►
From YouTube: ASP.NET Community Standup - April 24, 2018 - Super Early ASP.NET Core 2.2 Feature Planning
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
C
D
B
D
A
D
A
D
A
D
A
A
It
for
me,
it's
new
to
me,
so
they
have
a
really
cool
tech
community.
It's
like
it
was
fun
going
to
like
people
just
seemed
super
excited
at
the
conference
like
a
little
bit
more
than
conferences,
I
bet
like
when
you
go
to
a
conference
in
the
in
conferences.
I
been
to
people
are
like
yeah,
it's
cool
conference.
This
one
people
are
like
teach
me
things,
I
want
to
learn,
and
this
is
awesome.
It
was
cool,
that's
a
chance.
A
It
was
there,
wasn't
a
bit
like
happy
and
it's
not
like
intense
intense,
should
community
spotlight
thing
should
I.
Do
that.
Oh
I.
B
Little
wax
lyrical,
I,
don't
know.
No
I
was
I
tweeted.
Today
someone
had
an
idea
and
then
I
just
blatantly
stole
it.
I
had
some
packages
stolen
off
my
poet
recently
we
had.
We
live
in
a
circle
thing
with
like
ninety
houses,
and
we
had
some
aspiring
young
teenagers
come
and
steal
a
bunch
of
stuff
off
everyone's
porches
in
the
middle
of
the
week.
A
few
weeks
ago
and.
D
B
Were
following
the
delivery
truck
around
the
circus
around
the
circle
and
like
and
then
stealing
the
packages,
and
so
everyone
got
them
on
their
doorbell
cameras
and
God
knows
what
else
security
cameras
everyone
has
where
I
live.
So
that
was
funny.
He
said.
I
went
to
the
police,
I
think
the
most
interpret
I
think
the
most
valuable
thing
they
got
was
a
six-pack
of
infant
formula
from
one
of
our
neighbors
I.
Think
they
stole
I.
B
Think
I
think
we
had
some
cat
wipes
than
making
the
cat's
hair
look
pretty
nice,
so
they
would
have
enjoyed
that
anyway.
The
point
was
I
was
investigating
like
what
could
we
do
to
improve
the
physical
security
of
this
situation,
because
this
is
becoming
a
very
frequent
occurrence
and
no
matter
how
many,
how
much
footage
you
send
to
the
police
there's
only
so
much
they
can
do
and
I
tweeted
JJ
tweeted
fbiís
us
today
and
I
said.
B
Anchor
right,
and
so
you
would
have
a
steel
box,
you
have
four
holes
in
the
bottom.
You
bolt
that
into
your
concrete
porch,
so
it
can't
be
removed
and
then
you
would
be
an
Amazon
locker
right.
So
maybe
it's
a
part
of
your
Prime
subscription
or
some
extra
amount,
and
then
you
could
just
put
stuff
in
it.
So
you
can
buy
like
delivery,
lockers
for
your
home,
they're,
pretty
unwieldly
and
they
kind
of
look
like
a
post-office
box.
Right.
You
put
the
pass
on
the
top.
D
B
Take
that
idea
from
you
and
steal
it
I,
don't
care
because
ultimately,
I
will
be
the
one
who
benefits,
because
I
will
get
an
Amazon
Locker
on
my
friend
push
let's
say:
I
thought
that
would
be
really
cool,
but
you
know
we'll
see
what
happens
if
anyone
else
has
a
better
idea.
I
was
staring
the
front
of
my
house
going
well.
D
We
do
one
idea
which
already
exists
is
an
Amazon
locker
at.
B
D
A
D
B
D
B
Saying
I
don't
want
that
I,
don't
want
them,
opening
my
front
door
to
put
parcels
and
I
don't
want
them.
My
car
may
not
even
be
there
so
I
just
want
a
locker,
they
can
stick
it
in
and
then
I
can
open
it.
I
might.
D
B
B
D
B
B
So
I
can
imagine
they're
like
an
Amazon
locker,
that's
connected
to
your
Wi-Fi
like
so
you
get
alerts
when
it's
been
opened
and
like
and
stuffs
been
delivered
like
it's
be
like
a
smart
Locker
that
Amazon
controls
and
I
pay.
Some
part
and
portion
of
my
Amazon
Prime
Tours
I
would
totally
use
that
I
would
totally
totally
use
that
I
get
a
ridiculous
amount
of
stuff
delivered
to
my
house.
Sorry,
I!
Don't
want
to
go
out
in
the
world.
I
dig
around
the
world
at
lunch
today,
you'll
see
behind
me
here.
I.
A
B
I
went
for
a
walk
at
lunchtime
to
find
the
mythical
treehouse
this
thing
here,
there's
actually
three
of
them
there
in
like
a
this
little.
This
is
a
meeting
room.
You
can
check
out
it.
Microsomia
I
have
never
been
at
a
campus,
it's
actually
behind
I,
don't
know
if
you've
been
there
Scott
it's
behind
the
executive
briefing
Center
behind
the
conference
center
and
the.
A
B
D
B
Building
and
people
don't
like
you
to
know
it's
behind,
so
it's
actually
yeah.
So
it's
the
set
of
places
behind
that
in
Trier
I've
never
been
down
there.
The
cafe
was
very
nice:
I
wouldn't
had
lunch
there
and
then
I
walked
around
then
I
took
the
path
around
the
outside
of
the
campus
you're
starting
here
and
came
all
the
way
back
round.
It
was
very
pleasant.
It's
very
pleasant
so
anyway,
I'm
in
a
good
mood,
because
I
got
outside
the.
D
B
D
D
B
A
B
A
You
don't
want
to
want
to
optimize
that
so
your
builds
are
fast,
so
he
had
done
something
previously
that
I'd
spotlighted
using
tar
tarring
as
part
of
it,
and
then
somebody
said
here's
a
way
where,
if
you
structure
your
structure,
things
correctly
like
using
a
convention,
then
you
can
so
he
went
through
here.
So
you
can
write
something
where
you're
manually
writing
all
these
copy
statements.
But
that's
not
that's
not
super
nice.
A
Then
he
talked
about
the
previous
solution,
Aidid
using
tarballs,
and
then
his
new
solution
is
like
going
through
in
scripting
it
and
doing
it
using
kind
of
a
convention
based
approach.
So
this
is
based
on
a
comment
that
someone
else
had
previously
put
in
if
this
is
one
of
those
things
where,
if
you
are
doing
docker
based
releases,
this
sounds
like
something
you
should
read
in
detail
and
love.
A
So
this
the
benefit
is
that
docker
uses
layers
right.
So
when
you
do
docker
whatever
docker,
you
create
a
new
docker
image,
it
uses
kind
of
the
base
and
then
it
layers
things
on
top
of
it.
So
the
idea
is
that
your
sequential
builds
don't
have
to
like,
don't
have
to
restore
all
the
new
get
packages
and
all
that
kind
of
stuff
right,
you're,
not
you're.
It's
it's
because.
D
Though,
is
that
different
than
the
existing
solutions,
we
have
for
just
doing
a
multi-layer?
A
multi-step
build
like
you,
just
have
to
put
one
additional
step
between
the
copies,
and
then
it
makes
a
layer
and
that
layer
doesn't
get
invalidated.
You
do
the
restore,
and
then
you
copy
from
there
right,
yeah.
B
Loc
one
youjin:
yes,.
A
I
do
I
said
in
advance.
I
was
going
to
forget
today
to
do
that
and
I
blame
the
jet
lag
so
anyways.
So
this
is.
This
is
something
where
it
sounds
like
I,
honestly,
I
haven't
done
a
bunch
of
docker
base
like
you
know
it
solutions
like
this,
but
it
sounds
like
it
could
be
useful
for
people
so
there
it
is
moving
on
Joshi
talking
about
seven
things
worth
knowing
about
asp
net
core
logging.
So
just
going
through
and
running
down
things
you
know
so
there
it
is
talking
about
inbuilt
loggers,
etc.
A
D
D
D
B
Is
all
a
hang
on?
This
is
all
new
into
one.
So
HTTP
client
factory
is
the
new
thing.
Then
you
say
I
see
added
into
one
and
then
the
poly
extension
to
that
we
own
like
we
build
the
poly
extension.
It
doesn't
ship
it
in
the
box,
it's
a
serrated
edge
yeah,
but
you
add
that
and
then
you
can
easily
apply
poly
policies
to
HTTP
client
and
via
the
HP
client
factory.
So.
A
D
The
idea
here
is
I
scroll
down
just
a
smidge,
so
I
can
see
the
whole
simple
cast.
Client.
The
idea
here
is
that
I
want
to
be
able
to
have
an
HTTP
client.
I
can
go
and
have
a
simplest
amount
of
code
as
possible.
So
like
basically,
three
lines
there
and
on
get
shows
right,
go
and
get
it
and
read
it
out
as
a
list
of
shows
and
that's
like
the
minimum
amount
of
work
possible.
But
then
I
also
want
all
this
policy
stuff.
What
I
think
they
used
to
call
cross-cutting
concerns
right
like
well.
D
This
is
a
little
bit
of
a
flaky
API
I
have
found
it
to
be
a
little
flaky,
so
I'd
like
to
do
it
like
are
one
or
two
retries
and
then,
if
it
doesn't
retry
like
well,
what's
what's
a
retry
really
mean?
Does
that
mean
like
a
500
retry,
or
does
that
like
retried
twice
so
I
need
to
retry
and
then
back
off
then
retry
in
five
seconds,
like?
What
exactly?
Does
that
mean
all
of
that
policy?
I
could
write,
but
then
my
three
lines
of
code
would
be
40,
50,
60
and
well.
I.
B
Absolutely
I
think
this
is
gonna,
become
just
a
new
sort
of
default
idiom
that
you
see
in
your
a
subnet
core
applications,
because
if
you're
calling
out
to
anything
by
HTTP,
then
generally
you've
got
to
expect
that
sometimes
it's
gonna
fire
up
in
some
way.
So
yeah
just
like
databases,
it's
really
no
different
generally,
especially
if
you're,
using
something
like
sequel,
Azure
or
something
hosted.
It's
pretty
normal
over
the
course
of
you
know
any
type
of
traffic
that
you'll
get
some
transient
issues,
that's
just
the
reality
yeah
their.
D
D
B
Funny
we
we
did
some
customers
studies
recently
yeah.
It
seemed.
B
Yeah
we
did
some
customer
studies
just
on
that
point
recently.
You
know
Scott
I
think
you
might
have
you've
heard
some
of
this
we're
looking
at
doing
some
redesign
of
various.
You
know:
public
websites,
the
asp
net
website
and
whatnot
furniture,
and
so
we're
asking
like
developers
what
they
find
most
appealing
and
and
most.
A
B
Like
you
know,
as
you
add
features,
you
add
configuration
options,
you
had
API
those
things
have
to
be
able
used
together
and
there
are
patterns
that
have
to
be
learnt
and
they
were
like
those
those
things
don't
matter
as
long
as
there
are
good
docs
right
yeah.
We
expect
that
we
should
we're.
Gonna
have
to
read
some
docs
in
order
to
configure,
you
know,
feature
X
and
why
they're
working
other
than
a
way
I
want
them.
But
if
those
Doc's
don't
exist,
it
just
becomes
a
brick
wall,
and
so
that
was
kind
of
eye-opening.
B
D
B
D
Totally
agree
and
I
would
say
that
I'm
more
likely
to
use
something
and
I'm
more
likely
to
write
about
it.
If
you
know,
there's
great
talks-
and
this
case
here
it
was
extremely
easy
to
get
started
and
I
I
was
refactoring
via
subtraction
and
yeah.
That's
my!
He
like
I
love
that,
like
that's
the
magic
of
open
source
and
a
good
community,
is
hey
this
code,
I
wrote
I,
don't
need
any
of
that
and
you
remember
all
that
work
I
put
into
caching
for
my
my
podcast
site,
yeah.
B
D
And
there's
no
there's
half
dozen
different
ways
to
do
it.
I'll
go
and
I'll
show
how
I
wrote
the
code
that
has
multiple
double
checks
and
semaphores
and
semaphore
slam
and
all
that
and
then
I'll
go.
You
know,
that's
that's
it's
in
a
it's
in
a
freaking
class
called
show
database
which
it
doesn't
even
really
belong
there,
yeah
and
I.
You
know
I've
been
talking
to
David
about
what
the
best
way
to
yeah.
B
B
B
Every
request
so
I'm
using
I
memory
cache
because
I'm
not
using
HP
client
for
the
things
I'm
using
in
some
case,
but
that
makes
total
sense
if
you
can
do
it
in
a
cross-cutting
way,
and
you
can
apply
it
to
that
endpoint,
because
HP
client
factory
gives
you
typed
clients
right
that
you
can
tie
to
a
specific
host
or
whatever
it
is.
You
want
to
tie
it
to.
B
Then
you
got
a
really
nice
place
to
put
it
mmm-hmm,
but
to
be
fair,
also,
it's
worth
noting
that
there
are
lots
of
ways
to
cache,
and
sometimes
you
need
more
than
one
level
of
cache
and
all
those
things
are
fine.
We're
not
saying
you
should
do
all
your
caching
in
this
layer.
It's
just
another
thing
you
can
do
if
you
want
to.
D
D
A
A
B
Debbie
cleared
this
stuff
as
always
being
possible
with
HTTP
client.
This
was
part
of
the
original,
h-2b,
client
and
web
api
sort
of
mirrored
api
with
this
delegate
handler
thing,
but
having
a
central
place
to
configure
these
things
didn't
exist
right,
you
you
paste,
they
have
to
have
either
you'd
have
to
wrap
an
HP
client
yourself
or
you
know,
I'll
do
something
else,
but
now
we
have
a
single
place
to
put
it
in
a
snake
core
application,
which
is
nice
cool.
D
A
B
A
So
this
is
one
where
Dave
has
kind
of
a
similar
sort
of
thing,
but
he's
looking
at
how
things
are
specifically
wired
up
so
I
like
his
diagrams
in
here
we're
showing
kind
of
you
know
how
things
kind
of
call
in
to
each
other,
so
that's
nice
and
showing
you
know
like
how
mono
works
as
part
of
it
etc.
So,
to
say,
this
is
a
good
read
and
it's
there's
a
lot
of
things
that
I
kind
of
when
I
first
was
thinking
like.
A
Oh,
it's
all
webassembly,
it's
all
cross-compiling
or
whatever
it's
it's
important
to
kind
of
read
through
and
understand
what
what
is
actually
happening
so
so.
This
is
a
good
place
to
start
with
that.
Here's
the
release
post,
so
this
is
blazer
o2o
release
available.
So
so
this
is
pretty
neat.
This
is
target.
This
will
give
you
the
information
about
how
to
set
it
up,
and
you
know
it
started
building
it
as.
A
B
A
B
A
D
A
Nice
so
and
then
Lucan
teams
set
up
so
there's
a
lot
of
Doc's
in
here
also
a
rainier
who
had
done.
I
think
his
name's
rainier
strohbeck
saying
his
name
wrong
when
I'm
close,
he
did
the
learn,
blazer
site,
and
so
he
contributed
to
this
as
well.
So
this
is
kind
of
the
central.
You
know
blazer
Doc's
at
this
point,
so
pretty.
A
A
B
Brad
asks
why
are
they
on
a
separate
site
is
because
it's
not
officially
supported?
Yes,
that
is
why
putting
them
on
dock
stop.
Microsoft
comm
would
just
confuse
people,
it
is
a
project
not
a
product,
and
thus
it
is
on
its
own.
It
is
an
open-source
project
project
that
the
a
genetic
or
team
owns
and
that's
as
far
as
it
goes
right
now,.
A
So
this
is
this
is
from
tour
and
he's
showing.
So
the
idea
here
is
Redux
States
tour,
so
this
is
this
is
something
you'll
see
in
in
reacts,
there's
a
kind
of
a
Redux
styles
for
storing
states.
So
when
you're,
storing
state
on
the
client
and
and
your
flows
on
the
client,
how?
Because
you're
not
going
you're
not
doing
request
response
to
the
server
you're
stayed
on
the
client,
so
here
he's
implemented
a
Redux
style
state
management.
This
is
neat
to
hear
he's
showing
the
actual
flow.
A
You
can
kind
of
step
through
the
the
state
and
track
the
state
here
as
well,
which
I
would
use
it
just
for
that.
That
looks
crazy,
so
so
neat
this.
This
is
interesting.
I'm
gonna
be
watching
for
how
people
are
looking
into
building
these
kind
of
deeper
patterns,
I
mean.
Obviously
there
is
a
component
model
that
that
shipped-
and
that
has
that
as
part
of
the
templates,
but
then
people
build
sometimes
more
kind
of
complex
styles
of
how
you
manage
your
state.
So
that's
gonna
be
interesting
to
watch
how
that's
done
you.
A
Here's
something
I,
don't
I,
don't
know
if
they
mean.
If
you
want
to
comment
on
this
at
all
there's
this
kind
of
there
were
two
posts
about
library
manager.
So
this
is
kind
of
the.
This
is
the
right.
Click
manage
client-side
libraries
and
the
idea
here
is,
if
you
want
to
use
so
I'll,
try
and
say
it
the
right
way,
and
you
correct
me
if
I'm
wrong
here,
but
oh.
C
A
D
One
that
I
this
is
judge
this
made
me
think
about
when
I
was
working
in
big
big
corporate
world
and
everything
you
installed
on
your
machine
had
to
get
yet
to
get
permission
for,
and
you
just
want
to
go
and
get
something
from
a
CDN
and
bring
it
down
and
have
a
kind
of
a
package.json
for.
But
you
can't
put
a
note
on
that
machine,
or
you
don't
want
to
know
depends
see
for
NetApp.
D
B
Modi
khola
pac-man
yeah.
This
is
the
product
ization
of
pac-man
and
potentially
anyway,
and
yet
to
Scott's
point.
This
was
literally
just
about
what
does
it
mean
to
be
able
to
bring
in
heads
pac-man
there's
a
lot
of
images
there?
What
does
it
mean
to
be
able
to
just
bring
in
a
basic
client-side
libraries?
Not
frameworks
like
this
isn't
about
go.
D
B
D
A
B
You
know
our
web
forms
templates
in
our
static.
You
have
server-side
HTML
templates.
Basically,
we
only
need
sprinkles
of
client-side
files
to
get
those
functionality.
You
need
bootstrap,
you
need
a
crew,
you
need
jQuery
validate,
you
need
jQuery
validation
and
obtrusive,
and
that's
kind
of
it
and
you
might,
you
might
want
to
bring
in
a
couple
of
other
libraries,
typically
without
modules
or
dependencies
or
any
of
those
things
in
moment.
B
B
A
B
Will
get
the
stuff
into
a
node
modules
folder
and
then
it's
up
to
you
to
decide
which
part
of
the
module
you
actually
wanted
to
copy
into.
Whatever
part
of
your
app
Bower
was
the
same
web
pack,
you
know
relies
on
NPM.
So
it's
a
similar
thing
in
this
in
Lib
man
or
Pac.
Man
is
nothing
like
that.
It's
not
a
registry!
You
don't
have
to
upload
your
package
there.
B
It's
literally
just
a
tool
that
runs
in
your
project,
either
a
build
time
or
you
invoke
it
from
Visual
Studio
or
there
is
a
CLI
that
will
that
will
come
eventually,
hopefully,
and
you
who
can
you
you
tell
it
where
you
want
to
get
the
files
from
so
you
can
see
up
there.
It's
just
default
provider,
CDN
j/s,
it's
pluggable,
so
you
can
say
I
want
the
files
to
come
from
this
provider
called
CDN
GS,
which
is
you
can
imagine
just
downloads
them
from
that
website.
B
B
Numbers
aren't
turned
on
on
these
screenshots
that
that
you
only
want
you
know
those
files
and
so
yeah.
It's
really
just
a
it's
just
a
client
and
so
at
all.
In
order
to
make
it
simple
to
get
things,
they
fight,
client
side,
static
files
and
then
pop
pop
them
in
your
app
it
doesn't
do
dependency.
Is
it
doesn't?
There's
no
registry
to
upload
stuff
to
is
nothing
to
register.
It's
not
even
it's
even
less
than
Bower
was
because
it's
different
to
what
even
what
bow
was
bower,
usually
github.
A
B
Mean
yeah
I
mean
fullback,
obviously,
is
generally
a
good
pattern
if
you're
using
a
CDN,
but
even
more
than
that,
you
just
need
the
files
for
local
development,
stuff,
yeah
I
want
to
be
in
yet
other
users,
the
files
to
get
intellisense
and
there's
a
whole
bunch
of
reasons
why
you
just
want
the
files
in
your
app,
and
you
want
the
declaration
in
your
project
about
what
files
you
need.
So
typically,
this
workflow
here
is,
you
know
you
get
the
files
the
tool
brings
them
in,
and
then
you
check
them
in
right,
like
wherever.
B
Up
in
your
project
like,
in
this
case,
they're
going
into
double
double
root,
/lib
we're
just
kind
of
the
the
pattern
that
we've
established
for
a
genetic
or
you
commit
those
they
go
in
your
repo
like
this
is
a
part
of
your
build
system.
As
such,
you
can
choose
to
make
it
that
way
if
you
really
want
to,
and
as
you
mentioned
like
there's
nothing
about
the
existing
support
for
alternate
strategies
or
more
complex
strategies
or
feature-rich
strategies
for
this
is
going
away.
You
can
you
know
the
task
owner.
B
Explorer
is
still
in
visual
studio
with
gulp
and
grunt
support
the
by
our
templates
ship
with
webpack
and
that's
not
going
to
change
and
they
depend
on
node.
Node
still
is
part
of
the
vs
install
for
lots
of
reasons.
There's
lots
of
parts
of
vs
that
use
various
parts
of
node,
including
the
node
tooling.
B
D
D
B
B
A
B
B
B
A
All
right,
okay,
so
John
Hilton
did
a
MVC
versus
razor
pages,
a
quick
comparison.
Now
this
is
not
a
I
like
this
one,
better
or
whatever.
This
is
more
I
read
this
as
if
you
are
familiar
with
MVC,
here's
where
things
are
slightly
different
and
where
they're
the
same
with
razor
pages
etc.
So
this
is
kind
of
more
of
just
a
functional
here's.
How
a
request
comes
in
here's,
how
it
served.
A
B
C
A
B
B
B
Read
any
of
the
things
about
it,
you
would
realize
that
it
literally
has
nothing
to
do
with
any
of
those
things,
but
anyway
like
yeah,
though
it
is
old,
it
is,
and
it
was
interesting,
reading
people's
comments
that
you
know
that
I
haven't
seen
it
yet,
but
I'd
love
to
hear
more
I
mean
all
the
feedback
we've
had
from
people
who
use
it
has
been
overwhelmingly
than
the
majority
of
has
been
overwhelmingly
positive.
Mm-Hmm
I
know,
Scott
loves
it
yep.
A
I'm,
a
fan,
I
love
that
you
can
mix,
mix
and
match
and
I
died
same
thing.
Every
time
I've
presented
on
razor
pages
and
people
get
it,
you
know,
then
they
come
up
and
go.
Oh,
this
makes
sense
like
and
people
that
are
advanced
say
like
I,
see
how
this
is
a
lightweight
model.
That
I
can.
This
is
a
simpler
direct
mapping
for
serving
a
page
yeah.
C
A
Sense
for
a
lot
of
cases
and
I
have
had
people
that
said
I've
done
web
forms
to
have
MVC
scared.
Me
I
never
really
wrap
my
head
around
it,
and
this
thing
seems
like
this.
Can't
sig
I
need
a
controller
and
I
need
these
files
here
and
if
I
name
the
file
wrong,
something
doesn't
work
and
I,
don't
know
what
I
mean
it's.
This
is
just
a
little
bit
less
concept.
I.
D
D
Almost
like
a
restful
in
the
sense
of
like
when
we
all
figured
out
rest
as
a
community,
it
was
like,
oh
okay,
yeah,
rest
cool.
You
know
like
suddenly
WS
star,
dot,
star
and
those
things,
although
they
have
their
place.
Just
like
oh
yeah,
okay,
HTTP
for
a
couple
of
verbs
make
sense
cool
like
it.
It
suddenly
everything
melted
away.
Once
we
as
a
community
accepted
rest,
things
got
cool
in
the
API
space
and
I
feel,
like
razor
pages,
unapologetically
acknowledges
the
the
HTTP
of
it
all
yeah
and.
D
A
A
It's
a
new
hotness,
so
this
is
as
your
sample
code
samples,
Frenette
starting
with
a
basic
website
and
then
some
several
others
building
up
to
exploring
cosmos
DB.
So
these
are
like
these
are
official
as
your
code,
samples
for
dotnet,
so
neat
and
you
know
like
you-
can
go
through
and
check
them
out.
Yeah.
D
C
A
I
do
like
that
the
docs
they're,
not
just
like
MSDN
style
Docs,
where
it's
like
here's
the
reference.
It's
know,
here's
a
walkthrough,
here's
I
can
download
the
code
sample
here.
You
know
what
I
mean
it's
it's
it's.
This
is
like
everything
this
is
thought
through,
as
if
you
want
to
learn
this
thing,
let
me
guide
you
through
it.
So
yeah
there
you
go
so
that
is
the
well.
We
share
all
these
things
out
or
I.
A
A
D
B
B
C
A
B
D
B
B
Ready
I
have
no
idea;
this
is
gonna.
Do
now.
It's
just
hardly
not
gonna
work
that
tirely
did
nothing.
It
kind
of
tried
to
do
it
actually
right.
I'm
gonna,
try
one
more
I'm
gonna,
try
one
more
add
sauce
game
capture
devices
text,
webpage,
media
file,
streams,
widgets,
whiteboard
skype,
video
screen
capture,
just
like
gives
me
a
selection
thing
and
then
I
can
like
create
a
rectangle,
and
then
it
starts
capturing
that
as
a
stream,
but
it's
on
the
same
monitor
that
the
actual
tool
is
on.
B
So
it's
not
particularly
useful
because
I
have
to
like
capture
the
same
thing,
which
is
a
little
stupid.
Alright
I
gave
up
alright.
So
let's
go
back.
Let
me
just
share
my
screen.
Let
me
do
that.
Let's
go
here.
I
will
share
this.
No
one
noticed
any
of
that.
It's
all
fine!
Alright,
okay,
you
see
this
I.
Don't.
A
B
B
I
could
promise
that,
but
you
know
it'll
just
we'll
just
say
this
is
the
working
Tyler
forward
to
asp.net
core
4.7
working,
tired
I
can't
see
the
chat
anymore,
so
you'll
have
to
relay
the
chat
to
me.
If
there's
anything
useful
I
deleted
the
timeline
other
than
to
leave
it
as
2018,
so
we'd
want
to
ship
this
before
the
end
of
this
year
we
haven't
even
ship
2.1
yet,
but
2.1
is
is,
is
getting
close
like
with
the
RC,
is
shutting
down
right
now.
We're
in
like
super
super
super.
B
Thank
you
very
much,
particularly.
The
team
is
working
very
hard
to
get
it
out,
so
release
candidate
will
be
coming
shortly
and
then
hopefully,
there'll
be
no
issues
in
the
release
candidate
and
we
can
very
quickly
do
an
RT
em
after
that.
So
later
this
year,
sometime
much
later
this
year,
there's
a
couple
of
things
I'd
like
to
look
at
so
we've
had
some
feedback
around
a
template,
so
I
want
to
try
and
address.
B
Obviously,
the
bootstrap
moved
to
bootstrap
for
point
X,
which
isn't
just
templates
we
have
to
like
it's
actually
a
lot
more
complicated
than
top-paying
templates.
We
have
to
update
all
of
our
scaffolding
because
when
you
scaffold
stuff
into
your
project,
it
has
some
like
it
includes
like
bootstrap
classes
and
UI
and
crap
like
that.
B
So
we
have
to
be
able
to
know
that
oh
you're,
using
bootstrap
three,
let
me
put
in
the
bootstrap
three
basis,
scaffolds
versus
the
bootstrap
four
by
scaffolds,
etc,
etc,
live
man
we
talked
about,
so
you
know,
depending
on
what
way
that
goes.
We
would
obviously
update
the
templates
to
do
that.
This
one
is
a
little
more
large,
so
we
have
this
four
at
this
tension.
That's
been
on
forever,
like
pants
when
you've
been
here
for
long
enough
to
know
this
is
that
we
always
question
what
the
templates
are
for.
B
The
templates
are
getting
started
point
or
are
they
a
sample
app
and
so
we're
going
around
the
merry-go-round
on
this
again
and
we're
toying
around
with
the
idea
of
changing
our
web
app
template,
which
is
kind
of
the
full-blown
one
that
has
like
the
cookie
consent
and
all
the
other
new
stuff
in
in
2.1,
and
changing
that
to
be
a
sample
web
app
rather
than
the
starting
point.
That's
kind
of
the
default
starting
point
and
we
would
rather
have
the
web
app
be
more
like
the
old
empty,
empty
MVC
app.
B
So
we
used
to
have
this
back
in
the
MV
like
MVC,
4
and
5
days,
so
it'd
be
like
truly
empty,
which
was
like
literally
no
files
like
a
web,
the
config
file,
and
then
you
would
have
an
empty
MVC,
which
was
just
enough
to
get
the
MVC
stuff
in
there
and
a
control
of
that
like
said
hello.
So
there's
some
discussion
about
whether
we
change
the
default
experience
back
to
something
like
that.
B
B
But
if
you
do
want
the
fully
blown
thing,
maybe
you
can
get,
you
can
get
more
fully
functional
templates
from
like
the
vs
gallery
or
from,
and
you
get
installations
for
the.net
CLI,
and
so
we
had
these
back
nice
penis
or
two
days
we
have
the
starter
pack
templates.
I
know
if
you
remember
this
yeah,
like
a
show.
A
B
You
had
a
bunch
of
other
stuff,
so
we're
talking
about
whether
that
might
be
something
that's
worthwhile
doing
again.
I
think
this
might
be.
This
could
potentially
help
some
community
efforts
as
well.
We've
discussed
whether
it
would
be
you
know.
What
could
we
do
to
improve
the
visibility?
Also,
let
me
show
you
something
right
now.
Actually
so
I'm
gonna
open
up
the
s
in
the
one
monitor
I
can
open
up
EF
2017.
B
You
can
see
this
yep
good,
so
if
I
go
file
new
today,
folks
may
not
realize
because
I
overlooked
this,
like
four
four
versions
of
a
yes,
so
it
defaults
to
here
right,
so
you
I
can
create
new
application.
But
if
you
look
over
the
left
hand
side
here,
I'm
actually
inside
this
node
called
installed,
which
is
this
comes
in
vs,
but
then
there's
this
other
node
called
online
and
if
I
click
online
I'm
dropped
into
this
whole
other
universe,
row
of
community
templates.
B
B
Just
some
thoughts,
just
some
thoughts
and
that
might
help
some
of
the
people.
I
guess
says:
I'm
the
community
efforts
out
there
who
want
to
add,
like
alternate
templates
and
they
want
to
get
them
individual
studio
and
make
them
little
more
prominent.
If
we
could
kind
of
do
the
bootstrapping
work
so
that
they
are
prominently
shown
here
as
in
like
the
link
over
to
them
is
probably
show
and
then
maybe
that'll
help
folks
know
get
people
directed
to
the
templates
that
they're
publishing
to
the
gallery.
So
anyway,
that's
an
idea.
Yeah.
B
So
like
that
yeah
exactly
so
the
CLI,
that's
obviously
a
little
more
challenging,
you
would
probably
have
to
create
a
command
in
the
CLI
to
like.
Maybe
they
could
be
an
entry
when
you
type
dotnet
you
right
so
like
today.
If
you
type
O,
look
there's
what
I
did
earlier.
This
is
me
having
typed
net
new
write
H,
and
there
was
an
entry
right
here,
a
thing
here.
That
said,
you
know
type
net
new
online.
A
B
Something
and
then
it
gave
you
a
some
I,
don't
know
like
it
would
be
difficult.
There's
no
mean
because
you
really
need
an
interactive
experience.
You
can't
just
like
dump
the
entire
template
repository
you
kind
of
want
to
ask
them
like
what
are
you
interested
in
what
keywords
you're
looking
for
or
something
like
that
on
there,
but
that
could
certainly
be
something
that
could
be
investigated.
B
Okay,
so
in
the
web
applications
space.
So
if
I
go
back
up
the
the
major
investment
areas
that
we're
looking
at
or
around
api's
and
cloud
services
and
the
by
extension,
sort
of
running
those
in
a
micro-service
e
type
fashion.
So
what
are
the
type
of
concerns
that
come
into
application
development
that
you
want
to
be
able
to
solve
when
you
run
when
you
write
services
applications
and
when
you
run
them
in
a
small
e,
factored
maybe
container
deployed,
you
know
sort
of
pieces.
B
We
want
to
continue
our
heavy
investments
in
our
service,
so
kestrel
and
the
way
the
kestrel
integrates
with
your
casual
itself
and
the
way
it
integrates
with
the
various
front-end
proxies
the
way
integrates
with
the
various
environments
like
in
asia
and
then
performance,
which
is
always
something
that
we
want
to
continue
investing
in
I'm.
Sorry,
so
I
went
through
templates
in
the
web
app
space.
B
You
can
see
it's
kind
of
minimal
here,
there's
a
couple
things
that
we
want
to
light
up
because
it,
despite
our
major
focus
being
in
the
next
section,
we
want
to
revisit,
bundling
minification.
This
is
something
that
we've
kind
of
ignored,
ignored.
We
had.
We
tried
some
alternate
strategies
in
core
similar
to
the
client
side
manager
stuff,
and
this
hasn't
particularly
worked
out.
B
Well,
we
get
an
awful
lot
of
feedback
from
folks
who
just
want
something
as
simple
as
what
they
had
back
in
the
days
of
MVC
5
with
the
script
bundler
stuff,
where
you
would
configure
the
bundles
in
c-sharp
it
would.
It
would
bundle
at
runtime
when
you,
when
the
request
came
in
I'm
in
Center,
etc.
The
challenge
for
that
has
always
been
what
we
actually
used
to
do
the
minification
or
to
do
the
bundling
or
to
do
the
transpiling.
B
I
have
node
installed,
but
then,
if
they
do
want
to
bridge
that
into
the
wider
ecosystem,
that's
broadly
node
based
they
can,
and
we
would
have.
You
know
pretty
good
support
for
that.
By
way
about
JavaScript
services,
stuff
and
they're,
great,
the
beauty
of
that
is
that
it
would
still
continue
to
work,
whether
it's
running
as
middleware
at
runtime
or
whether
it's
running
at
Build
time
through
them
as
build,
or
indeed
you
invoke,
go
through
a
CLI
tool.
So
know
it's
an
idea.
B
As
I
said,
we
keep
getting
a
lot
of
feedback
about
this
area
and
we
don't.
This
is
kind
of
the
best
strategy
that
we've
come
up
with
so
far
that
we
think
there
weren't
some
investigation
in
terms
of
razor
tooling.
The
next
one
we
really
want
to
try
and
bite
off
is
the
razor
vs
in
when
you're,
using
razor
of
Visual
Studio,
getting
navigation
and
refactoring
to
work.
B
So
the
review-
and
you
see
sharp
file
and
you
rename
a
member
and
that
member
is
used
from
a
razor
file
that
actually
renames
the
member
in
all
your
razor
files.
We
have
never
had
this
ability
in
aspx
or
razor
for
the
17
years
that
this
product
has
been
around
more
now
18
years,
whatever
it
is,
but
we
have
some
work
that
we've
started
on
this
and
we're
trying
to
get
all
the
stars
to
align
across
the
various
teams
would
have
to
do
that
work
to
have
it
light
up.
B
Sometime
later
this
year,
Cloudant
services-
this
is
really
the
big
one.
This
is
the
big
workload
so
dispatcher.
This
is
something
we
started
into
one
and
then
we
stopped
we
want
to
bring
it
back,
which
is
basically
we
call
it
global
routing.
This
is
the
ability
to
have
something
that
runs
at
the
beginning
of
the
pipeline
that
understands
endpoints
like
the
URLs
and
the
metadata
and
route
token
capturing
and
all
that
type
of
stuff,
so
that
you,
a
random
middleware,
can
say
Jim.
B
It
generate
me
the
URL
to
this
endpoint
please
or
attach
this
metadata
to
the
current
endpoint,
that's
going
to
be
hit
when
you
get
there.
So
we've
done
a
lot
of
the
work
on
this
already,
but
it
just
wasn't.
It
was
just
too
big
to
fit
into
one.
So
we
want
to
page
that
back
game
for
2.2
and
it
enables
a
whole
bunch
of
other
scenarios.
We
want
to
finish
up
the
API
controller
convention
stuff
that
we
started
in
2.1.
B
So
this
is
the
API
controller
attribute,
there's
a
whole
bunch
of
more
stuff
that
we
could
infer
automatically
about
your
API
just
by
adding
some
more
conventions,
and
then
we
want
to
make
that
pluggable
so
that
you
can
write
your
own
convention.
So
we
want
to
continue
that
work
in
box
support
for
open
API.
So,
rather
than
having
to
you,
know
again
put
a
whole
bunch
of
attributes
and
stuff
install
an
external
library.
B
We
want
to
figure
out
a
good
way
of
getting
a
nice
inbox
solution
for
swagger,
slash,
open,
API,
endpoint
generation
happening,
and
then
we
want
to
go
and
generate
we'd
love
to
be
able
to
do
the
end
to
end
there
and
actually
generate
API
clients
for
you
from
that
endpoint
as
well.
Now
again,
both
these
things
would
be
fully
pluggable.
You
know
for
all
the
various
reasons
that
we
usually
have
there's
some
existing
solutions
that
Microsoft
the
wider
Microsoft
owns
around
this
area.
B
There's
the
auto
rest
tool
that
one
of
the
azure
teams
owns,
which
is
used
for
a
lot
of
the
azure
api
stuff,
I,
believe
that's
implemented
with
node.
Today,
though,
so
we
and
we
want,
we
kind
of-
want
something
more
natively
integrated
in
with
the
build
experience
so
msbuild
and
dotnet
based
so
we're
doing
some
early
investigation
into
that.
So
they
you
know,
as
it
says
there
you'll
be
able
to
then
generate
c-sharp
clients
and
typescript
clients
and
then
whatever
all
the
client
that
you
want
very
easily
based
on
your
web
api
s--
health
checks.
B
Again,
this
is
something
we
have
code
for
right
now,
it's
up
on
our
repos,
but
we've
never
actually
shipped
it.
So
that's
something
that
we
want
to
circle
back
around
and
actually
turn
that
into
a
part
of
the
product
distributed.
Configuration
is
a
really
good
one.
It's
a
concern
when
doing
stuff
in
highly
distributed
environments.
The
configuration
system
we
have
today
has
some
limitations
when
it
comes
to
doing
stuff
in
a
distributed
fashion.
B
We
need
to
solve
the
inbox
API
off
story,
so
we've
done
a
whole
bunch
of
work
around
this
previously,
but
we
haven't
shipped
it
about
shipping
a
tote,
some
type
of
token
server
in
the
box,
so
that
you
can
issue
token
to
authorize
your
AP
is
for
spire
scenarios
and
rich
clients,
Uno's,
etc,
and
also
for
front-end
back-end
scenarios.
You
might
have
a
front-end
web
application
that
talks
about
back
in
API
and
you
want
to
bet
all
of
those.
So
we
have
a
rough
idea
of
where
we
can
add
value
here.
B
Similarly,
for
front-end
and
back-end
those
type
of
things,
so
that's
what
we
think
we
can
add
some
value
there.
So
we've
got
some
interesting
ideas
around
them.
We
have
this
proxy
middleware
that
has
been.
We
have
a
repo
for
it,
but
we've
never
shipped
it,
and
there
is
some
code
in
there
and
we
every
now
and
then
people
log
issues
and
but
we
haven't
shipped
it.
So
we
want
to
again.
This
is
a
very
popular
and
frequent
concern
when
building
distributed.
B
We
have
this
pretty
cool
thing
which
I
don't
have
much
to
say
about
right
now,
but
we
have
some
ideas
around
a
better
developer
experience
for
built
for
interacting
with
Web
API,
so
while
you're
building
them
for
HTTP
API,
while
you're
building
them
today
in
vs,
when
you
build
your
web
api
app
when
you
hit
f5,
it
just
opens
the
browser
and
you
get
a
404.
It's
not
particularly
compelling.
So
we
have
some
some
nice
ideas
here.
We
think
that
might
make
that
development
experience.
B
Much
nicer
and
you
know
I
mean
postman,
is
a
fantastic
tool.
We
use
it
in
our
workshops
to
kind
of
help
people
build
api's.
We
think
we
could
build
a
nice
sort
of
inner
loop
experience
here,
as
we
call
it
out
there,
while
I'm
coding,
my
API
I
can
have
a
command
line,
experience
or
a
visual
experience
or
something
that
lets
me
easily
hit
that
API
navigate
around
it.
You
know
this
this
plugs
into
the
dispatcher
work
it
plugs
into
the
opening.
B
I
am
point
work
and
you
could
potentially
point
it
at
other
backends
as
well
and
get
a
nice
experience
out
of
that.
So
you
know
very
early
thinking
on
some
stuff
there,
but
we
want
to
investigate
that
and
then
lastly,
for
signal,
are
we
want
to
do
the
Java
client
work
as
our
next
client
after
two
point
one,
because
that
seems
to
be
the
most
frequently
asked
for
client
after
c-sharp
and
type
scripts
us
JavaScript
in
the
so
and
of
servers.
B
We
really
need
to
get
HT
be
to
support
and
kestrel
again
we
started
this
work
in
2.1
and
then
had
to
pull
it
out
just
due
to
lack
of
time
to
fix
to
get
it
finished.
So
we
want
to
do
that
again.
Continuing
the
work
on
span.
We
did
a
whole
bunch
of
that
in
2.1,
but
there's
more
that
we
can
do
so.
We
want
to
do
that
adding
generic
host
support
for
kestrel.
B
All
this
is
I.
Think
Fowler's
been
in
here
and
been
writing
in
my
issue.
So
connection
adapters
is
an
existing
feature
in
kestrel
and
we
have
this
new
IDE.
This
new
epic
called
bedrock,
which
is
this:
how
do
we
build
a
nicely
programmable
server
on
top
of
kestrel,
and
this
is
doing
that
work
into
to
the
next
step
of
that
working?
We
still
have
some
unknowns
as
it
comes
to
our
non-pipelined
request.
Throughput,
so
I've
actually
got
some
stats
up
here.
B
So
this
is
like
our
dashboard
for
performance
right
now
for
2.1,
and
you
can
see
most
of
the
stuff
is
green
or
yellow,
which
is
great,
but
the
one
that
is
red
is
Jason
and
Jason
is
interesting
because
it's
it's
the
cheapest
or
sort
of
Mike,
most
micro
benchmark.
We
have
that
represents
doing
requests
to
kestrel
and
a
spinet
core
that
are
pipelined
right.
So
plain
text
for
people
who
remember
is
actually
pipelines
requests
it
sends
15
or
16
requests
at
a
time
to
the
server.
B
Who
can
then
process
16
very
quickly
and
then
send
all
the
answers
back
in
the
same
order,
but
in
sort
of
one
chunk.
So
the
server
gets
to
optimize
a
lot
of
stuff
that,
typically
you
don't
get
to
in
real-world
scenarios.
Jason
is
the
next
sort
of
one
along
the
line
in
the
tech
and
power
test,
and
it
is
non
pipeline.
Just
just
like
normal
request.
The
client
sends
a
request,
it
waits
for
the
response
and
then
a
sends
another
request
and
it
maxes
out
to
256
clients
Inc
concurrently.
B
So
our
performance
here
is
something
that
we're
still
working
really
hard
on.
We've
got
a
we've,
got
a
nut
to
crack
here
and
we
haven't
cracked
it.
Yet,
as
you
move
towards
the
other
end
of
the
spectrum,
like
the
fortunes
test,
which
does
a
whole
bunch
of
database
access
ranges,
HTML,
etc,
etc.
Then
we
make
up
all
that
gap
again
and
in
2.1
we're
gonna
have
some
really
really
good
numbers
for
our
Fortunes
tests.
B
You
can
already
see
the
like
EF
fortunes
versus
2.0
is
31%
faster
and
the
raw
test
is
27,
nil,
nearly
28%
faster-
and
this
doesn't
include
the
new,
a
do
de
net
providers
that
we
have
coming
online
into
one
from
the
community
as
well.
If
the
Postgres
driver,
those
new
drivers
will
double
or
even
triple
those
numbers
again,
and
so
we
expect
the
fortunes
test
for
2.1
to
really
you
know
to
get
a
massive
boost
on
our
standings
like
if
we
go
over
to
check
and
power.
B
B
A
C
A
A
B
So
that's
the
idea,
like
this
I
mean
plain
text
you
can
see
a
spinet
core
is
still
well
above
the
fold
here
like
we're
in
the
top
20,
but
we're
well
off
what
the
top
performers
are
now.
But
the
thing
to
keep
in
mind
is
that
basement
core
is
is
a
framework,
whereas
these
are
service.
Alright,
these
are
these
are
just
testing
the
servers
and
you
can
see
that
by
their
classification
is
because
they're
classified
as
platforms.
So
there
are
a
couple
in
here
that
are
micro
frameworks.
B
So,
like
rust,
has
a
couple
of
really
interesting.
New
service,
like
Tokyo
I,
haven't
actually
looked
at
their
code.
So
I
can't
comment
on
what
the
code
looks
like,
but
I
can
tell
you
rapido.
It
is
actually
quite
nice
to
look
at
so
is
undertow
which
are
Java
based,
but
we
do
have
a
platform
level
test
coming
in
2.1
that
will
be
submitting
to
tech
empower,
and
currently
we
expect
that
to
land
about
here
in
between
the
reactor
and
h2o.
It's
about
6
million
requests
per
second
now.
B
B
David's
github
I
think
it's
cold.
I
already
has
platform
level
taken
power,
so
in
the
coming
weeks,
as
we
shut
down
four
to
one
RC
will
be
submitting
this
to
take
on
power
because
second
power
requires
the
benchmarks
submitted:
quote-unquote
production
ready
and
we
don't
have
a
go,
live
license
until
RC
generally,
and
so
we
won't
submit
our
two
one
updates
until
we
get
to
RC,
but
that's
going
to
be
happening
shortly,
so
we'll
submit
all
this
as
part
of
the
RC
release.
B
So
if
we
look
at
platform
level,
tech
and
power,
I
think
which
one
is
it?
Let
me
have
a
look.
It's
a
bunch
of
code
here.
Some
of
this
code,
I
wrote,
but
we
don't
the
one.
That's
interesting.
I
think
is
this
plain
text
or
I?
Think
Oh
sebastian
is
telling
me
it
is.
Oh
we
have
it
in
the
thing
now.
Oh
there
we
go
thanks,
Sebastian,
let's
try
that
here
we
go
so
we
could
poured
it
over.
Okay,
I
think
wait.
Is
this
hit
Sebastian?
Are
you
sure
it's
this
right?
B
Sorry
he
gave
me
the
wrong
way,
yeah
that
again
right,
so
here's
the
platform
edge
mode.
So,
if
I
look
at
I,
think
it's
HTTP
application,
that's
the
basic
it's
at
the
base
class.
This
is
the
okay.
Maybe
it's
benchmark
application?
Okay.
So
this
is
what
your
benchmark
app
would
look
like.
This
is
what
programming
directly
against
the
server
looks
like
all
right.
D
B
B
B
Is
similar
to
what
the
other
server
based
test
too?
So
you
can
here
you
get
a
method
when
this,
when
the
HTTP
start
line
is
sent
to
you,
you
get
a
method
that
tells
you
well
here's
an
enum
that
represents
the
method
by
user.
To
get
is
it
a
post
is
the
version
of
HTTP
that
was
sent
and
then
here's
a
bunch
of
spans
which
are
basically
just
pointers
into
memory.
Remember
you
get
a
bunch
of
pointers
to
memory
to
tell
you
here's
the
path.
Is
the
query
string?
B
No,
there's
no
collections
here,
like
there's
none
of
that
stuff.
This
is
literally
the
server
level
constants,
and
so
you
can
see
the
dispatching
if
the
path
starts
with
what
we
expect
for
the
plaintext
path
which
you
can
see
up
here
is
slash
plaintext
then
we
run,
we
say
is
plaintext
and
we
say
it's
Jason
tree.
B
So
then,
when
we
get
down
to
the
let's
actually
process
the
request,
it's
as
simple
as
well
was
it
plaintext
run
plaintext
was
it
Jason
and
Jason
and
then,
if
we
go
down
and
look
at
what
the
plaintext
one
looks
like
this
is
how
you
respond
to
a
plain
text,
so
lion
79.
We
create
a
buffer
Rider
to
ride.
Out
to
the
response
we
write
out
the
HTTP
near
the
start
line.
It
should
be
one
one.
Okay,
we
start
adding
a
couple
of
server
a
couple
of
headers
to
the
response.
B
You
know
server
header,
you
need
a
date
header,
you
need
a
content,
type
header,
a
content,
length
header
and
then
you
go
ahead
and-
and
you
actually
write
the
the
content
here
and
then,
where
is
it?
You
write
the
content
down
here.
Sorry
and
then
you
commit
that
and
that's
your
if
that's
it
like
that.
That
is
what
it
looks
like:
okay,
Jason,
it's
similar
except
you're,
writing
a
JSON
response.
Instead
of
writing
plain
text
so
down
here,
when
you
create
the
payload,
you
run
it
through.
B
The
Jason
serializer
in
this
case
I
think
we're
using
the
utf-8
JSON
library
from
nice
yeah,
which
is
a
non
allocating
jason
serializer,
which
is
why
serialize,
unsafe,
yeah
and
so
yeah,
so
that's
the
and
then
like
the
results.
I
can
actually
show
you
the
results.
So
we
have.
We
run
these
in
our
if
I
go
to
the
Pink's,
the
last
page,
you
know
this
is
all
this
is
the
public
dashboard?
Anyone
can
view
this
aka
EMS,
/ht
net,
slash
benchmarks.
This
is
the
custom
benchmarks.
B
Dashboard
I
can
choose
what
I
want
to
run
so
I
can
say.
Let
me
show
it
show
me
the
plain
text
platform
which
is
what's
being
shown
now
versus
plain
text
and
let's
look
at
it
for
HTTP
sockets
on
our
physical
servers
running
on
Linux
and
windows.
Alright,
so
we
can
see
comparison
if
I
uncheck
that
I
get
both
right.
Okay,
so
over
on
here,
requests
per
second
plain
text
platform,
which
is
green
one,
that
is
on
Windows
on
sockets.
B
So
you
can
see
that's
getting
like
three:
what's
that,
like
3.4
million
versus
full
asp
net
core
with
middleware,
which
is
sort
of
1.9
million,
so
you
can
see
this.
This
shows
you
what
the
overhead
of
the
actual
a
spinette
core
part
of
the
stack
ears
with
middleware
animation
and
service
container
and
all
that
type
of
stuff.
So
this
is
actually
for
no
other
reason.
This
is
a
nice
visualization
of
how
much
it
costs
to
go
from
kestrel
to
running
middleware,
alright
and
then
MVC.
B
B
How
you
complain-
and
this
is
this-
is
kind
of
what
taking
power.
One
of
the
things
I
was
striving
to
do
is
that
if
I
go
to
the
actual
benchmarks,
this
is
one
of
the
other
reasons.
Why,
having
a
by
having
a
by
having
a
platform
level
benchmark
is
important
because
they
do
this
diff
so
to
see
this
framework
over
think
I
can
do
it
here.
If
you
have
a
platform
level
benchmark
for
the
what's
marked
as
the
platform,
the
name
of
the
platform,
they
can
tell
you
the
overhead
of
the
framework
that
makes
sense.
B
B
A
B
That
was
a
nice
diversion
NCM,
so
I
think
I
mentioned
a
couple
weeks
ago.
We
had
to
pull
the
ANC
em
in
process
support
from
two
one:
it
just
wasn't
ready
and
so
we're
gonna
we're
gonna
continue
work
on
that
and
we
actually
hope
to
ship
that
before
2.2
we
might
do
that
as
like
an
out-of-band
thing,
we'll
see
how
we
go,
and
so
part
of
that
is
using
a
new,
side-by-side
versioning
scheme
for
the
asp
net
core
module.
B
This
is
the
part
of
is
this
is
the
bit
that
we
ship
in
is
or
that
you
install
on
is
I
should
say
that
lets
you
run
a
split
core
by
enabling
side
by
side
versioning
of
that
and
deployment.
Then
we
can.
We
can
more
easily
deliver
new
features
without
risking
you
know,
existing
applications
on
the
server,
so
we're
gonna
do
a
bunch
of
work
around
that
and
then
we're
going
to
have
a
look
at
doing
better
integration
with
Ingenix
and
apache.
B
Because
you
get
all
the
log
messages
searchable.
It's
filter
ball
all
that
type
of
stuff,
but
we're
gonna
implement
a
different
way.
We're
going
to
look
at
how
we
do
some
of
our
light
up
stuff
in
Azure
and
sort
of
revisit
that
and
see
if
there's
some
some
more
powerful
ways
that
we
can
do
that,
because
right
now,
we've
got
some
notations
around
that,
and
then
we've
got
some
other
stuff
here
about
figuring
out
certain
critical
events.
B
How
can
we
make
sure
that,
if
you're
running
in
Azure
and
you
something
goes
wrong
with
your
application,
how
can
you
very
quickly
turn
on
something
that
will
help
you
get?
You
know
information
about
what
happened
right
so
like
in
the
old
framework.
We
had
this
egw
ability,
so
you
could
just
go
and
like
remotely
turn
on
ET
w
events,
and
you
get
all
this
extra
trace
information.
We
want
to
do
something
like
that.
It's
not
exactly
that
phrase
Burnett
core
as
well
for
the
platform
we've
got
some
amp
32
work.
B
B
I
think
I
tweeted
a
few
weeks
ago
that
we
did
a
bench
a
benchmark
of
a
split
core
on
a
Raspberry
Pi
and
we
got
like
I
know:
48,000
requests
per
second
or
something
so
the
performance
is
pretty
good,
but
the
startup
performance
is
pretty
abysmal
and
this
is
what
the
cross-chaining
improves.
It
are
pre
jets
a
whole
bunch
of
stuff
to
make
this
the
app
start
up.
Much
faster,
which
is
kind
of
cool
I
mean.
B
Then
we
have
a
bunch
of
other
little
stuff,
so
some
more
performance
stuff
we
want
to
do,
and
some
of
this
is
infrastructure,
it's
kind
of
boring
and
we
the.
Lastly
a
couple
of
miscellaneous
things.
We
want
to
add
more
of
encounters
for
tracing
in
the
in
the
in
the
framework,
and
we
want
to
address
one
of
the
number
one
feedback
items
we
get
about.
Our
caching
infrastructure,
which
is
people,
want
the
old
output
caching
functionality.
B
B
So
if
you
send
you've
like
control
f5
in
the
browser
and
send
a
request
with
the
ignore
cache
header,
then
our
request,
caching,
with
a
whirl,
will
go
okay
and
the
three
run
the
requests
that
people
want
to
be
able
to
configure
on
the
server
and
say
I
want
you
to
cache
this
for
60
seconds.
No
matter
what
and
we
don't
have
that
capability
right
does.
A
B
B
A
B
A
B
That
the
stuff
in
the
spa
template
I'm
guessing,
probably
is
yeah
I,
don't
know
no
plan
that
I
know
of,
but
definitely
we
should
there's
no
reason
we
shouldn't
like
I
I'm
of
the
opinion
they
would,
if
we're
gonna
integrate
with
their
party
stuff.
We
don't
want
to
get
too
far
behind,
because
it's
just
debt
that
we
don't
help
anyone.
So
a.
B
A
B
Not
simple,
otherwise
we
would
have
done
already.
We
are
working
on
Razer
support
for
vs
for
Mac
and
I.
Think,
and
you
know,
that's
been
in
the
preview.
That's
out
right
now
doing
it
for
vs
code
is
a
lot
more
complicated
because
we
don't
have
the
vs
editor,
which
has
that
multi-language
capability,
but
we
are
in
discussions
with
the
vs
co
team
to
try
and
figure
out
what
the
path
forward
there
is.
B
Oh,
yes,
sorry
yeah,
so
we
actually
have
a
prototype
of
that
already.
So
yes,
that
would
be
on
the
list
as
well
the
C++
client
for
Sigma.
We
have
something
very
basic
in
there
already
it
won't
ship
as
part.
What's
your
point
one,
but
it
might
be
that
that
ends
up
getting
finished
for
duty
or
it
might
be
that
it
just
progresses,
but
not
as
completely
finished,
we'll
see
cool.