►
From YouTube: Silk.NET Coffee & Code Catchups
Description
Join the Silk.NET team for a quick catchup. This is a more informal stream wherein we'll be discussing what we've been doing on the Silk.NET project, what we've been doing outside of Silk.NET, bouncing ideas off each-other, and (most importantly!) provides us an opportunity to answer any questions the community may have.
These streams are impromptu and the VOD will be unlisted as a result.
GitHub: https://github.com/dotnet/Silk.NET
Discord: https://discord.gg/DTHHXRt
Background music by HOME, curated by Tyler McVicker.
A
A
B
Well,
yeah,
I'm
not
on
the
stream
either.
So
I
have
no
idea
if
it's
working.
B
People
are
watching
this
stream.
Hopefully
this
gives
an
idea
of
just
how
informal
this
one's
going
to
be
compared
to
our
last
one.
B
Okay.
Well,
I
was
gonna
have
background
music,
but
it's
annoying
me
at
the
moment,
so
I'm
gonna
pause
it
so
yeah
I
mean
for
anyone
watching.
This
is
just
a
quick
catch
up
for
the
silk
devs,
primarily
just
just
to
catch
up
on
what
we're
doing
on
3.0,
as
well
as
get
any
input
from
the
community
and
answer
your
questions.
B
If
you
have
no
idea
what
sil.net
is,
this
is
going
to
mean
absolutely
nothing
to
you,
but
obviously
this
is
just
an
impromptu
stream,
so
I
won't
keep
it
up
on
the
channel.
This
is
just
for
now
I
mean
the
link.
It
will
still
be
there,
but
it'll
be
unlisted.
If
you
have
the
link
so
yeah
yeah,
kai
how's
it
going
just
generally,
not
necessarily.
B
I'm
not
I'm
not
doing
bad
at
all.
I
mean
it's
yeah.
I've
had
enjoy
I've
been
enjoying
the
long
weekend.
It's
it's
been.
Nice
had
a
bbq
over
the
weekend
had
burgers
for
breakfast
this
morning
as
well,
which
yeah.
B
Been
an
unhealthy
weekend,
but
yeah
yeah,
it's.
B
C
Well,
just
come
back
so
not
much
yet,
but
in
general
I've
started
working
on
the
symbols
at
the
mission
layer
or,
like
I
don't
know,
we
still
haven't
decided
what
to
like
call
this,
but
basically
the
thing
that
turns
the
symbols
into
strings
and
I
think
it's
going
pretty
well
yeah
I've,
just
I
I
think
I'd
add
pinkview
in
teams.
So
I
think
so.
I've
just
put
up
the
teams
in
the
in
the
team's
channel.
I
mean
you
know.
C
B
Okay,
so
yeah,
I
guess
I
mean
we
were
talking
yesterday
in
text,
but
I
mean
it
got
a
bit
unwieldy
because
there
was
like
three
different
conversations
going
up
oh
before
before
I
could
carry
on
furball.
How
are
you
doing
if
you
wanted
to
pitch
in
it's
nice
to
have
you
here.
B
Or
not,
I
guess
no
worries
I
haven't
got
there.
I've
got
the
text
channel
up.
C
B
C
Up
yeah,
I
would
open
this
one,
oh
yeah,
and
that's
just
like
in
the
last
like
five
minutes,
I've
just
wanted
to
fix
ci
and
then
I
stumbled
across
the
website
deployment
script
and
so
I've
I'm
throwing
together
a
quick,
quick
script
to
deploy
the
3.0
docs
to
somewhere
else.
B
Oh
wait:
the
3.0
docks!
Well,
where
are
you
what,
where
are
you
thinking
of
deploying
them,
then.
C
I
was
just
going
to,
I
haven't,
actually
figured
out
how
to
best
do
it
yet,
but
I
was
thinking
to
just
push
the
3.0
dogs
to
like
a
private
cf,
half
player
kv,
and
then
I
could
just
like
host
it
at.
Like.
I
don't
know
like
dogs,
rio,
dots,
kaya,
the
tech
or
something
I
don't
know
just
something.
So
I
can
look
at
what
I
have
documented.
What
I
have
documented.
B
C
Yeah
so
yeah
just
just
throwing
that
together.
B
So
I
I
guess
to
catch
up
on
what
was
going
on
yesterday,
you
were
looking
at
the
silk
touch
symbol,
stuff.
I
shall
pull
up
that
pull
request,
so
we
can
look
through.
B
Yeah,
I
remember
having
a
whiz
through
that.
I
don't
remember
if
I
looked
too
in
depth
about
it,
but
you
know
what
I'm
sure
it's
fine!
It's
your
code,
not
mine,.
C
Well,
well,
you
know,
I
I
mean
I've,
I've
put
lots
of
effort
into
building
all
of
the
tests
so
not
much.
I
can
hopefully
do
wrong
so
yeah.
B
I
do
love
tests,
that's
something
that
we
did
not
have
pretty
much
at
all
in
2.0,
yeah.
C
B
Yeah
2.0
was
such
a
rush
development
cycle.
It
feels
good
to
actually
take
our
time
and
think
about
things,
because
that's
something
I
don't
think
we
did
last
time.
It's
actually
a
thing.
Yeah
I
mean
we've
got
a
really
usable
library,
I
think,
but
I
mean
it's
just
you
know.
If
we
want
to
make
it
properly
usable,
then
yeah.
B
B
Seems
to
be
just
about
holding
itself
together
yeah
I
mean
at
this
point:
we've
got
community
contributions
coming
in,
which
seems
to
be
doing
most
of
the
improvement
side
of
things
and
then
from
my
side,
I'm
just
running
a
generator
every
month
fixing
it
when
it
falls
over.
C
B
I
think
our
philosophy
from
the
start
was
always
to
move
fast
and
break
stuff,
and
we
sort
of
lost
that
in
two
point
actually
well,
probably
because
we
don't
want
to
break
outside
of
a
major
release,
which
I
think
is
a
sensible
stance
as
much
as
people
in
the
community
aren't
a
fan
of
it
but
yeah.
I
think
3.0
gives
us
an
opportunity
to
redesign
everything,
rethink
everything
and
make
it
a
bit
more
stain.
B
Right
this
pull
request.
I
will
look
at
in
a
second
I've.
Just
got
some
discord
messages
I
wanna
handle.
I.
B
Talk
anything
else.
C
Not
really,
oh,
actually,
I
was
wondering
earlier
today.
Is
it
like
a
german
thing
to
like
have
easter
and
like
have
like
a
bunny,
bring
chocolate
and
things
like
that
or
something.
B
Yeah
yeah
we
have
the
easter
bunny,
that's
a
oh
yeah,
there's
a
thing
here
as
well:
I'm
not
sure
where
it
originates
from
I
mean
we
could
spend
this
stream.
Looking
at
the
origins.
C
B
B
I
mean
yeah,
I
think
if
it
hasn't
reached
some
parts
of
the
globe,
then
I'm
sure
there's
a
giant
chocolate
conglomerate
waiting
to
push
it.
B
Do
love
cadburys?
Do
you
love
capris,
right?
Okay,
I
think
I'm
about
ready,
I'm
just
gonna
try
and
pull
up
what
I've
got.
Oh
actually
now
that
could
be
a
bit
risky.
Let's
do
it
anyway
and
see
if
the
stream
reveals
any
personal
information.
This
is
the
button
awesome.
Okay,
so
on
the
stream
at
the
moment,
is
the
pulling
quest.
Let's
do
a
bit
of
a
live
review
thing
as
you
ping
this
over
me
right
as
the
stream
was
starting.
B
Yeah
so,
admittedly,
I
think
on
on
the
structure
side
of
things
I
mean
I
sort
of
talked
to
you
in
in
the
discord
about
how
I
think
we
misinterpreted
what
we
named
things
with
regards.
B
And
shared
infrastructure
I
mean
I
do
think
it's
good-
that
we've
got
the
symbols
package,
because
that
way
that
can
be
used
standalone
if
any
generators
want
to
use
that
yeah.
It's
just
where
the
emitter.
I
I
think
it's
fine
for
development
to
have
the
emitter
package
actually
do
the
c-sharp
emitting,
but
then
we'll
just
refactor
it
later
on
anything.
B
C
C
But
yeah
I
haven't,
I
just
haven't
bothered
with
that.
Yet
it
is
always
kind
of
approval,
but
I
have
like
an
outstanding
pr
and
like
back
then
like
major
refactors
like
using
records
or
everything
isn't
like
possible
because
it
breaks
the
other
pr
and
it's
all
such
a
pain.
C
B
The
other
one
yeah
like
when
this
this
pr
needs
this
pr,
but
when
that's
merged
retargets
it
to
the
main
branch,
yeah,
hey
yeah,
this
all
seemed
fine,
because
I
don't
really
know
what
to
complain
about.
I.
B
I
mean
actually
you've
written
unit
tests.
Let's
just
look
at
the
unit
test
and
see
what's
going
on
yeah,
oh
so
you
have
got
with
mock
at
the
end.
The
mock
package.
C
Yeah
I
did
go
with
mark
just
a
few
like
that's
just
used,
I'm
just
using
that
to
like
create
a
create,
like
a
instance
of
the
symbol
visitor
because
it
is
abstract.
I
can't
do
that
directly
and
then
I
can
just
use
it
to
verify
that
the
method
is
called
and
like
it's
basically
just
a
glorified
method
counter
for
us,
but
you
know
it
works.
B
Yeah
it
works,
I
mean
we
don't
need
anything
else
other
than
that.
To
be
honest,
at
least
I
don't
think
so.
B
A
B
Without
the
ridiculous
amount
of
time
to
convert
it
to
a
string,
as
well
as
the
more
specialized
use
case,.
C
And
like
a
big
problem
is
also
with
current
suck
patch
that
I
basically
gave
no
when
it
came
to
like
spacing.
We
just
used
like
the
that,
like
normalized
white
space
stuff
and
that's
not
great,
like
it's
really
slow
and
I
don't
know
how
to
improve
it.
And
so
here
all
of
the
generation
just
includes
white
space
and
like
the
test
tests
that,
like
the
white
space,
is
there
and
it
makes
code
like
slightly
more
ugly.
C
But
it
also
means
that
the
constructed
syntax
can
just
be
turned
into
a
string
directly
without
any
like
extra
steps.
And
that's
pretty
pretty
quick.
B
B
C
The
stream
is
slightly
lagging,
but
you
can't
use
name
up
there
because,
because
the
method
is
protected,
and
so
you
can't
like
c
sharp
just,
doesn't
allow
you
to
use
a
type
and
then
like
access
the
the
protected
method.
B
C
Awesome,
the
stream
is
like
slightly
lagging,
but
you
just
slightly
one
of
the
things
I
was
commenting
about
also,
and
there
yeah,
the
the
field
type
is
visited.
If
you
look
at
that
at
that
method,.
C
Yeah
there
you
can
like
see,
they
can
see
like
the
fields
look
kind
of
funny,
because
they
basically
they
need
a
type
right
like
the
field,
obviously
needs
to
type,
and
I
wasn't
so
sure
how
to
represent
the
type
so
for
now.
C
I
just
did
I
just
used
type
symbol,
which
you
know
would
be
a
struct
symbol
or
like
a
class
symbol,
but
that's
not
really
that
reasonable,
because
well
you
know
what
what
would
we
do
if,
like
we
had
like
a
like
integer
right
that
that's
like
not
one
of
our
types,
so
we
would
like
somehow
to
reference
it
and
like
it
like
works
like
I
can
do
like
new
struct
symbol
and
then
just
give
it
like
in
as
identifier
but
that's
kind
of
ugly.
C
So
I
thought
like
what
did
I
think
again.
I
have
to
actually
open
my
own
comment
again.
C
I
think
I
said
we
would
need
like
a
different
type,
probably
like
an
external
yeah
external
type
reference,
which
would
just
mean
you
know
reference
some
other
type,
and
it
would
probably
do
that
like
by
namespace
and
like
full
name,
probably
but
yeah.
I
wasn't.
C
I
wasn't
so
sure
how
to
like
do
this,
because
the
current
way
is
kind
of
ugly
and
also
there's
no
way
to
like
represent,
for
example,
in
builds
right,
like
ins
like
as
keyword
as
an
inbuilt
type
and
like
there's
no
need
to
like
reference
that
by
name
so
yeah,
not
not
so
sure.
B
I
mean
if
we,
if
we
really
want,
we
could
do
like
a
type
reference
symbol
or
like
a
primitive
type
symbol,
but
I
think
that
I
mean
if
it
works,
then
I
guess
it's
fine,
because
I
know
in
I
know,
in
silk:
church
2.0.
We
like
globally,
fully
qualified
everything
which
I
presume
we'll
need
to
do
for
this
as
well
for
safety.
C
Yeah,
I
would
I
would
like
to
do
this
for
safety
everywhere,
but
not
for,
like
prototypes,
like
what
I
really
don't
want
is
like
global
and
then
like
the
two
colons
and
then
system.string
right,
like
I
think,
that's
super
ugly
and
like
kind
of
necessary
so
yeah,
I
kind
of
feel
like
having
some
kind
of
like
external
riff
type
reference
or
something
that's.
Instead,
there
is
probably
good
mean.
B
Yeah,
that
makes
sense,
I
mean,
to
be
honest,
I
mean
just
focusing
on
the
usage.
I
think
this
all
looks
really
clean.
To
be
honest,
especially
like
what
was
the
other
that
I
scroll
past.
B
Yeah
this
the
ones
where
it
actually
uses
the
symbols
to
convert
it
to
a
string.
I
think
that
looks
really
good,
so
I've
probably
missed
a
bunch
of
the
other
pl.
Now
I'm
actually
paying
more
attention
for
the
purposes
of
this
stream,
but
yeah.
I
think
I
think
the
actual
code
you've
written
is
relatively
small.
Most
of
it's
just
tests,
which
is
good
to
see
you
know
it's
not
really.
C
Confident
that,
like
I
could
like
test
some
edge
cases
here,
but
I've
also
kind
of
built
everything
so
like
if
you
really
want
to
produce
like
invalid
c
sharp,
then
like
go
ahead,
for
example
like
identifiers,
don't
actually
validate
any
like
identifier
rules,
so
like
no
spaces,
no,
like
weird
characters
like
you,
can
just
put
a
space
and
a
struct
name,
and
you
know
you'll
get
invalid
c-sharp
but
yeah.
B
C
I
mean
I
mean
this
is
by
no
means
like
if
you,
if
you
put
this
on
like
a
web
server,
then
you
guaranteed
have
the
first
c-sharp
injection
vulnerability
ever,
but
you
know
we're
not
planning
to
do
that
anyway.
So.
B
Yeah
I
mean
this
is
really
really.
I
think
our
symbols
are
more
of
a
specialized
kind
of
thing.
I
doubt
this
is
going
to
be
as
much
as
I
am
thinking
about
other
generators
who
might
want
to
use
it.
I
don't
think
it's
that
important
that
this
is
like
a
general
purpose
thing,
because
I
think,
if
that
is,
if
we
do
go
for
a
general
purpose,
generator
framework,
then
the
scope
creeps
very
quickly.
C
For
sure
I
mean
part
of
the
thing
that
I
think
the
emitter
is
named
c.
Sharp
emitter
is
just
because
you
know
you
can't
name
the
class
the
same
as
the
namespace,
and
so
I
just
named
it
c
sharp
emitter,
because
you
know
not
really
like
trying
to
say.
Oh,
we
should
make
like
a
you
know.
Visual
basic
emitter,
like
I
said
yesterday,
just
you
know
just
naming
is
hard
and.
B
So
I
I
guess,
looking
at
this
field,
snippet
is
it.
I
noticed
that
it's
just
hard-coding
the
public
keyword
into
the
generator
code.
Are
we
just?
Is
that
just
for
prototype
purposes?
Do
you
see
us
needing
to
change
that
out?
Is
it.
C
Read
fields
but
didn't
really
like
see
if,
like
when
we
use
that
ever.
B
Bit
fields:
oh,
we
use
it
in
balkan
video.
I
think
that's
probably
a.
A
C
C
C
So
in
my
like
nodes
in
my
like
tracker,
I
put
it
so
I
will
add
a
struct
layout
thing
to
the
single
layer
and
then
that
will
can
be
extended
to
your
bit
fields,
which
can
then
which
just
get
its
properties
but
like
the
underlying
fields
that
is
behind
like
bit
fields
that
wouldn't
actually
be
represented
in
the
symbol
layer
at
all.
So
everything
that's
like
in
the
symbol
layer
should
be
like
public.
I
guess
I
think
that
makes
sense.
At
least
I'm
not
sure.
B
No
yeah,
well
I
mean
if
we,
if
we
have
it
so
the
so
you
only
sort
of
create
public
fields
and
then
just
say:
oh,
this
field
is
actually
a
bit
field
and
then
the
emitter
just
emits
all
the
private
magic
that
it
feels
is
necessary
under
the
hood.
Then
I
suppose
it'll
be
fine,
I
mean
is
it
it
is
a
bit
limiting,
but
I
think
I
I
don't
think
we
need
to
care.
C
C
But
I
kind
of
don't
see
a
case
where
you
need
the
like
the
overloader
or
the
scraper
to
actually
generate
a
private
view.
C
B
B
B
C
B
Just
happened,
oh
I've
got
noise
suppression
on
so
it
could
just
be
that
I
could.
It
could
be
waffling
so
much
that
rtx
voice
thinks
it's
just
background.
Chatter.
C
Yeah,
what
looks
like
the?
What?
What
do
you
like
did
you
have
any
thoughts?
How
you
could
like
resolve
that,
because
I
think
we
like
agree
but
also
like
how?
How
can
we
like,
like?
What
do
we
name
things
that,
because,
like
to
me,
it
kind
of
makes
sense
to
name
the
thing
that
turns
a
symbol
into
a
string?
B
Yes,
yeah.
I
agree
the
emitter
was
just
well
to
be
honest.
This
the
naming
came
from.
It
was
sort
of
mirroring
cs
win32,
because
they've
got
a
scraper
and
an
emitter,
and
then
the
overloader
was
just
well
that's
self-explanatory.
So
we
can
rename.
C
B
C
C
You
know
a
module
in
the
initializer
or,
like
you,
know,
the
pinwheel
functions
themselves
and
like
the
assembly
resolver,
and
all
that
and
I'm
not
sure
how
that
can
like
be
represented
in
the
simple
layer
like
I
thought
we
would
just
like
have
like
a
native
call
symbol
and
then
all
of
the
native
core
symbols
would
be
like
found.
Somehow
I
don't
know,
and
then
you
know
the
the
part
that
actually
generates
the
rogue
stuff
would
just
you
know,
see
all
of
them
and
and
say:
okay.
C
B
A
B
Suppose
it
is
baking
a
lot
into
like
the
symbol
emitter
as
I'm
just
going
to
refer
to
it,
but
at
the
same
time
the
symbols
are
meant
to
be
specific
to
our
use
case.
So
I
don't
know
and
there's
also
there's
also
the
fact
that
I
mean
every
binding
will
probably
be
different
in
its
own
right
and
it's
just
like
a
native
cool
symbol
like
it.
It
would
generate
a
lot
of
code
under
the
hood
it
would.
It
would
be
just
like.
C
Yeah,
I
mean
that's
the
other
big
part
that
I'm
still
completely
missing
like
how
we
can
like
how
our
configuration
will
like
work.
Like
that's
completely.
I
don't
know
like.
I
know
what,
like
the
the
user-facing
stuff
looks
like
with
the
attributes
and
and
the
configuration
file
and
all
of
that,
but
how
it
can
like
look
to
the
stages
is
completely
lost
on
me.
B
C
Yeah
I
mean
it's
pretty
clear
thing
that
we
will
have
like
one
like
mood
configuration
file
and
then
it'll
just
get
like
more
specific
from
there
and,
like
I
kind
of
know
how
I
can
like
do
that,
like
I
can
probably
write
some
custom
configuration
loading
code
for
folder
for
the
die
stuff,
like
I
don't
think,
that's
so
problematic,
but
I,
like
attributes,
are
kind
of
like
specific
to
methods
of
to,
like
you
know
anything
you
put
them
onto
types
to
whatever
and
then
yeah.
I
don't
know
like.
C
Probably
probably
have
to
do
something
like
the
razna
api
does,
which
has
like
a
method
that
you
can
like
give
your
own
workspace
and
the
like
symbol.
You
want
configuration
for
to
and
then
it'll
like
figure
out.
The
right,
editor
config,
because,
like
editor
config,
is
also
kind
of
complicated.
You
know
and
then
like
figure
that
out-
and
we
probably
need
to
do
something
similar.
B
Yeah
and
well
I
I
mean
it,
though
the
way
I
saw
it
is
that
the
attributes
they
weren't
technically
configuration,
I
mean
why
they
are,
I
mean
I
just
sort
of
saw
that,
like
this
attribute
activates
x,
part
of
the
generator.
B
So
I
don't
know
I
mean
I,
I
definitely
wouldn't
fuse
like
the
configuration
file
and
the
attribute
quote-unquote
configuration
together,
but
how
we're
actually
going
to
express
it.
I
don't
know
I
mean
to
be
honest.
I
just
I
just
saw
something
similar
to
roslyn,
where
we
just
had
a
list
of
attributes
and
then
we'd
pick
through
that.
C
C
A
B
C
Yeah,
so
that's
like
why
I
mean
like
if
it
you
give
like
some
method,
like
I
don't
know,
some
utility
method,
the
the
method,
symbol
and
then
it
figures
like
out
okay.
So
this
has
these
attributes
to
apply
that
and
then
and
then
look
where
the
method
and
what
like
file
that
is
defined
and
then
like
go
up
in
the
tree
from
there
and
look
for
any
like
configuration
file.
That's
like
named
some
specific
pattern.
C
B
B
It
does
I
mean,
to
be
honest,
I
think
that
even
is
a
bit
over
complicated.
To
be
honest,
I
mean
the
way
I
see
it
is
that
the
configuration
file
like
the
json
and
the
editor
config.
That
is
the
what
like,
what
the
generator
does
and
no
that's
the
how
sorry
it's
yes,
it's
that
it's
the
how,
rather
than
the
what
so
it's,
how
the
generator
does,
what
it
does
and
then
the
attribute
just
says
what
what
it
does
like
it
generates
a
native
call
using
whatever
I
don't
know.
B
A
C
Symbol
because
because
then
we
have
to
use
roslyn
everywhere
and
like
as
much
as
I
love
the
muslin
api.
I'd
really
rather
not
leak
it
into
the
overloading
stages
because
it
makes
like
overloading
as
complicated
as
is
but
like
leaking
the
rosten
stuff
into
it
is
like,
like
that's.
That
would
complicate
things
a
lot.
I
think.
B
I
mean
we
could
we
could
just
like
pass
the
roslin
stuff
into
our
own
representation
and
do
that
when
we're
when
we're
actually
operating
on
it.
I
just
don't
think
that
that
that
mood
that
seeing
attributes
in
the
configuration
file
as
one
it
doesn't
really
make
sense,
because
they
I
mean
fundamentally,
they
will
be
used
for
different
things.
B
C
I
mean
that
that's
that's
obvious,
so
like
we,
because
I
think
the
the
scraper
like
it
would,
for
example,
tell
us
you
know,
you
know,
tell
us
like
a
bunch
of
native
methods
and
then
tell
I
guess,
like
tell
the
old
time
like
the
the
symbol
like
okay
and
you
have
to
be
like
look
like
from
this,
you
know
have
to
be
loyal
from
like
this
native
file
like
from,
like
I
don't
know,
vulcan
one
dll
or
whatever,
and
then
you
know
it'll
it'll
like
go
like
that,
so
there's
no
like
an
attribute
involved
and
then
for
like
the
rustling
sources,
so
that
that's
really
only
that
only
concerns
cases
where
the
source
generator
is
used.
B
Yeah,
so
to
be
completely
honest,
I
don't,
I
think
the
scraper
will
be
the
only
thing
that
uses
the
actual
configuration
file.
I
don't
really
see
a
sp,
I
mean
yeah,
because
the
overloader
is
activated
using
attributes
which
the
scraper
will
add
to
the
symbols.
It
creates.
C
C
For
example,
then
that
would
have
to
be
converted
into
a
method
declaration
or
like
into
a
method
symbol,
and
we
just
have
to
get
like
a
property
or
something
that
tells
the
emitter
where
to
load
it
from
and
the
scraper
would
do
itself
so,
for
example,
vulcan
which
the
scraper
imports
vulcan,
which
at
like
the
scraper,
would
just
say.
Oh
this
has
to
be
loaded
from
you
know,
vulcan
1.0,
so
I
think
that's
probably
how
that
would
work
and
then
there's
no
configuration
involved
in
the
overloader
and
the
screen
and
emitter.
I
guess.
B
Yeah
I
mean,
admittedly,
I
was
a
bit
in
and
out
there.
The
words
didn't
really
register.
I
guess
it's
a
monday,
but
yeah.
B
But
but
yeah
I
mean,
to
be
honest,
I
mean
the
way
I
see
it.
The
yeah
the
attributes
are
sort
of
verbs.
It's
saying
it
I
mean
they're
doing
words,
whereas
the
configuration
file
is
just
sort
of
like
well
sort
of
an
adjective.
I
guess
I
mean
to
be
honest.
If,
if
we
really
wanted
to,
we
could
have
the
entire
scraper
configuration
represented
as
attributes,
but
I
just
think
that'll
be
that'll,
be
unholy.
A
B
C
C
B
Yes,
I
mean
if
everything
goes
well
in
the
current
version
of
the
proposal
that
is
gone
entirely
in
favor
of
the
I
forget
what
it's
called.
Let
me
pull
it
up.
B
B
Yeah,
so
we
in
3.0
we
just
have
use
dynamic
library
which
takes
multiple
names
and
tries
them
all
on
whatever
platform.
Oh.
C
Yeah,
that
makes
sense,
so
I
guess
then
we
just
have
like
an
array
or
whatever
right
like
just
something
that
then
tells
the
the
bit
that
generates
the
code
to
invoke
the
native
library
like
where
to
load
it
from
and
then
like
not
no
like
attribute
is
required.
C
B
Guess,
well,
I
suppose,
if
I
mean
with
how
shared
infrastructure
is
working,
I
mean
if
we
just
have
a
native
call
symbol
and
that
doesn't
gain
meaning
until
we
actually
reach
the
point
where
we
generate
the
actual
p
invoke
stuff
like
well,
for
example,
name
containers
which
I
know
won't
exist,
but
that's
just
an
example
right
and
if
that
doesn't
get
generated,
we
can
just
well.
B
C
Yeah
exactly
and
then
just
all
the
informations
on
the
symbols,
and
we
have
no
worries
it'll,
just
be
like
a
huge
pain
for,
like
you
know,
like
setting
up
the
roster
and
import
thing
which,
like
has
to
convert
rosalind
some
bulls
into
into
our
symbols.
But
that's
a
that's
a
worry
for
for
another
day.
Yes,.
B
B
Oh,
hang
on
my
sound's
gone.
Bear
with
me.
C
I'm
doing
my
best
yeah,
I'm
I'm
still
here.
I
actually
can't
hear
me,
unfortunately
getting
his
his
headphones
have
broken
again.
C
Actually,
a
new
person
just
joined
that
that's
really
that's
that's:
output,
oh
and
and
the
person's
gone
again
welp.
I'm
just
I've
just
opened
the
stream
to
see
if
anyone's,
actually
watching
us.
Well,
I
I
guess
we're
just
waiting
yeah.
C
Someone
of
the
now
against
seven
people
who
are
watching
has
has
anything
to
say
about
now.
Eight
people,
actually
that's
that's
awkward,
has
anything
to
say
about
you
know
I
don't
know
their
hopes
their
dreams
for
for
3.0.
B
B
C
B
C
B
The
scraper
is
basically
well,
none
of
our
codes
actually
started
being
written.
I
mean
we've
we've,
we
invoking
clan
shop
at
the
moment
and
we
are
generating
an
xml
file.
What
we
need
to
do,
then,
is
read
that
xml
file
and
create
symbols
from
it
and
then
have
a
pipeline
of
renaming
those
symbols.
C
C
Yeah,
exactly
and
like
we
can
actually
work
with
it
right,
yeah
because
like
when
we
have
everything
that
clank
shop
has
in
our
symbol
layer.
It's
like
super
easy
to,
like
add
renaming,
add
overloading
at
all
those
like
nice
things,
but
it's
just
you
know,
building
that
that
bridge
first
yeah.
A
C
Ideally,
once
you
have
like
the
bare
minimum
so
like
that
that
be
like,
you
know,
maybe
triple
the
time
we
need
for
building
the
dead
minimum
wire
product.
We
can
probably
like.
That's
probably
then
like
a
preview
one
where
we
have
like
you,
know
the
basic
vk
type.
Again,
we
have
the
all
the
methods
we
have
like
basic
overloads,
which
we
have
today
so,
like
you
know,
span
overloading
strings
things
like
that,
and
we
have
the
improved
loading.
C
So
like
that's
like
my
main
thing
that
I
want
to
get
out
is
is
as
soon
as
possible,
get
people
to
test
the
new
loading
mechanisms
which
have
no
little
to
no
overhead.
Like
that's
the
that's
the
main
goal
for
preview
one,
but
I
guess
way
out
for
now,
like
I'm
hoping
to
like
in
in
six
weeks,
get
the
get
the
mvp
that
can
just
like.
Why.
B
B
C
C
But
yeah
like
at
like
six
weeks,
seems
visible
because
at
the
pace
that
I'm
moving
right
now,
I
can
probably
get
like
all
of
the
mission
done
in
like
two
weeks.
Hopefully.
C
B
B
Yeah,
the
the
the
vk
video
stuff
that
is
actually
generic,
clang
sharp,
is
generating
bindings
today
for
vk
video
using
our
configuration.
We've
got
that
file,
so
we
d
and
from
there
it's
just
reading
the
xml
and
creating
symbols
from
it,
which
I
think
is
it's
a
fairly
simple
task.
It's
just
a
lot
of
it.
C
Well,
yeah,
I'm
not
so
confident
there,
but
because
I'm
just
like
this
is
the
typical
kind
of
integration
problem
right,
like
I'm,
building
all
the
simple
layer
to
be
nice
for,
like
my
main
work,
so
emission
and
like
overloading
and
then
eventually
I'll
sit
on
the
other
side
and
then
hate
myself
for
building
the
symbols
in
this
extremely
stupid
way.
That
is
horrible
for
building
the
xml
posit.
But
you
know
so
so
I'm
I'm
kind
of
worried
there,
but
yeah
future
future
problems.
Future
holders.
B
C
Yeah
I
mean
I
think,
since
the
two
like
since
I
made
silk
touch
like
as
like
back,
then
I
was
still
kind
of
like
back
then
I
felt
like
this
was
great
code,
because
I
understood
everything
and
I
was
like
you
know
I
was
I
was
writing
the
code
every
day,
and
so
I
understood
everything
and
so
from
from
like
today
I
like
you
know
I
don't
get
anything
there
anymore.
So
you
know
at
the
time
we
felt
like
that
was
good
code,
but
like
at
least
I
thought
like
that
was
amazing
code.
You
know.
C
B
C
C
And
then
I
looked
it
up
and
actually
I'm
output.
B
Oh
kai's
gone
you're
right.
C
Well,
this
code
just
went
out,
I'm
not
sure
why,
but
this
could
I
I
was
searching
for
oh
nine
out
of
seven
in
this
chord
and
it
just
I
don't
know
just
didn't
want
to
just
didn't
want
to
be
anymore.
C
I
wish
I
could
find
the
the
explanation
for
that
code,
but
it
doesn't
matter
yeah,
but
I'm
I'm.
You
know,
I'm
pretty
confident
in
the
current
code
base
you
know,
might
might
be
a
problem
with
clang
sharp,
but.
B
Yeah
I
mean
bank
drop
is
relatively
nice
and
I
mean
I'm
always
here
to
ask
questions.
I
mean
most
of
my
silt.net
time
at
the
moment
is
spent
holding
two
point
x
together,
because
I
I
really
don't
have
a
lot
of
free
time
on
my
hands,
but
I
mean
I
always
hate
to
answer
questions.
Should
you.
C
B
C
Actually,
did
you
ever
complete
the
xml
stuff
for
playing
track,
or
do
you
like
just
have
a
fork
and
like
mostly.
B
Xml
output
is
completely
done.
We
there
are
a
few
bugs
which
are
documented
on
our
repo
and
I've
mentioned
tanner
in
the
issues.
So
we
will
need
to
pr
a
few
things
back
into
clan.
B
I
think
any
clan
shop,
I
think
any
clan
shop
changes
we
do
should
be
peeled
back
into
clan
shop
at
the
earliest
possible
opportunity.
C
B
I
don't
know,
I
don't
really
feel
like
there's
a
lot.
We
will
need
to
add
to
plunger.
I
mean
it
gives
us
most
of
the
information
we
need
and
if
we
need
more
information,
then
we
can
just
add
what
I
call
xml
markers,
because
I
mean
the
xml
that
clang
sharp
outputs,
it's
just
marginally
decorated
c
sharp
code,
but
it's
yeah.
A
B
Did
stuff
we
can
parse
and
it
gives
us
all
the
information
we
need.
Obviously,
clan
shot
will
never
be
considered
with
oh
considered,
with
overloading,
with
renaming
or
anything
like
that.
That
will
be
entirely
so,
in
fact,
that's
all
this
is
which
scraper
will
be
doing
to
be
able
to
be
on
that.
C
I
I
wanted
to
yeah
you
mentioned
that
before
regarding,
like
renaming
and
stuff
like,
is
there
any
reason
why
we're
moving
that
into
like
the
scraper?
Does
the
script
know
anything
there
that
it
can't
like
pass
on,
I
mean
like
naming
is,
is
a
good
thing
to
have
just
isolated.
C
A
B
Could
split
it
up?
I
guess
there's
no
reason
I
can
think
of.
A
B
I
mean
yes,
yes,
it's
a
different
kind
of
scope.
Group,
yes
yeah!
I
don't
know
it
just
depends
how
much
we
want
to
do.
I
mean
we
could
split
up
and
isolate
it,
but
then
we
gotta
come
up
with
a
proposal
for
how
to
configure
that
that
generator
stage
and
all
that
jazz,
which
I
don't
know
if
we
want
to
do.
C
B
A
B
Yeah,
I'm
just
I.
I
agree
that
to
be
fair,
so
yeah,
I
would
say,
make
the
symbol
renaming
its
own
thing.
Oh
I'm
trying
to
think
of
what
else
it
does.
I
mean
to
be
honest
that
all
the
scraper
really
does
is
take
the
information
we
get
from
clemson
and
create
symbols
for
it.
What
it
was
then
going
to
do
is
run
it
through
a
modification
pipeline
which
was
originally
called
xml
mods,
because
this
was
back
before
we
had
the
symbol
system
that
I
designed
this.
B
B
To
separate
them,
to
be
honest,
it's
just
yeah
make
sure
you
don't
document
it
when
you
do
come
to
that
yeah
and,
to
be
honest,
I
think
that
will
give
us
a
cleaner
sense
of
progression
as
well.
I
mean
it
will
when
we
will
be
able
to
see
by
what
stages
we
have
working
where
the
actual
generation
pipeline
is
in
terms
of
progress.
C
Yeah
for
sure
I've
also,
you
you've
linked
me
this
this
this
issue
regarding
the
tests.
Yes,.
B
Yeah,
so
there
are
two,
so
those
were
the
two
issues
I
was
talking
about
on
our
repo
for
clan
shop.
So
at
the
moment
we've
got
a
bug
that
oh.
B
So
we've
got
a
bug
in
clan
shop
upstream,
where
it's
just
spitting
out
pure
c
c-sharp
codes
into
xml,
which
isn't
great,
and
then
we've
also
got
another
issue
open
for
actually
generating
tests
in
xml
and
using
those
tests
to
check
that
everything's,
blittable
and
whatnot.
So
it
has
that
compile-time
validation
that
our
types
are
actually
correct,
which
is
something
that
we
really
could
have
done
within
2.0x.
C
C
B
B
Of
terror
fx
for
tests,
so
I
shall
send
you
a
few
links
there.
B
B
It's
not
going
to
compile
c
code
and
like
it
I
actually
test
the
native
call.
If
we
wanted
that
which,
to
be
honest,
I
think
I
think
would
be
kind
of
cool.
I
mean
that
would
be.
C
I
mean
I
still
dream
of
the
cls
stuff
ported
to
c
or
like
automatically
ported
to
shop.
That
would
be
amazing,
but
yeah,
oh
okay.
So
that's
what
it
like.
Usually.
C
B
C
C
B
Well,
yeah
it
would,
it
really
would
have
been
good,
but
I
think
ultimately,
this
is
just
like
increasing
the
testing.
We
have
against
against
silk
touch
itself,
which
is
needed
because
we
need
this
generator
to
work
properly
and
really
well
for
sure
and
then
later
down
the
line,
we
could
look
at
like
auto,
generating
c
code,
compiling
it
on
the
fly
and
then
p
invoking
into
that
to
like
sort
of
test
the
whole
like
a
full-on
integration
test.
B
But
again
that's
something
I
can
look
at
when
everything's
a
bit
more
established
yeah,
because
I
do
not
have
a
time
I
I
do
not
have
a
lot
of
free
time
and
I
will
sort
of
spend
it
where
my
expertise
is
best
spent.
I
guess
yeah.
C
B
Well,
I
I
jump.
I
jump
around
languages
where
I
work
anyway,
I
mean
I
was
doing
python
full-time
for
like
six
months,
then
I
was
doing
c
and
c
plus
plus
full
time
for
six
months.
Then
I
was
doing
rust
full
time
and
now
I'm
doing
c
full
time
so
I'll
sort
of
jump
around
a
bit.
C
C
B
B
B
Well,
because
I'd
imagine:
okay,
well,
there's
nothing
more
there,
but
yeah
that
I
mean
that's
just
like
generating
a
dispatch
table
from
an
opengl
loader
but
looks
things
which
so
yeah
it's
just.
I
wouldn't
worry
about
the
extension
too
much
I
mean
you
can
just
treat.
C
Yeah
yeah!
Well,
I
just
wasn't
so
I
was
wasn't
so
sure
what
was
like
the
file
names
but
yeah.
B
Oh,
I
just
realized.
Oh
no,
the
background
music
on
the
stream
stopped
that's
unfortunate.
Let's
try
and
get
that
back.
B
A
B
Okay,
well,
I
I
trust
that
there
are
people
in
the
chat
to
complain
if
that's
crap,
but
yeah,
so
I
mean
yeah.
I
think
we've
covered
a
lot
there
is
there
anything
else
on
the
generator
you
wanted
to
discuss
while
we're
here.
I.
C
Don't
think
so
I
mean
I
don't
know,
are
you
like
available
for
today,
like
later
today?
Maybe.
C
B
C
Like
like
today,
there
will
be
like
a
two,
I
think
or
like
one.
I
don't
know
like
a
handful
of
pr's
that
guy
kind
of
blocking,
so
I
was
hoping
to
like
get
those
through
quickly
like
especially
like
the
refactoring
into
into
records
like
that's
like
that
needs
to
go
through
before
I
can
do
anything
else.
Yeah.
B
Yeah,
of
course,
yeah
I'll
be
around
for
reviews,
and
I
suppose,
if
there's
any
like
design
discussions
we
want
to
have,
we
can
just
whip
up
a
stream
like
this
really
impromptu.
I
mean
I'm
very
aware
that
I
don't
want
to
pollute
the
the
youtube
channel,
which
is
why
I've
got
it
set
so
that
when
the
stream
ends,
the
video
goes
unlisted,
but.
B
You'd
hope
so
yeah
yeah,
okay
cool.
I
think
that,
but.
A
B
Yeah,
okay,
cool
cool.
Well,
thank
you
for
that.
I
think
that
was
a
good
catch
up.
I
sort
of
run
through
what
what
we're
thinking
in
terms
of
3.0
and
generator
design
and
yeah
it'll
be
good
to
do
these
every
now
and
then
I
mean
there's
no
fixed
schedule,
but
it's
sort
of
like
a
touch
base
and
yeah
just
make
sure
we're
on
the
same
page
with
the
generator.
I
guess
yeah
cool.
B
Yes,
and
the
same
goes
to
everyone
watching,
thank
you
for
your
time
as
well.
Thank
you
for
joining
us.
Alright,.