►
From YouTube: .NET Core Design Reviews: Backlog Review
Description
I this meeting, we'll just walk the list of open issues and PRs.
A
Alrighty
so
today
we'll
just
do
a
quick
review
of
a
backlog
which
is
basically
looking
at
the
open
issues
and
open
PRS
to
make
sure
that
we
do
something
with
them.
We
don't
want
to
discuss
the
individual
PR
so
much.
You
just
want
to
basically
make
a
decision
whether
we
should
move
them
forward
or
you
know
close
them,
but
before
we
go
into
that,
there's
one
issue
that
I
think
we
already
have
consensus
on
it.
We
just
want
to
close.
A
Which
is
the
secure
exception
proposal?
So
there
was
some
discussion
recently
about
what
we
do
with
exceptions
in
the
context
of
app
models
like
server-side
@f
models,
where
exceptions
themselves
could
leak
information
that
is
considered
security
sensitive,
and
there
was
a
proposal
to
deal
with
that
by
introducing
a
new
exception,
we're
changing
the
way
messages
are
being
retrieved
by
basically
introducing
a
new
property,
and
we
talked
about
this
with
various
people.
You
can
seize
annoyed
inside
this
15
participants.
In
this
thread
alone.
There
was
some
little
conversation
and
there
was
the
original
issue
then
raised.
A
There
was
a
bunch
of
communication
back
and
forth
as
well.
It
seems
the
general
consensus
from
the
security
experts
is
that
if
you
leak
any
information,
in
any
exception,
you're
already
in
trouble,
so
the
safest
thing
you
can
do
it
in
that
model.
Is
this
allow
any
exceptions
to
be
propagated
and
that's
really
relatively
easy,
giving
the
exception
design
that
we
have?
You
can
either
register
handlers
for
unhandled
exceptions
based
on
the
app
model,
or
you
can
just
have
a
global
try-catch
handler
and
do
something
meaningful
there.
A
If
your
developers
are
basically
leaking
information
by
logging
them
by
doing
like
exception,
to
string
or
grab
the
message
invited
into
the
response,
which
is
sometimes
eve
of
PHP
sides,
then
the
security
experts
argue
that
you're
in
trouble,
because
there's
no
way
the
fabric
can
of
a
protection
against
that.
A
The
only
actionable
feedback
here
would
be
that
we
basically
remove
all
the
useful
information
from
exception
objects
like
type
the
stack
trace
the
message,
and
at
that
point
it
becomes
basically
a
win32
API
model
where
you
just
have
an
error
code
and
the
only
really
helpful
error
code
would
be
Boop.
Something
went
wrong
which
is
essentially,
you
know,
II
fail
in
the
calm
days,
which
is
the
a
business
which
is
basically
a
framework
that
we
don't
want
to
provide
so
I
think
on
our
side.
A
We
will
just
close
issue,
as
you
know,
by
design
or
won't
fix,
depending
on
how
you
see
that
and
I
will
probly
do
this
later,
because
there's
nowhere.
You
have
to
do
this
on
the
side.
How
do
I
do
this
I
just
close
this.
A
And
what
I've
basically
done
is
I
opened
up
and
then
basically
this
is
other
one
that
was
related,
which
triggered
the
whole
discussion,
which
was
the
duplicate
key,
and
for
this
one
we
will
basically
just
merge
it
in
now
giving
up
giving
this
consensus.
So
what
I've
done
is
like
this
list
here.
It's
essentially
all
the
the
issues
and
gate
of
the
demarc
is
ready
for
sorry,
add
markers,
yeah,
Marcus,
ready
for
API
review.
C
Pretty
old
animals,
it's
sort
of
all-encompassing,
we
already
confessed
parts
of
it,
how
the
other
parts
that
have
already
been
ocellated
by
changes
that
john
hannah
stand,
making
our
gloria
changes.
I
see
so
at
this
point,
I'm,
not
sure
what
the
concrete
actually
I'm
would
being
actually
I
for
this
one
is
we
probably
need
to
revisit
it,
see
what
pieces
haven't
been
done
and
then
discuss
those
if
there
are
any
that
require
discussion
that.
A
B
A
A
A
A
B
B
C
A
B
A
Discussion
on
Twitter
mocking
our
team
showed
me
a
screenshot
of
his
twitter
app.
There
was
showing
an
exception.
It
was
embedded
cast,
exception
cannot
convert
objects
to
type
of
twitter
status,
page
or
whatever,
and
and
I
made
fun
of
that
and
said
well.
At
least
we
show
you
the
types
and
looks
like
a
duplicate
who
should
handle
this.
One,
then,
is
that
you,
because
you're
ready
engaged
with
that
did.
A
C
B
B
C
A
But
funny
enough,
I
think
we
added
some
things
where
you
can
get
the
actual
video
as
it
stored
in
the
set,
because
if
you
have,
for
example,
strings
with-
and
you
say
like
all
in
at
ignore
case,
they
still
want
to
know
what
the
original
casing
was,
as
was
as
it
was
recorded
in
the
set,
and
I
think
they
may
not
actually
have
july
based
things.
So
it
may
be
worth
looking
into.
I.
C
B
A
C
To
their
addiction
rates,
the
same
reason
they
exist
on
Initial
D
tables,
there's
synchronization
involved
internally.
So
in
order
to
get
the
comic
creator
of
it,
it
needs
to
build,
in
that
said,
there's
still
value
in
having
them
on
her
for
simplicity
and
ease
of
use,
theirs
values
having
them
on
naan
bread
same
times
as
well.
C
B
C
C
B
B
C
A
A
B
A
That's
true
for
pretty
much
any
feature
work.
We
do
unless
it's
in
the
new
self-contained
thing,
but
on
which
we
don't
have
that
many
so
alright.
So
that
was
that
one.
This
one
is
an
interesting
one,
because.
B
B
A
B
A
All
right,
these
are
all
the
ones
that
are
marquez/ap
I
edition,
but
Donald
mark
is
ready
for
review.
So
these
are
probably
one
said,
may
not
be
ready
or
we
just
have
marked
them.
Tell
it
like
you're
doing
a
very
good
job
around,
so
we
could
either
walk.
This
is
amit
could
look
at.
I
think
this
one
probably
just
open
for
the
giggles,
but.
A
A
A
A
Bottom,
basically,
what
I'm
trying
to
do
is
like
avoid
extra
items
on
our
side.
Like
is
this
something
where
the
guy
who
opened
the
PR
has
to
deliver
work?
Because
I
would
just
take
a
note
and
say
you
know
you
want
to
come
back
to
us,
make
sure
that
it
be
able
East
somebody
assigned
on
our
side,
and
if
we
have
somebody
assigned
everything
has
been
done.
We
just
need
to
review
that
would
just
tag
it,
but
just
making
sure
that
it's
not
in
some
stale
stage,
so
that
we.
A
A
B
A
There's
a
good
point.
I
just
read
the
first
sentence:
no
you're
right.
We
did
not.
How
do
I
actually
that's
a
good
point.
I
haff
that
section
of
my
dog.
B
A
A
B
A
B
B
Distributions,
like
our
subpoena
dancing
class
project,
belongs
by
the
way.
I
think
that's
true
about
the
one.
You
know
you
disclose
a
minute
victory,
which
is
that
every
comparators
where
they
have
to
be
one
of
the
basic
leads
you
logical,
like
cultural
wire
comparisons.
Things
like
that.
That's
something
to
be
should
be
baking
into
the
cold
air
inside
to
make
sure
corvex
not
happening.
Yeah.
C
Especially,
we
have
you
got
packages
for
everything
for
a
classic
there
unexplored,
that
people
use
them
and
improve
them.
Look
a
great
place
to
have
bcl
like
librarians
that
people
could
have
all
venues
in
taking
editor
their
applications
that
we
don't
have
to
consider
its
aboard
in
this
part
of
the
core
watching.
A
A
A
B
A
All
right
steps
a
to
the
distribution
of
generator.
What
the
club
yeah
that
seems
like
something
is
fairly
specialized,
so.
B
B
A
A
A
A
His
goals
may
be
to
get
them
to
donate,
to
point
0,
because,
if
he's
still
targeting
to
ponder,
we
just
had
the
methods
at
all
which
I
would
ignore,
but
I
think
for
my
point
of
view,
the
performance
benefit
would
be
still
available
right
now.
The
question
whether
we
maybe
should
put
these
guys
but
specializing
array
seems
like
a
good
idea
in
some
of
the
cases,
but
that
I
would
rather
tell
tie
this
to
the
link
performance
features
because
that's
that's
the
thing.
A
C
B
C
A
A
B
C
A
A
B
C
B
A
The
I
think
the
biggest
problem
is
that,
even
if
you
can
do
things
in
place
in
the
new
array
or
I,
you
dig
that
you
would
basically
say
you
know,
let's
say,
array,
dot
and
let's
say
convert,
and
then
you
say
the
ends
to
array
right.
The
two
are
a
call
it
because
the
last
one
does.
But
let's
say
you
wanted
actually
on
the
way
back
right,
then
you
have
to
do
two
already
at
the
end,
but
to
make
an
optimizer
to
array
has
to
create
a
new
array.
A
A
A
B
A
A
A
B
C
B
A
A
A
Even
take
a
little
factor
into
account,
I'm
actually.
Well,
that's
why
I'm
saying
this
be
could
have
maybe
I
review.
That
could
be
very
quick
to
say.
If
you
don't
need
it,
because
we
don't,
we
can
support
it
all
right
and
option
your
loved
one
or
a
results
from
transform
blog
yeah.
C
B
B
C
And
today,
basically,
we
have
ordering
implicitly
enabled
data
goes
into
a
transform
block,
regardless
of
whether
Carol
ISM
is
employed
by
the
transport
water.
Not
data
comes
out
to
the
same
order.
That
forces
on
apparel
is
amazing.
There's
no
parallelism.
That's
easy
right!
Data
computer
picked
up.
If
there's
parallelism,
we
have
to
implicitly
the
lower
reordering
buffer
at
the
end,
to
ensure
that
the
data
can
sound
right
order.
The.
C
I
was
asking
what
they
how
they
thought.
It
should
be
exposed,
I
think
more
about
it,
but
I
think
I
would
probably
add
an
an
option
to
one
of
the
existing
them.
There
are
these
options
bags
you
passive.
That
would
probably
just
be
another:
either
heating
value
or
property,
or
something
on
one
of
those.
C
B
A
A
A
B
A
Because,
at
some
point
I
think
I
saw
the
api
is
like
very
likely
voters,
but
you
would
basically
allow
people
to
set
all
the
properties
with
one
method.
I
think
they
at
some
point
had
option
of
tea
overloads,
so
they
couldn't
differentiate
between.
I
want
this
property
to
be
now
versus
I.
Don't
want
up
to
the
property
at
all
then.
A
B
B
A
B
A
A
A
Yeah
I
just
yeah.
Basically,
what
I
would
think
is
you,
which
is
marked
as
as
a
PR,
a
very
fragrant
view
talk
about
whatever
we
would
be
want
to
add
this
type,
regardless
of
the
specific
surface
area
right
because
I
mean,
as
you
said
like,
should
we
just
tell
people
like
we
have
existing
characters?
You
can.
C
A
B
A
C
A
A
B
B
B
A
At
all
the
stringy
PS
and
try
to
reconcile
Thank
You
Jimmy,
you
could
just
have
a
menu
methods
convention
that
just
does
ignore
case
right,
which
is
basically
what
David
said
will
be
reasonable.
I,
don't
like
the
bull
versions
that
we
had
before
because
six,
what
does
that
meet
night?
Is
it
the
case
sensitive
or
is
a
case-insensitive
like
an
I?
Remember
you.
A
Generic
ingham
parts
matt
said
actually,
if
this
one
probably
has
enough
data
to
actually
talk
about
this
inner
beauty,
we
can
at
least
say
as
it's
a
direction
you
want
to
go
after
it.
If
so,
what
that
would
look
like
alright
implemented
generic
enum
parcel
method
at
this
one
I'm
always
surprised
to
be
done.
There
never
had
that
one.
Anybody
know
why
we
never
ended
up
doing
that.
Oh
we're
not
right.
B
A
B
A
So
what
I
want
to
do
is
I
want
to
make
sure
that
I
can
basically
what
I,
what
I
generally
want
to
do
and
I
think
I
never
did
a
good
job.
But
basically
the
idea
is
that,
but
if
I
peer
reviews,
things
that
I
can
actually
put
an
agenda
and
I
make
sure
that
be
I
cluster
them
accordingly,
so
that
the
meeting
itself
has
some
better
focus,
then
just
a
random
macro
could
be
aware.
We
just
brushed
through
like
20
items,
three
items
that
make
sense.
A
C
Yeah,
there's
a
definitely
value
in
such
methods.
The
problem
is:
there's
also
a
fair
amount
of
policy
assume
exactly
how
you
want
things
to
behave,
which
I
think
makes
it
tricky
to
include
core
libraries.
I
mean
I'm
blog
post
reference.
I
think
I
have
like
six
different
implementations
that
have
pros
and
cons.
C
C
A
Should
we
just
basically
have
a
meeting
to
establish
a
principle
to
say
locally,
there's
many
ways
to
do
it,
but
just
pick
one
and
then,
if
you
really
care
about
the
other
ones,
you
just
come
with
the
other
ones
yourself,
which
there's
a
better
way
to
not
give
people
any
implementation.
Because
you.
C
C
A
C
A
B
C
C
B
A
A
Alright,
so
then
it
is
review.
The
first
page
is
apparently
like
four
more
pages
to
go.
I
will
probably
hopefully
spend
some
time
this
week
to
actually
do
that,
but
I
think
we
have
enough
things
on
the
backlog,
not
actually
skin
side
up
a
bunch
of
meanings
to
look
at
these
things,
tech,
n
box
and
then
just
make
sure
that
we
can
actually
close
in
them
is
anything
more
than
you
guys
expect
to
show
up
anytime
soon,
like
the
major
thing
today
to
review
nothing
major.