
is
that
previously
we
considered
every
installation,
something
that
you
might
want
to
reach
back,
maybe
and
get,
and
we
finally
realized
come
on.
You
know
you're
not
going
to
go
back
for
patch
releases,
and
so
basically
what
we're
doing
is
cleaning
the
machine
if
you
use
an
installer
now,
if
you
put
it
on
yourself,
you
have
controls,
you
can
still
control
having
every
patch
or
a
set
of
patches
if
you
want,
but
if.
B
Just
have
normal
stuff
for
you.
We
assume
that
normal
is
actually
more
like
the
highest
patch
right
and
then
the
minor
versions,
and
so
we're
doing
that
on
both
the
SDK
and
the
runtime.
And
then
the
SDK
is
a
little
confusing
because
we
use
the
third
position,
which
other
people
consider
a
patch.
The
first
hundreds
position
of
that
is
our
minor
so
that
we
can
stay
lined
up
on
the
the
first
two
positions
with
the
runtime.
B
So
anyway,
we
will
drop
it
down
to
having
in
the
SDK
one
per
version
band
and
then,
as
we
work
on
the
runtime
I,
don't
know,
I,
don't
think
that
work
is
in
preview,
3
I'm,
not
sure
about
that,
but
we're
working
on
both
of
those
being
less
impactful
in
people's
machines
and
in
the
meantime,
there's
windows
windows,
add
remove
programs
and
I'm
I'm.
Sorry,
we
don't
have
a
better
answer.
We're
also
working
on
that.
But
that's
gonna
be
a
little
slow.
B
B
B
Out
because
we'd
like
to
do
is
be
able
to
make
that
something.
I've
got
a
global
tour.
We
just
haven't
quite
found
the
resources
yet
so
hey
somebody
wants
to
write
a
global
to
love.
It
helps
people
out
there's
a
lot
of
folks
in
this
building.
That
would
be
immensely
grateful
and
a
lot
of
other
people
that,
when
they
do
dotnet
info,
see
a
frightening
list
of
things
on
their
machines,
so
yeah.
C
B
The
problem
is
that
we're
hesitant
to
do
that,
because
if
somebody
has
a
global
JSON
that
pins
them
back
to
a
previous
version,
then
we
don't
know
that
and
we
think
they
actually
want
that
version.
So
for
the
Installer
to
just
go
back
to
what
would
be
logical
for
a
global
Jason
on
the
SDK
or
would
be
targeted
for
a
runtime.
That's
that's
a
pretty
big
hammer
and
we've
been.
We've
has
women.
B
We
have
not
decided
to
do
that
yet
so,
as
of
right
now,
we're
looking
at
a
different
way
to
clean
machines
and
then
going
forward
having
less
on
the
machine,
which
also
means
that
going
into
add/remove
programs
is
not
quite
as
painful.
If
you
just
once
a
year,
you
have
one
or
two
to
remove,
instead
of
as
many
as
people
running
a
lot
of
previews.
However,
right
now
and
it's
those
machines
that
we're
really
battering
a
little
bit
and
you're.
B
A
One
one
insertion
that
I
heard,
which
I
think
is
reasonable
is
to
say
most
people
get
on
their
core
by
a
Visual,
Studio,
I,
guess
I'm
you
and
saw
an
update
and
you
get
it
there
then
PS
itself,
basically
nukes
the
old
one,
because
if
they
added
it
to
your
box
while
channeling,
you
just
want
to
get
the
latest
version.
If
you
manually
install
it
sure,
let's
take
it
around
but
like
well.
If
it
comes
from,
you
know
studio,
it
seems
reasonable
for
us
to
upgrade
that
and.
B
That's
the
story:
we're
working
on
for
Visual
Studio
I,
get
kind
of
my
head
gets
stuck
in
the
CLI,
sometimes
so
I
was
talking
about
if
you're,
installing
the
SDK
like
you're
using
Visual
Studio
code
or
some
other
situation,
where
you're
installing
it
separately,
but
Visual
Studio
will
clean
up
after
itself,
pretty
pretty
completely
we're
going
to.
Actually
we've
got
plans
for
that.
Should
that
should
be
good
going
forward.
Speaking.
A
C
A
A
B
A
You
know
the
color
theme
like
it
looks
a
little
bit
dated
and
a
little
bit
busy,
so
there's
a
facelift
coming
for
the
blog,
which
makes
it
look
more
like
this,
where
you
know
you
have
a
cleaner
representation.
It
looks
more
similar
to
the
other
microsoft
properties.
You
have
already
seen
Microsoft
comm
the
dotnet
side,
for
example-
and
you
know
the
header
will
be
part
of
this
experience
and
then
you
have
a
consistent,
color
theme
and
then
you
know
the
actual
blog
posts
themselves
look
much
nicer.
A
They
are
much
more
focused
on
the
individual
wrote
the
blog
posts.
They
have
a
picture
for
example,
so
you
can
actually
recognize
your
famous
sorry,
your
favorite
people
quickly
against
it
and
say:
oh
this
guy,
don't
like
so
I,
don't
read
the
blog
post,
and
so
it's
kind
of
interesting
to
see
how
people
will
react
to
that
and
we
actually
do
fun,
believe
it
or
not,
actually
take
feedback.
So
there's
a
server
you
can
fill
out.
A
It
can
tell
us
what
you
think
about
it,
including
colors,
everybody's
favorite
topic
and
then
there's
also
like
the
typical
question
of
like
well:
whatever
is
the
old
URLs?
Well,
the
US
will
change.
You
have
a
simplified
URL,
but
you
also
have
redirect.
So
if
you
have,
you
know,
bookmark
your
favorite
posts
or
whatever
this
should
work,
because
we
will
just
redirect
them
and
save
your
readers.
If
you
use
some
newsreader,
there
are
lots
of
continue
to
work
and
in
the
process
we
also
clean
up
a
little
bit
like
I.
A
There's
actually
a
list
of
blocks
the
view
reconcile
because
we
had
a
large
number
of
blog
posts
here
and
so
a
block
block
owners
across
various
areas.
So
some
of
them
are
merged
and
some
of
them
are
just
staying.
Different
are
just
getting
renamed
to
make
it
easier
for
people
to
find
what
it's
up.
So,
for
example,
the
SF
seven-plus
xamarin
is
no
mobile.
A
A
C
C
B
C
B
D
E
B
D
B
C
D
D
C
C
At
the
very
bottom,
I
didn't
put
it
on
hold
er.
So
what
this
is
about
is
our
roll
forward
behavior,
and
so
we
already
had
roll
forward
for
patch
version
and
minor
version.
So
this
is
the
idea
that
you
know
if
you've
got
a
2.0
app
and
you
try
and
run
it
on
a
machine
that
only
has
2.1
that
will
work
and
over
on
2.1
without
any
configuration
and
I
should
make
this
stuff.
C
B
C
That
was
our
behavior,
that's
our
existing
behavior,
and
so
we
started
to
have
these
discussions
about
what
we
should
do
with
major
version.
Now
we
didn't
ever
discuss
or
not
at
length
major
version
row
forward
from
1
X
2
to
X,
just
because
there
were
several
aspects
about
the
product
that
were
not
amenable
to
roll
forward,
mostly
with
respect
to
the
way
asp.net
worked
well,
there
was
project
JSON.
C
C
So
anyway,
but
we've
fixed
all
the
issues
that
prevent
us
from
doing
this
in
the
past,
but
we
have
still
decided
to
not
auto
roll
forward
across
major
versions,
because
that
gives
us
into
that
excuse,
the
worse
of
the
the
use
of
this
term
compat
hell
that
we've
had
with
done
and
framework.
So
we
feel
like
giving
developers,
choice
and
and
users
choice
with
opt
in
roll
forward
for
major
versions
is
a
good
plan
right,
but
to
leave
it
as
opt
in
only
as
opposed
to
something
automatic.
So
that's,
basically
what
this
document
says.
C
I
still
think
it's
the
right
plan,
but
so
just
to
cover
that
today
we
we
don't
do
any
roll
forward
between
release
and
stable
versions
and
in
any
direction.
So
the
first
thing
that
we're
going
to
is
make
it
so
that
if
you
built
your
app
against,
you
know
the
very
last
build
of
3.0
preview
and
you
and
then
you
take
that
application
or
global
tool
and
run
them
a
machine
that
has
the
three
of
stable
version
that
will
now
work
which
wouldn't
have
worked
in
the
past.
C
C
A
That's
just
a
long
experience
from
don't
let
fame
work.
The
problem
of
the
binder
has
never
been
that
it
allows
things
it
shouldn't
allow
at.
The
bottom
has
always
been
that
it
is
a
lot
of
things
that
it
should
have
allowed.
I
think
that
if
it's
a
general
like
while
development
and
dotnet
framework
can
suck
at
times
is
that
well.
C
B
C
B
C
B
This
was
simply
it
would,
then
it
would
roll
for
and
I
absolutely
think,
there's
no
question
that
there
needs
to
be
an
opt-in
and
that
we
would
we
would
teach
people
with
certain
deployments,
because
it's
not
really
global
tools.
It's
the
fact,
you're
deploying
to
a
very
large
universe
of
people,
and
you
know
nothing
about
the
state
of
their
machines.
B
If
you're
in-house,
you
know
a
little
bit
more
about
the
state
of
the
machine,
you're
more
likely
to
want
a
phone
call
if
something
goes
wrong,
so
we
kind
of
those
two
extremes
and
making
sure
what
we
have
in
out
strategies
that
work
for
those
extremes
and
letting
the
people
in
the
middle
figure
out
do
I
want
a
phone
call
as
to
why
this
isn't
working,
because
somebody
took
two:
don't
went
off
their
machine
or
never
had
it
and
clearly
for
global
tools.
No,
we
just
do
your
best,
because
we
don't
want.
B
D
C
A
C
C
A
B
B
A
C
A
Say
that
this
sounds
very
familiar
and
no
I
think
we
could
I
mean
we
have
features
that
kind
of
do
that
right,
I
mean
so
though,
so
the
two
things
that
I
think
I
was
talking
about
quickly
is
that
the
philosophy
of
the
CLR
has
always
been
to
kind
of
abstract
away
the
hardware.
Right
so
basically
say
you
write
manage
code
and
then
the
JIT
knows
best
and
we'll
explore
the
hardware
to
the
to
the
best
capability
which
is
kind
of
what
jets
are
good
at.
A
You
don't
have
to
decide
in
your
bill
lab
or
heart
that
you're
running
on
right.
The
downside
of
that
approach
is
that
well,
JIT
is
something
happens
at
runtime,
so
you
cannot
spend
five
hours
determining
the
perfect
set
of
machine
instructions,
because
then
it
takes
forever
for
your
program
to
actually
execute,
and
so
there's
this
trait
of
and
then
what
we
did
in
the
last
five
years.
A
However,
we
don't
want
you
to
tie
yourself
the
particular
cindy
spec,
like
sse2
sse3,
a
b
x
whatever,
so
how
about
we
just
give
you
an
abstract
data
type
that
kind
of
says:
well,
here's
a
simply
register,
which
we
call
vector
of
T
and
you
don't
know
the
length
up
front,
but
it's
static
on
the
machine
because
it's
tied
to
your
processor
and
then
you
just
vectorize
your
stuff,
but
then
that's
also
not
the
best
case
in
some
cases.
So
the
next
step
we
did
was
the
hunter
intrinsic
stuff.
A
A
If,
if
you
support
sse
to
do
this
and
then
the
JIT
is
able
to
just
kill
all
the
code
that
it
knows,
aesthetically
unreachable,
because
these
properties
are
considered
constants
and
then
you
know,
there's
no,
not
much
heavy
lifting
required
for
the
JIT
and
you
have
pretty
pretty
much
100%
control
of
what
machine
instructions
you're
using
me
run
on
that
hardware
and
so
I
think
for
for
graphic
processors.
We
haven't
done
much
yet,
but
I
would
imagine
that
we
would
go
over
a
similar
route
of
that.
You
know
stages
of
control.
B
C
So
what
this
was
is
some
done
at
core
2.1.
We
added
support
for
32,
which
is
a
prerequisite
to
running
on
raspberry
PI's
and
then
in
Donna,
core
3.
Oh
we've
added
support
for
programming
all
the
pins
there
on
our
Raspberry
Pi,
which
is
awesome,
so
that's
like
GPIO
PWM,
SPI,
i2c
and
I,
probably
missed
at
least
one,
and
so
as
part
of
that,
one
of
the
things
we
we
quickly
realized
is
just
providing
an
API
for
these
pins
is
necessary,
but
not
sufficient
for
any
kind
of
developer
success.
C
What
you
need
is
each
one
of
these
little
devices,
which
are
mostly
called
peripherals,
actually
basically,
there's
a
whole
bunch
of
bit
twiddling
and
numeric
magic
that
you
have
to
do
in
order
to
work
with
these
because
they're
actually
extremely
low-level
they're.
Actually,
it's
more
or
less
the
same
level
of
low-level
nough
spean
voc,
if
not
actually
to
some
degree,
one
step
below
that.
C
So
we
created
these
things
called
we're
calling
device
bindings,
which
are
basically
a
single
class
that
allows
you
to
interact
in
just
a
very
nice
c-sharp
dotnet
way
with
these
devices,
so
that
you
can,
you
know,
get
data
off
of
a
a
you
know:
temperature
sensor,
you
can
write
to
an
LCD
panel
that
kind
of
thing,
and
so
we've
got.
You
know
a
over
a
dozen
of
these.
Now
this
this
particular
one.
C
The
brick
pi3
thing
is
awesome
because
it
enables
you
it's
a
little
device
that
you
can
put
on
to
Lego
Mindstorms
and
it
instead
of
having
the
proprietary,
Lego
Mindstorms
thingamajig
you've
got
a
Raspberry
Pi,
and
then
you
can
program
the
Lego
Mindstorms
in
c-sharp
by
the
PI.
But
you
can
see
this.
One
is
all
about
running
the
servo
motor
thing.
A
C
Yeah
I
mean
most
most
done
at
developers
are
not
used
to
writing
code
like
that,
but
I'm
sure
many
of
them
are
capable,
but
it's
just
there's
two
different
problems.
One
is
the
skill
level
required
and
then
the
other
one
is
kind
of
more
the
domain
expertise.
So
what
happens?
Is
you
end
up
reading
these
data
sheets
right?
So
it's
this
PDF.
That
basically
says
this
is
this.
Is
the
data
that
this
device
expects?
This
is
the
data
that
the
device
will
produce.
There's
like
eight
channels
on
this
particular
chip.
C
You
have
to
ask
for
data
from
each
one
of
the
channels
and
this
order
and
blah
blah
blah
so
go
just
before
we
get
off
this
topic
just
go
to
one
of
the
ones.
The
one
that
I
liked
is
I
mean
liked
all
of
them,
but
the
one
that
had
a
nice
picture.
It's
not
this
one,
it
yeah!
So
if
go
back
to
devices,
it
was
the
one
that
not
the
character,
LCD.
Well,
I
can't
tell
with
all
these
names
you
have
to
go
to
the
readme.
You
got
a
little
bit
further
down.
C
C
B
C
C
One
of
the
things
we've
been
asking
people
to
do
with
their
contributions
is
submit
these
Fritzing
diagrams.
That's
what
these
pictures
are
called,
so
it
shows
the
PI,
but
then
it
shows
exactly
how
its
wired
up
and
what
the
device
looks
like.
So
this
particular
one
is
an
eight
by
eight
matrix
module,
as
is
fairly
obvious,
and
so
it
allows
you
to
light
those
things
up.
This
particular
device
allows
I,
don't
know
if
it's
unlimited,
but
certainly
fairly
extensive
daisy-chaining.
C
So
you
can
have
you
know
eight
of
these
things,
for
example,
but
you
only
have
to
connect
up
to
the
first
one
right
and
then
I
think
I
don't
looked
at
this
particular
code,
but
the
the
this
device
binding
makes
it
so
that
you
can
talk
to
any
of
them.
Just
super
nice,
so
you
could
you
know
you
could
build
a
case
for
these
things
and
then
you
know
you
could
have
like
a
readout
of
some
kind
on
those.
So
very
nice
and.
C
C
C
Think
over
three-quarters
of
them
are
done
by
the
community.
So
when
we
built
this
repo,
this
outcome
was
absolutely
our
intention
and
it's
it's
happening
so
I
actually
built
the
first
one,
which
was
the
mCP
3008,
that's
kind
of
how
we
built
up
this
this
model
and
I'm
very
happy
with
where
this
is
headed.
This.
A
C
C
Is
that
yeah,
so
I
think
that
the
announcements
I
was
going
to
point?
You
are
they're,
actually
written
they're
being
interviewed
currently,
so
we
can't
we
can't
actually
point
to
them,
but
there's
a
couple
things
one
is
that
we're
adopting
a
different,
a
new
policy
for
alpine?
So
basically
the
idea
is,
is
we
want
to,
but
by
the
way
should
delete
the
parts
as
dropping
a
bun
two
for
three?
Oh
we're
actually
probably
not
going
to
do
that
plan.
So.
B
C
Okay,
so
with
alpine,
this
is
my
favorite
Linux
distro,
and
so
we
are
more
and
more
committing
to
it.
So
we
want
to
offer
docker
images
with
the
latest
Alpine
super
quickly
after
the
new
Alpine
comes
out,
so
as
of
this
year,
they're
in
January
alpine
3:9
shipped,
where
we
believe
will
have
support
for
that
in
probably
March,
which
is
which
is
good,
but
then
we
have
to
figure
out
what
we're
gonna
do
with
support
for
older
Alpine
versions.
C
So
our
plan-
and
this
is
what
the
document
says
is-
will
we
aim
to
support
new
Alpine
versions
within
sixty
days,
hopefully
much
sooner,
and
then
we
will
support
the
older
Alpine
version
for
a
quarter
so
with
that's
three
months
before
dropping
support
for
it
from
an
from
a
docker
image.
Standpoint.
Well,
so
still
support
it
from
like
a
dotnet
core
support
policy
standpoint,
but
we
won't
produce
images
for
it
anymore.
C
C
There
will
actually
be
a
different
URL
for
them
than
the
one
today,
oh
they're,
the
ones
that
we
use.
They
will
all
forward
to
the
new
place,
so
none
of
the
links
will
be
broken
and
we'll
be
using
NCR
more
like
a
global
CDN
for
all
of
our
docker
images.
So
you're
from
statements
will
changed
to
point
to
a
new
registry,
which
is
the
the
NCR
registry,
and
we
will
continue
pushing
updates
to
the
floating
tags
on
docker
hub.
C
So,
like
the
the
you
know,
like
22.1
sdk,
which
is
a
floating
tag,
we'll
continue
pushing
new
images
for
that
tag.
If
you
use
three-part
tags,
then
you'll
have
to
move
to
MCR
to
get
a
new
three-part
tag,
so
people
who
don't
know
anything
about
docker
now
that
probably
made
any
sense,
but
we
are
moving
to
NCR
in
the
very
short
future.
Those
are
those
are
the
two
things
and
we're.
B
E
C
Yeah,
so
the
first
question
is:
are
the
GPO
yeah
basically
can
these
can
this
work
on
other
single
board
computers
other
than
Raspberry
Pi?
The
answer
is
yes,
and
it's
yes
in
a
couple
of
different
ways.
One
is
that
actually
it
talks
about
it
on
the
homepage
of
this
repo.
You
know
so
we
support
a
few
different
drivers.
So
one
is,
we
do
have
specific
drivers
for
humming
board
and
raspberry
pi,
we're
actually
kind
of
moving
away
a
little
bit
from
writing
these
bored
specific
drivers.
We're
gonna
probably
make
Sisyphus
our
default
implementation.
C
C
So
there's
this
new
daemon
in
Linux
called
Lib
GPIO
D
our
plan
is
we
actually
have
an
implementation
already
that
we
hope
to
book
out
in
either
preview,
3
or
preview
for
of
donek
or
3,
and
it
is
tremendously
faster
than
Sisyphus,
but
is
just
as
compatible
Sisyphus
and
actually
supports
more
features
than
Sisyphus.
So
as
long
as
your
device
supports
either
Sisyphus
or
a
live
GPIO
D,
then
our
API
should
work.
C
B
I'm
not
good
for
them.
Yeah
I
know
from
a
c-sharp
perspective,
elet
Phillip
answer
this
from
an
f-sharp
perspective.
The
answer
I
mean
I.
Think
what
you're
looking
for
is
that
void
is
a
rather
imperfect
way
to
handle,
not
returning
a
value.
It
creates
the
mess
of
function,
action
being
parallel
and
just
opens
up
a
slew
of
issues,
and
the
answer
is
right:
now
we
don't
have
plans.
B
So
Rosalyn
is
where
you
would
put
an
issue
like
a
bug
or
something
specific
to
the
way
that
c-sharp
behaves
in
Visual,
Studio
or
those
types
of
things.
But
in
terms
of
you
want
to
see
something
looked
at
for
c-sharp
itself.
The
c-sharp
Lang
is
a
great
place
to
put
that.
My
guess
is,
this
is
up
there,
but
I
think
it'd
be
great
for
you
to
either
put
it
up
there
or
make
a
comment
on
an
issue.
B
If
it's
already
there,
it's
it's
a
great
way
to
communicate
to
the
whole
language
design
team
rather
than
just
just
a
couple
of
us,
and
so
that's
what
I
would
suggest
on
that.
One
I
understand
why
that
would
be
absolutely
fantastic
to
have,
and
it
nothing
may
or
may
not
be
the
right
way
to
think
about
that.
Just
because
that's
not
exactly
what
it
is
in
VB
and
it
looks
like
there
may
be
nothing,
but
a
type
that
can
handle
it
and
I
don't
know.
Does
F
sharp
have
a
we.
B
B
That's
a
teach,
that's
t-shirt
worthy,
so
absolutely
no
I
think
they'd
be
great
to
continue
that
conversation
on
c-sharp
lang,
so
that
would
be
totally
awesome.
It
sounds
like
f-sharp
sort
of
has
that
if
there's
some
aspect
that
f-sharp
doesn't
have
I
think
it
would
be
great
to
talk
about
that
on
there
sharp
line.
So
we
can
broaden
the
conversations.
B
A
So
I've
talked
about
this
a
lot
like
you
know.
We
often
talk
about
how
open
source
or
design
processes
and
how
much
of
the
stuff
is
exposed.
So
here's
an
example
of
something
that
is
that
is
totally
in
flight
right
now,
and
it's
probably
one
of
the
more
popular
API
reviews
we
have
ever
done,
I
think
in
terms
of
just
he
was
concurrently
and
so
here's
the
notes
that
are
that
I
uploaded,
which
is
you
know,
includes
both.
A
You
know
just
the
notes
we
talked
about
in
the
meeting,
but
also
links
to
the
to
the
to
the
core
FX
design
issue,
as
well
as
the
YouTube
thing.
If
you
want
to
watch
it,
our
video
of
us
renting,
the
meter
is
actually
highly
entertaining
because
it
just
mean
a
room.
Everybody
else
on
the
call,
so
it
looks
kind
of
ridiculous.
But
the
interesting
thing
is,
you
have
all
the
cool
people
on
the
call.
So
there's
a
bunch
of
really
insightful
conversations
that
are
happening
and
the
same
thing
is
when
you
go
on
YouTube.
A
If
you
can
actually
watch
the
video
on
the
side,
you
also
have
the
the
chats.
If
you
watch
the
video,
you
see
the
chat
happening
in
real
time
like
in
the
point
in
the
video,
so
you
get
a
sense
of
what
the
community
directions
are
and
like
what
what
what
we
responded
with
and
what
else
right.
So
it's
it's
pretty
in-depth.
So
if
you
care
about
JSON
civilization,
I
would
suggest
you
check
it
out.
A
It
is
something
that
we're
shooting
for
4-3
or
we
will
probably
not
have
the
Swiss
Army
knife
civilization
that
Jason
Lynette
provides.
But
our
goal
is
to
have
what
we
call
an
MVP
Minimum
Viable
Product
for
adjacent
civilizer,
so
that
we
can
eventually
say
you
know
asp
net
core
with
web
api.
You
can
use
effectively
without
Jason
dotnet
and
have
meaningful
civilization
happening
for
you
automatically
and
the
reason
we
built
this
whole
JSON
library.
A
Now
you
know
this
reader
there's
a
writer
there's
a
Dom
there's
a
civilizer
is
because
we
want
people
give
more
flexibility,
so
the
platform's
more
independent
of
Jason
on
net,
because
right
now,
they're
kind
of
tied
at
the
hip
and
the
other
part
of
it
is
we
really
want
to
go
more
after
high
throughput
applications.
We're
having
a
Swiss
Army
knife
that
Jason
net
is
not
what
you
care
about.
You
care
more
about
throughput,
and
so
we
all
knew
JSON.
A
B
B
What,
at
the
same
time,
the
new
Jason
stuff
will
allow
some
of
the
the
framework
pieces
you
depend
on
to
go
faster
and
to
have
less
dependencies
that
can
get
in
the
way
so
we're
there
version
of
Jason
net
conflicted
with
your
version
of
Jason
net.
So
there's
a
lot
of
goodness
in
this.
This
work
we
sort
of
didn't
set
that
out
by
explaining
why
we
were
doing
it.
Yes,.
D
A
A
You
want
to
go
to
the
motivation.
There
is
actually
so
much
of
the
issue
that
yeah.
If
I
could
spell
it
correctly,
it
designs.
There
is
an
acceptable
puzzle
to
to
do
the
Jason
whatever.
Maybe
it's
tough,
you
know
no,
actually,
it
might
have
been,
even
though
it's
been
actually
I'm
thinking
about
it.
It
was
an
announcement.
It
would
be
here,
yeah
yeah
yeah,
it
was
Vince
again.
Our
floral
is
a
thing
and
no
I
just
have
to
find
it.
Well,
you
can
just
type
Jason
or
yeah.
You
can
do
that
too.
A
There
we
go
so
this
one
basically
has
all
motivation,
and
it
also
has
summer
in
here
quote
from
James,
because
he
was
asked
you
know
what
do
you
think
about
that?
And
and
so
we
we
are
very
deliberate
about
that
choice.
It's
not
the
typical
Microsoft
and
I
ain't
sure
anything
like
that.
I
mean
we
have
supported
Jason
Lynette
for
a
very
long
time,
and
we
it's
not
that
we
on
unhappy
love
it
in
any
meaningful
way.
C
A
A
D
I
was
just
saying:
the
Jason
stuff
is
definitely
interesting
for
us
and
the
upside,
because
we're
starting
to
look
at
moving
our
entire
tooling
out
of
process
and
the
best
way
to
do
that.
Like
I
mean
there's
a
standard
language
server
protocol,
but
it
uses
JSON
and
it's
it's
serializing.
Data
back
and
forth
for
intellisense
is
something
that
you
want
to
keep
as
allocation
free
and
as
fast
as
possible,
and
this
is
sort
of
the
way
forward
for
that.
So
yeah.
C
A
A
Basically,
we
talked
about
the
basic
API
saved
the
converter
options,
but
then
we
scroll
down
more
there's
this
whole
attributed
programming
model
right,
like
the.
How
do
you
indeed
like?
How
do
you
mark
your
types?
What
behavior
do
you
get
right
and
we
haven't
covered
any
of
that,
so
I
think
the
next
time
like
if
I
would
have
to
guess
I
would
say,
there's
probably
two
more
rounds
of
or
reviews,
okay
and
so
we'll
just
cover
it.
And
then
the
question
is:
what
follow-ups
do
we
have
or
design
changes?
C
A
Twitter,
like
I
retweeted
from
that.
Well
like
basically,
we
meet
every
Tuesday
10:00
to
12:00,
Pacific,
Standard,
Time,
obviously,
and
we
we
try
to
reserve
the
first
hour
for
just
going
over
the
guitar
backlog
on
the
second
or
for
reviews,
unless
there's
bulky
items
like
Jason,
maybe
just
book
the
full
two
hours
for
just
this
one
thing
and
then
there's
also
you
know
one
of
the
be
set
up
a
fee
if
we
have
to
so.
For
example,
last
week
Thursday
we
met
for
talking
about
pipelines
and
so
I
tweeted
I.
A
Think
a
few
weeks
ago,
our
schedule
for
the
next
six
or
seven
reviews
I
wish
she
did
this
one
little
that
account
and
I
will
probably
keep
doing
that.
The
very
unfortunate
thing
is:
there's
no
calendar
on
YouTube,
so
I
wish
there
would
be
a
way
for
us
to
share
that.
If
you
go
to
the
YouTube
channel
for
dotnet
foundation
at
the
very
top,
you
see
like
the
next
three
or
four
things
that
happen
to
be
scheduled,
but
there's
no
calendar
view.
A
C
C
C
C
Is
it
so
the
thing
I
said
it
because
I
talked
about
this
on
the
call
this
isn't
about
migration?
It's
about
usability
right,
there's,
zero,
my
personal
feeling,
I'm,
not
the
PM,
for
this
I
feel
like
there's
zero
migration
requirement
for
this
feature,
meaning
that
that's
not
what
this
is
right.
But
there
is
a
usability
report.
The.
B
E
E
C
Out
which
is
I
think
they
can
also
be
incredibly
different,
but
but
if
one
is
like
10x,
more
and
10x
isn't
even
the
right
number
if
one
of
them
is
if
Jason
net
is
significantly
more
usable
and
intuitive
right
than
this
one.
That
I
think
we
have
to
discuss
that,
but
I
don't
think
the
patterns
have
to
be
no.
A
Think
the
thing
that
matters-
the
most
is:
what
kind
of
types
can
you
see
relies
on?
How
do
these
types
have
to
look
in
order
to
be
sterilized?
It
was
so,
for
example,
James
has
done
a
great
job
of
pretty
much
handling
any
time
you
throw
I
did
write
if
the
time
is
immutable.
He
handles
that
if
there's
a
constructor
with
mention
parameters,
I
think
he
mentions
that
and
like
the
problem
is
that
each
of
these
things
are
usually
like
in
create.
A
You
know,
increasing
usability,
but
they
also
make
the
designer
still
that
much
harder,
and
so,
if
you
want
to
off
too
much
for
throughput,
then
it
becomes
this
okay.
We
can
only
support
that
many
scenarios,
which
ones
are
we
picking
to
support
yeah
and
in
you
know,
which
one
most
of
them
do.
We
think
are
actually
impactful
versus
just
nice
to
have.
A
Yeah
I
mean
there's
this
plenty
of
areas
and
like
the
nice
thing
right
now
is
like
it's
all
in
flight
and
if
you
follow
the
API
reviews
repo,
for
example,
or
the
issues
on
core
FX,
there's
plenty
of
highly
skilled
individuals.
Commenting
on
this,
you
know
both
from
the
community
as
well
as
for
Microsoft
right.
There's.
People
trying
to
use,
for
example,
the
JSON
work
in
MVC,
the
guy
that
writes
most
of
the
ABC
implementation,
is
commenting
on
this
Ryan.