►
From YouTube: GraphiQL Working Group - 2023-07-11
Description
GraphQL is a query language for APIs and a runtime for fulfilling those queries with your existing data. GraphQL provides a complete and understandable description of the data in your API, gives clients the power to ask for exactly what they need and nothing more, makes it easier to evolve APIs over time, and enables powerful developer tools. Get Started Here: https://graphql.org/
A
A
B
B
C
A
B
Yeah
yeah
I'm
excited
about
this
fix
for
this.
This
cursor
jumper
jumping
thing:
let's,
let's
start
there,
maybe
we
can
get
this
in.
C
Oh,
a
bunch
of
users
confirmed
it
for
both
CDN
and
module
side.
So
let
me
release
it.
I
hope.
That's!
Okay!
I
was
a
little
quick
to
pull
the
trigger
on
that,
but
it
just
seemed
like
oh.
C
Yeah
I
there's
enough
user
like
I
got
so
many
heart
reacts
and
positive
reacts
and
people
were
saying
a
word
from
the
module
side
and
from
CDN
side
and
it's
it
only
required
a
change
to
plugins
so
and
a
breaking
change,
but
a
a
positive
breaking
change.
You
know
in
a
way
that
simplifies
things.
It
allows
us
to
get
away
from
the
pattern
of
asking
them
to
invoke
their
own
set
State
and,
and
then
we
we
solve
it
internally
in
the
plugins,
so
it
it
leads
to
a
much
smoother
implementation
for
plug-ins.
C
You
can
look
in
the
changelog
for
like
Explorer
plug-in
if
you're
interested
it
just.
B
C
A
B
Here's
the
only
concern
I
have
is
that
there
was
one
commenter
who
said
that.
Let
me
find
it
here.
B
The
issue
yeah
28.59
says
that
they
disabled
the
plugin,
the
Explorer
plugin,
which
was
you
know
what
we
thought
was
the
issue
and
still
might
be
so
they
said
they
disabled
it
and
then
it's.
They
still
saw
it
happening,
and
so
my
the
reason
that
I
backed
away
from
this
was
because
I
thought
it
had
to
do
with
those
context,
layers
and
and
complication
and
re-rendering
and
all
of
the
memos
and
callbacks
that
are
a
part
of
that.
That
context
stack
so.
C
B
C
Well,
we
had
several
places
where
people
were
seeing
the
issue
and
then
they
installed
it
and
it
fixed
the
issue.
It's
all
like
the
canary
release
and
they
said
yeah
that
does
it
so.
C
I
mean
if
it
doesn't
solve
the
issue,
then
we'll
figure
out
something
else,
but
the
plugins
internally
were
doing
yeah
this
stuff.
With
passing
the
query
to
both
the
plug-in
and
to
graphical,
you
know,
and
yeah
and
I,
think
we
just
want
to
avoid
that
and
avoid
I
think
it's
also
partially
could
be
a
result
of
instantiating.
The
plugins
within
the
component
life
cycle,
I.
Think
like
instantiating,
it's
in
returning
the
function
that
just
passes
the
props
in
I
think
I.
C
Think
this
approach
to
having
them
move
that
instantiation
out
of
the
react
life
cycle
and
only
blending
it
into
the
react
life
cycle
where
needed
should
help
in
most
cases,
and
if
this
solves
the
issue
for
some
people,
then
that's
great.
If
we
see
it
again,
then
yeah,
it
might
be
a
deeper
issue
to
address
yeah
so
but
I
wanted
to
make
sure
that
we
had
something
before
before.
Dimitri
finishes:
The
Monaco,
rewrite
he's
on
holiday
right
now,
but
yeah
yeah.
So
I
think
this
will
make
things
I
think
this.
C
To
recreate
the
issue
in
the
webpack
example,
where
we
install
the
graphical
Explorer
plugin
but
I
think
the
problem
is,
if
there's
other
users
that
are
passing
in
the
query,
value
and
set
state
with
other
implementations,
it
could
be
that
that's
also
causing
the
issue.
So
even
if
you're,
not
using
the
plugins,
just
this
behavior
of
of
control
of
turning
graphical
into
a
controlled
component
seems
to
cause
issues.
It's
always
caused
issues
even
before
the
rewrite
but
I
think
now
because
of
changes
in
the
life
cycle.
B
It's
a
it's
a
great,
it's
a
great
fix
and
it
you're
right
that
it
doesn't.
It
doesn't
solve
the
problem
of
users
being
able
to
pass
both
the
current
query
value
and
the
center
for
the
query
value
around
on
their
own.
Just.
A
C
Yeah
yeah
there
there
used
to
be
like
early
I,
think.
Maybe
I
wrote
this
a
while
ago
or
something,
but
there
used
to
be
a
way
where,
when
you
did
pass
that
it
kind
of
automatically
moved
into
a
controlled
component,
State
you'll
see
this
sometimes
with
some
of
the
editor
come
like
just
like
the
generic
react.
Monaco
react
code,
mirror
editor
components
will
have
this
Behavior.
You
see
it
with
with
other
types
of
complex
components
as
well,
where
there's
an
input
and
an
unchange
Handler.
C
You
know
that
there's
like
a
controlled
State
and
then
an
uncontrolled
State,
and
when
you
put
when
you're
passing
in
both
the
value
and
onset
and
you're
controlling
this
you're
trying
to
control
the
query
State
outside
of
the
component,
then
that's
what
you
would
call
like
a
controlled
component
versus
uncontrolled.
It's
it's
even
it's
just
that
way,
even
with
form
components.
For
example,
you
know
so
like.
If
you
have
a
you
know,
I
I
think
that's
an
oversimplification.
C
How
of
how
controlled
and
uncontrolled
State
works
but
yeah
it's
a
thing
and
it's
a
thing
where
maybe
we
want
to
yeah
like
I
as
I
as
I
recall,
the
on
change
Handler
should
be
used
for
handle
it
like
firing,
an
event
when
the
change
happens
and
not
for
well
I,
guess
that
would
be
part
of
controlling.
B
C
B
B
C
A
C
Pull
some
of
that
logic
out
and
pull
it
and
have
it.
The
other
logic
happened
by
using
hooks
that
that
also
hooked
to
the
editor
event.
So,
instead
of
so,
there
would
just
be
a
simple
way
where
it's
handling
the
query
there
and
then,
if
you're,
using
some
of
these
other
features
that
like
so
like,
for
example,
the
the
whole
logic
of
collecting
the
operation
names
unchanged,
it'd
be
cool
to
pull
that
out
of
the
query
editor
itself.
Somehow
things
like
that
I
think
would
really
help
simplify
it.
C
B
This
change
is
drastically
with
Monaco.
This
is
this
whole
all
of
this
all
of
this
complication,
not
all
of
it.
Much
of
this
complication
goes
away
because
we're
not
we
don't
have
to
track.
B
C
Yeah,
that's
the
Monaco
is
very
based
on
its
event
emitters
and
you
can
handle
the
events
in
multiple
places,
though
the
tricky
thing
with
that
is
dealing
with
race
conditions
and
not
being
able
to
guarantee
the
order
in
which
events,
fire
and
whatnot.
But
there
are
yeah.
C
There
are
some
ways
that
we
can
yeah
cut
down
on
the
complexity
and
some
of
the
you
know
the
loading
of
modules
and
things
like
that
that
we
have
to
do
in
code
number
five,
so
yeah
I'm
hoping
to
cut
down
on
some
of
that
complexity.
Another
thing
that
I
think
is
going
to
help
cut
down
on
complexity
and
graphical
react
is
I
plan
on.
C
Finally,
as
we
had
charted
out
years
ago,
with
graphical
plugins
to
move
the
history,
the
query
history
tool
and
the
doc
Explorer
tool
into
their
own
plugins,
yeah.
A
C
This
will
cut
down
on
the
complexity
of
graphical
react,
I,
think
and
also
allow
us
to
show
some
like
Advanced
plug-in
integration.
C
Examples
you
know
for
people
to
work
from
and
it
might
add
it
might
require
us
to
extend
the
API
a
bit
I.
Don't
think,
though,
like
I
think
we
can
have
our
own,
we
can
move
those
contacts
to
the
plug-in
itself
and
just
use
the
editor
context
separately
like
because
that
context
will
still
be
wrapped
in
the
editor
context.
So
what
that
means
is
we
won't
need
to
like
have
something
where
plugins
can
provide
a
context.
That's
globally
loaded
I.
C
C
So
if
we
need
to
handle
you
know
the
documentation,
click
event,
then
we
can
show
how
Doc
Explorer
does
that
and
then
other
people
can
use
that
to
say:
okay
now
I'm
going
to
make
my
doc
explorer
that
handles
when
you
click
on,
you
know
a
symbol
in
the
the
hover
popover
that
opens
that
doc,
Explorer
pane,
you
know
and
those
kinds
of
tricky
things
that
seem
they
seem
difficult
to
do
with
plugins,
but
because,
when
you
create
a
plug-in,
you
have
already
the
whole
editor
contacts
and
the
operations
you
know,
like
all
the
different
contacts
available,
you
don't
need
to
do
anything.
C
B
C
B
Not
I,
don't
think
that
that's
entirely
true,
maybe
that's
just
based
on
the
way
that
it's
used
now,
but
so,
for
instance,
here's
an
example
right.
The
history
context
is
used
in
the
main
executor
when
you
try
to
run
an
operation
I'm,
trying
I'm
tracking
it
down
right.
C
Now,
yes,
but
here's
the
thing
is
we
can
move
that
out,
because
now
we
can
track
operation
change
using
the
graphical
react
SDK
within
the
history
plug-in
content.
Do
you
see
what
I
mean
so
within
the
plug-in
history?
We
can
just
use
hooks
to
handle
those
events
rather
than
baking
that
logic
into
the
events
itself.
A
C
Will
really
simplify
things
and
it'll
on
it'll
create
more
layers
of
domain
separation,
where
you
know
we're
not
baking
in
the
the
logic
of
history
and
the
logic
that
that
sounds:
fancy
the
logic
of
the
dock
Explorer
into
the
execution
and
scheme
loading.
These
other
things
that
that
need
to
be.
You
know
they
should
be
independent,
you
know,
and
then
it
kind
of
just
it
will
lead
to
a
kind
of
decoupling
where
the
logic
of
the
history
or
doc
Explorer
component
won't
be
in
graphical
SDK
at
all.
C
But
we'll
have
those
be
installed
by
default
with
graphical
package,
I'm,
not
sure
what
we'll
do
for
like
an
override
there,
but
we'll
figure
out
something,
and
that
also
means
we
can
pull
some
of
those
big.
Some
of
those
doc.
Explorer
and
history
explore
specific
props
out
into
the
plug-in
props
and
then
show
people
how
to
instantiate
that
in
a
in
an
easier
way,
perhaps
or.
A
C
B
Now
that
I'm,
looking
at
this
in
a
little
more
detail,
it
looks
like
Thomas
already
has
the
doc,
Explorer
and
history.
It
looks
like
they're
set
up
pretty
well
together,
like
yeah.
C
We
should
be
able
to
expose
like
the
event
we
need
to,
you
know,
add
a
use
effect
hook
that
waits
for
the
you
know
the
you
know,
whatever
I
I
don't
know,
I,
don't
know
the
refactor
that
well,
the
code
base
is
still
a
little
confusing
for
me
at
times,
but
I
will
have
some
way
to
be
able
to
tell
you
know
this
is
executing.
This
is
finished.
C
Executing
Etc
I
have
I
hope
to
get
even
more
advanced
with
this,
because
when
adding
plugins
for
say
like
results
right,
the
results,
pane
I
would
like
I,
and
maybe
this
isn't
doesn't
make
sense
as
a
plug-in.
Maybe
it's.
It
should
be
a
core
use
case,
but
users
have
requested
a
more
Apollo
Studio
like
well
I
shouldn't.
C
Compare
it
to
that,
but,
like
a
more
iterative
result,
friendly
results,
pain
for
like
websocket
subscriptions
right
or
whatever
transport
they're
using
for
subscriptions,
so
we
should
be
able
to
get
like
a
little
pain
that
iterates
examples
of
each
result,
and
you
know
whatever
metadata
came
with
that
payload
Etc.
Another
plug-in
I
want
to
have
is
a
user's
I.
C
Want
users
want
to
have
is
like
a
a
kind
of
network
like
results
or
network
data
page
or
tab,
rather
so
that
when
you,
when
a
result,
when
you're,
when
a
a
operation
is
completed,
you
can
see
the
headers
and
other
additional
data,
even
though
it's
technically
their
your
developer
toolbar.
B
C
C
It
could
be
plug-ins,
so
yeah,
there's
there's
a
lot
of
possibilities,
so
I
want
to
introduce
a
plug-in
type
key
that,
like
an
enum
that
we
can
build
on
and
right
now,
what
we're
we
have
the
toolbar
plugins
so
by
default,
if
you
don't
provide
a
plug-in
type,
it's
a
toolbar
plug-in,
but
otherwise
you
can
provide
other
ones
and
then
there
might
be
like
the
ability
to
do
override.
So
if
it's
like
a
eventually
we'll
have
non-display
oriented
plug-ins,
so
you
know,
maybe
there
won't
be
a
icon
required.
C
For
example,
in
the
you
know,
just
the
or
overload
would
be
the
term
for
that
in
Titan
script.
So
there's
there's
a
few
different
tricks.
You
know
like
how
beat
plugins
have
several
different
I've
been
working
with,
beat
plugins
a
lot
so
they've
been
giving
me
inspiration,
not
a
not
exactly
applicable
but
I
like
the
thing
like
that,
there's
different
types
of
signatures
they
support
and
that
some
handy
things
there,
but
either
way
yeah
and
also
I
figured
out.
This
is
I.
C
Think
what's
going,
and
this
is
from
the
graphco
Berlin
feedback.
We
had
agreco
Berlin
Meetup
and
we
got
to
talk
to
a
lot
of
users
like
me,
and
and
and
Thomas
in
person
got
to
talk
to
a
bunch
of
users
who
you
know
are
using
it
in
a
way
where
it's
facing
commercially
to
their
users
right.
C
C
C
It
in
xjs,
for
example,
there's
some
things
that
I
want
to
test
more,
but
yeah,
so
with
plugins
I
found
that
actually
without
adding
in
a
way
that
actually
reduces
complexity
in
the
repo,
it's
a
I
think
it's
pr3326
or
3328
yeah
I
know
it's
weird
to
actually
be
kind
of
yeah.
C
Three,
three,
two
eight
yeah,
so
here
I'm
I'm,
going
to
propose
creating
this
a
simple,
just,
a
starting
point
of
a
graphical
plug-in,
utils
package
that
provides
just
for
now
a
v
plugin
and
a
shared
TS
config,
and
then
eventually
we
can
have
a
create
graphical
plugin.
We
can
also
the
V
plug-in,
does
quite
a
bit
and
it
just
kind
of
front
load,
some
some
conventions
and
expectations.
C
There's
a
couple
things
I'm
going
to
tweak
from.
What's
in
that
PR
like,
for
example,
we've
decided
with
the
plug-in
plugins
at
020
that
we're
just
going
to
publish
the
UMD
bundle
at
index.umd.js
just
to
make
it
consistent
and
simple.
Instead
of
like
Dash
plug-in
Dash
name,
you
know
whatever,
so
that
that
should
cut
down
on
some
things,
but
the
the
beat
plugin
will
allow
us
to
encapsulate
a
lot
of
conventions
and
will
lead
to
like
a
simple
V
config
file.
C
That's
like
three
lines,
long,
four
lines,
long,
that's
just
like
load,
this
beat
plugin
and
then
you
like
I,
use
a
config
plug-in
approach
where
it
actually
configures
everything
for
you
and
even
is
able
to
load
your
package.json
and
and
add
externals
in
different
cases,
and
things
like
that.
So
it
will
take
care
of
some
of
that
and
what
I
hope
to
do.
I
went
way
down
this
road
and
I'm
sure
you've
been
down.
C
It
too
is
is
to
get
the
beat
Dev
mode
to
actually
load
a
like
a
graphical
instance
with
your
plugin
already
in
it
and
I've
like
oh
yeah.
C
This
is
one
of
the
hardest
parts
of
Mount
veed.
Is
that
it's
very
opinionated
about
this
index
HTML
stuff
and
what
I
don't
want
is
to
have
to
have
this
like
create
graphical,
plug-in
CLI
like
install
this
index.html
that
you
have
to
go
and
manually,
update
or
or
try
to
manually
mess
with,
like
what
I
prefer
to
do,
and
there
is
this
opportunity
we
have.
That
could
be
really
nice
where
there's
this
middleware.
That
allows
you
to
rewrite
parts
of
the
HTML.
C
It's
meant
for
like
text
replacement,
but
I
think
we
could
just
use
that,
to
just
add
everything
we
need
and
your
plugin
in
this
rewrite
script
where
it
just
like
replace
main
and
hat.
You
know
replace
head
with
this
replace
main
with
this,
and
just
have
this
like
like
a
function
that
generates
the
the
preview
output
with
your
plugin
like
already
installed,
with
no
configuration,
that's
the
goal,
I
think
it's
possible,
or
or
maybe
just
like
one
small
stub
file
where
you
instantiate
graphical
with
your
plugin.
C
You
know,
and
we
might
have
to
do
it
just
with
esm.
It
would
be
nice
to
do
it
with
you
UMD,
but,
like
it's
probably
going
to
be
easier
to
do,
as
you
know,
with
v,
because
V
prefers
esm
because
of
es
build
and
how
role
is
configured
so,
but
yeah
I
think
it's
a
good
opportunity
for
us
and
it
also
leads
to
consistent,
build
outputs
and
then
also
what
I
want
to
do
is
have
like
about
a
simple
validation
tool
that
validates
that
your
package.json
follows
the
conventions.
C
So,
if
you're
specifying
exports
and
things
like
that,
that
everything
is
where
it's
supposed
to
be
I'm.
Looking
for
an
existing
library
that
might
already
exist
for
this
that
validates
like,
like
pre-publish,
you
know,
you've
built
the
bundles
for
your
your
scripts
you're
about
to
publish,
let's
make
sure
that
what
you
have
in
your
package.json
exports
matches
what's
actually
there
and
that
you're
not
accidentally
publishing
for
files
that
aren't
there
I'd
like
to
add
this
other
parts
of
the
refocus.
C
We've
had
issues,
you
know
where
we
changed
something
with
the
bundler
and
a
file
gets
renamed.
An
extension
is
renamed
or
whatever,
and
and
yeah
it'd
be
nice
to
have
some
assurances
there.
C
But
at
least
yeah
like
I,
think
this
would
this
would
be
helpful,
I'm,
adding
plug-in
docs
here
too
and
I
hope
to
make
it
I
want
to
ease
the
process
of
people
publishing
their
own
plugins,
because
technically,
you've
been
able
to,
like
users,
have
been
able
to
publish
plugins
outside
of
our
mono
repo
this
whole
time,
but
the
people
at
graphql
Berlin
were
like
how
like
we're.
Not
we
weren't
sure
if
it's
possible.
C
We
assume
that,
because
there's
no,
we
couldn't
find
any
docs
on
plugins,
which
is
like
yeah
docs,
like
there's,
always
missing
docs
for
everything,
with
this
much
complexity.
So
if
we
just
start
with
some
plug-in
docs
and
say
Hey,
you
can
just
create
a
new
repo
and
publish
it.
This
way,
you
can't
use
our
graphical
scope,
but
you
can
publish
anything
that
you
want
and
then
this
is
how
you
would
install
it.
C
This
is
how
you
would
install
your
own,
published,
graphical
plugin,
and,
if
you
want
to
you,
can
call
it
graphical,
Dash,
plug-in
Dash
whatever
on
npm
and
that
might
help
us
find
it
more
easily
if
we
ever
want
to
have
a
plug-in
registry
at
some
point.
But
for
now
we're
not
going
to
be
doing
anything
with
that,
like
users
are
talking
about
the
dynamic
import
idea,
we
talked
about
years
ago
as
well,
and
that
would
be
cool
to
allow
users
to
install
plugins
on
their
own
time.
C
C
You
know,
like
most
of
our
users
need
want
to
have
you
know,
plugins,
pre-configured
and
shipped
based
on
their
Frameworks
needs
or
their
platforms
needs
and
things
that
tie
into
their
framework
or
their
platform,
and
we've
exposed
this
user
settings
dialogue
as
a
way
to
give
users
more
control,
and
we
can
add
to
that.
But
user-defined
plugins
is
a
little
bit
far
down
the
road.
C
In
my
mind,
you
know
we
need
to
be
able
to
do
more,
with
plugins,
create
a
more
consistent,
easy
to
set
up
environment
for
adding
plugins,
make
our
mono
repo
friendlier
for
people
who
want
to
contribute
plugins,
but
also
make
it
clear
to
them
that
they
can
publish
their
own
without
coming
to
our
mono
repo
I.
Think
that's
going
to
be
important
because
I
think
in
our
mono
repo
we're
going
to
get
to
the
point
once
we
it.
C
You
know,
expose
night
stocks
for
this,
that
P
that
we're
going
to
be
getting
like,
I,
think
copy.
You
know
export
or
the
code
export
plug-in,
it's
probably
on
the
on
the
edge
of
what
we
should
be
supporting
and
not
right.
B
C
Right
I've
actually
thought
about
creating
my
own
graphical
Community,
mono
repo,
that's
under
my
name.
That's
not
officially
graphql
Foundation
supported,
that's
just
the
the
user
land
for
everything.
We
won't
officially
support
with
a
big,
clear
disclaimer
of
saying
these.
These
are
not
officially
maintained
things,
but
this
is
this
is
where
you
know,
there's.
C
Yes,
yes,
that
there's
a
community
of
people
that
can
be
invited
to
up,
and
then
that's
where
we
finish
the
playground
parody
I
feel
like
when
it
comes
to
like
create,
you
know,
create
graphical,
like
function
and
server
middlewares
to
automatically
load.
Your
graphical
we've
had
people
who
were
multiple
people
requests
like
a
just
a
plain
Docker
image
for
loading
graphical,
which
feels
silly
to
me,
but
I
mean
eventually
that
could
be
a
more
complex
thing
with
the
back
end.
C
If
people
want
to
do
that
now
that
we
have
the
ability
to
do,
store,
query
history
and
other
things,
asynchronously,
like
the
the
local
storage
adapter,
could
could
be
turned
into
like
a
network
interface.
For
example,
if
we
wanted
there's,
there's
a
bunch
of
options
there,
but
either
way
yeah
it
could
be
its
own
user
land,
and
that's
where,
like
some
of
these
plugins,
could
go
people
don't
want
to
just
publish
their
own
repo,
which
is
another
thing
they
can
do.
C
For
example,
I
could
show
you
know
Gatsby
or
Heroku
or
her
no,
not
Heroku.
All
these
different
platforms,
how
to
you
know,
publish
their
own
graphical
plugins
from
their
repos
that
they
support
you
know
from
their
from
their
own
ecosystem
and.
B
High
graph
is
another
one:
they
have
a
they're
they're
still
on
the
previous
version
of
graphical,
as.
B
Remember-
and
they
have
they
have
an
internal
plug-in
that
Jamie
Barton
was
explaining
this
to
me
recently.
They
have
an
internal
plug-in
that
they're
using
to
share
like
tap
state
or
not
tap
state
they're
on
the
first
version.
So
it's
operate
like
document
editor
state,
so
that
you
can
just
share
that
around
the
team
without.
C
C
That's
a
great
responsibility
for
platforms
to
figure
out
themselves,
but
you
know
like
like
I
and
several
users
have
proposed
what
if
there
was
like
a
reference
implementation,
graphql
server
that
that
could
be
used
to
share
you
know,
State
and
user.
Can
you
know,
persist
user
configuration
across
instances.
B
B
An
easier
version
of
I
think
an
easier
version.
The
easier
version
of
that
initially
would
be
to
write
a
plug-in
that
uses
that
you
authenticate
through
GitHub,
and
then
it
writes
it
to
a
gist
with
some.
B
C
B
Thing
Ricky
one
thing
I
want
to
mention
before
before
I
forget
is
that
we
were
talking
about
making
graphical.
We
were
talking
about
the
complexities
around
beats
index.html
file
and
how
that's
really
just
a
a
pain.
So
there's
a
there's,
a
PR
2925.
That's
in
draft
that
I
filed
last
year,
where
I
wrote
a
v
plugin
that
allows
real-time
transformation
of
that
index.html
file.
C
B
A
C
Yeah
the
method
I
was
planning
on
using
the
I
think
it's
just
called:
transform,
HTML.
C
A
A
C
But
I
I
decided
to
make
it
simple
or
for
now
now
so
in
32
30
is
about
build
output
and
in
Auto
configuration
and
just
some
simple
starting
things.
For
now,
it
sets
up
the
plugins
for
you
and
everything.
So
all
you
need
is
just
this
one
plugin
to
get
started,
but
then
yeah
I
was
hoping
to
do.
This
basically
was
just
to
create
replacement
script
for
to
transform
index.html,
but
I
decided,
I,
swapped
I
switched
contacts
to
the
bug
fix.
C
C
What
I'll
do
is
I'll
have
it
I'll,
maybe
I'll
create
two
different
versions
of
the
plugin
one,
that's
for
outside
consumers,
or
maybe
it
will
have
like
a
config
flag
where,
whether,
if
you're
inside
graphical
mono
repo,
it
would
load
the
local
locally
built
graphical
and
then,
if
you're
outside
the
mono
repo,
it
would
load
the
latest
release
of
graphical
or
it
would
load
it
from
your
Dev
dependencies
or
something
you
know,
and
then
that
way
you
could,
you
could
be
in
a
standalone
repo,
but
have
the
full
graphical
experience
with
just
your
plugin
I,
like
I
like
the
way
you
went
about
this,
though
I
think
the
one
thing
I'll
have
to
figure
out
for
that
outside
experience
is
like,
let's
see
yeah
like
how
do
we
an
external
graphical
plug-in
in
a
standalone
repo
we'll
have
to
figure
out
like?
C
B
Jamie,
Barton
and
I
are
working
on
a
project
outside
of
graph
base
or
I
mean
sort
of
tangentially
related
to
graph
base.
We're
working
on
a
project
that
uses
graphical
and
we're
writing
a
plug-in
for
it.
B
So
that
plug-in
can
be
it
works.
You
know
we're
gonna,
it's
gonna,
take
a
little
while
probably
take
us
where
I
think
we're
kind
of
hoping
to
get
it
ready
before
September
19th
before
the
conference,
but
it'll
be
a
good
opportunity
for
me
to
get
back
into
the
plug-in
I
mean
I
haven't
looked
at,
I
haven't
looked
at
this
for
I'm,
just
looking
at
the
reap
I'm.
Just
looking
at
this
PR
right
now,
I
haven't
looked
at
this
for
a
long
time.
B
It'll
be
a
good
opportunity
to
to,
while
we're
actually
building
a
new
plug-in.
So
this
plug-in
is
to
share
tab
State,
basically
right,
so
everything
that's
awesome.
You
can
just
share
that.
You
can
share
the
state
and
it
just
gives
you
a
URL
that
you
can
hit
and
it'll
load
graphical
for
you
with
the
state
of
the
tab
or
tabs
like
as
as
you
left
it.
B
C
It'll,
just
hopefully
that
can
help
enable
that,
with
this
with
this
plugin
I
just
linked
in
the
chat.
So
so,
if
we,
so
this
is
the
starting
point.
It's
just
a
build
configuration
and
plug
in
Auto
configuration
it
handles
outputting,
your
your
DTS
file
and
everything
and
sets
up
the
whole
build
configuration.
But,
as
you
can
see
in
the
checklist,
my
next
step
was
to
do
this:
transform
HTML
to
use
that
transform
HTML
hook.
C
Similarly,
so
that
I
think
should
make
it
much
easier
for
you
all
to
to
make
this
custom
plugin.
Another
thing
that
came
up
at
graphql
Berlin
was
people
have
have
hit
some
some
like
the
ceiling
of
design
customization
to
some
degree
that
there's
a
feeling
that
the
CSS
variables
aren't
aren't
dynamic
or
aren't
replete
enough
to
be
able
to
override
everything
and.
C
B
Had
this
with
this
was
early
on
after
we
launched
two
I'm
not
even
going
to
bother
trying
to
find
the
issue
or
the
Discord
conversation
but
Patrick
when
he
was
working
on
strawberry
and
he
was
trying
to
integrate
the
new.
This
must
have
been
August
or
September
of
last
year.
He
was
trying
to
get
graphical,
2
working
and
he
just
wanted
to
change
the
background
code.
You
know
he
just
wanted
to
get
rid
of
that
like
that.
C
Look
yeah:
this
is
something
that
we
should
definitely
have
documentation
for
as
well.
So
alongside
plug-ins
I'd
like
to
create
a
theming,
the
MD
file
that
documents
all
the
CSS
variables
and
gives
like
an
example
of
a
custom
theme,
and
then
we
can
use
that
as
the
barometer
of
saying
like
Okay,
let's
override
everything
we
can,
let's
make
it
look
completely
different
for
a
different
design.
C
Hypothetical
Design
This
is
how
I
was
originally
setting
it
up
with
Gatsby,
but
I
didn't
have
a
product
designer
in
time
to
help
me
with
it
was
was
we
were
going
to
have
a
Gatsby
design
and
the
graphical
official
design
and
have
both
at
the
same
like
release
the
graphical
design
and
then
the
Gatsby
one
that
looked
completely
different
or,
as
you
know,
at
least
color
typography
wise
as
possible,
so
yeah
I
think
we
could.
C
This
is
something
that
we
can
just
improve
on
and
we
can
maybe
just
create
a
theme
example
ourselves
to
test
the
limits,
or
maybe
just
work
closely
with
Patrick
too,
because
Patrick's
already
a
co-maintainer
now,
and
so
you
know
just
where
there's
someone
already
has
a
an
entirely
different.
B
C
Sailor
was
this:
is
this
company
from
Rockwell
Poland,
that
was
at
graphco
Berlin,
and
they
have
a
highly
customized
implementation
and
they
were
talking
about
some
of
the
limits
that
they
ran
into
with
CSS
variables
and
they
had
to
use
some
of
their.
They
had
to
do
some
manual
CSS
override
type
things
that
were
tricky.
B
B
Well,
something
I
brought
up
early
on
when
design
was
that
the
design
is
working
on
transparent
layers.
These
are
alpha
layers
that
are
on
top
of
each
other.
So,
even
if
you
wanted
to
change
a
blue
button
to
Red,
you
and
you
say
red-
make
it
red
it
it's
because
of
the
way
that
Alpha
is
used,
it's
not
red,
you
don't
get
that
color
and.
B
Was
suggesting
that
we
drop
this
you
the
use
of
alpha
entirely
because
you're
getting
too
many
shades
you
like
start
layering
these
neutral
out
these
Alpha.
You
stack
these
neutral
Alpha
colors
and
you
end
up
with
infinite
variations,
depending
on
how
many
layers
you're
using
like
a
button,
it's
on
top
of
a
give
I.
C
Remember
this
conversation
now,
yes,
there's
ways
without.
B
Over
without
over
without
grabbing
the
CSS
class
names
and
and
manually
overriding
them,
like
you
suggest,
like
you
said
that
sailor
was
was
likely
doing,
there's
no
way
to
just
toss
in
a
theme
object
with
a
dozen
of
your
own
colors,
like
semantically,
named
to
match
the
variables
there's
no
way
to
override
it
and
have
it
match.
Even
if
you
had
a
figma
file
where
you
were
like
replacing
them,
it
wouldn't
look
the
same
in
the
app
because
of
the
use
of
the
of
the
alpha
color.
So
theming
is
it's
incredibly
okay,
I.
C
What
would
it
be?
I'm
gonna
sound,
really
silly
when
I
try
to
hsl
and
not
hsla
right
like?
Is
that
what
we
do
now
in
CSS
like
so
like
you
would
have
the
option
of
doing
one
or
other
right
that
you
could.
You
could
provide
Alpha
Channel
colors.
If
you
want
or
you
could
just
have
normal,
not
you
know
hsl
style,
colors
or
rgba
or
whatever.
B
I,
think
that
we
should
you
know
one
one
thing
that
we
might
think
about
doing
is
is
just
vastly
simplifying
the
palettes
for
the
dark
and
the
light
theme
like
just
go
like
vers
cell
level,
simple,
like
they're,
no
Alphas,
no,
nothing
and
then
provide
overrides
so
late
is
stellate
uses
these
Alpha,
like
they
do
it
Alpha
like
in
their
own
design,
which
is
why
it
was
done
this
way
last
year
and
and
I
don't
know
other
than
feeding
in.
C
What
I'm
saying
is:
is
there
a
way
that
we
could
make
it
so
that
the
CSS
variables
can
be
overridden
for
normal,
this
normal
colors
without
alphas,
and
but
then
we
still
provide
the
alphas,
but
if
we
just
use
flat,
you
know
non-alpha
channels.
That's
fine!.
B
They're
combined
right
now,
so
if
you
look
at,
let
me
share
my
screen.
C
C
B
B
C
Get
yeah:
this
is
what
I'm
talking
about
that.
We
could
provide
a
background,
color
or,
or
you
know,
bat
you
know
background
like
whatever
variable
and
that,
if
you
want
you
can
provide
an
HS,
hlsa,
color
and
like
when
we
might
still
provide
that
in
our
in
our
implementation
of
the
theme,
but
you
would
be
able
to
just
make
that
you
know
like
a
hex
if
you
want
it
or
whatever
like
without
having.
B
C
B
B
C
If
you
want
to
have
variables
that
are
used
for
your
own
Nuance
HS,
hsla
colors,
you
provide
those
as
variables.
So
you
have
background
color
and
it's
you
know
a
variable.
That
is
a
combination
of
your
own
variables
that
that
you're,
using
for
your
Alpha
Channel
stuff,
but
you
but
the
it's
not
baked
into
the
theme.
Yeah
yeah,
exactly.
B
C
It's
not
baked
into
the
theme
implementation
details
like
this
yeah
I
see
what
you
mean
yeah.
This
is
something
we
should
have
done
a
while
ago
and
it's
something
we
can
fix
and
we
can
document
and
we
can
say
here's
how
to
do
a
very
plain
override
of
a
you
know.
Let's
say
Versailles
Style,
let's
make
graphical,
looks
like
look
like
it's
embedded
in
the
Verso
dashboard,
for
example,
and
we
showed
just.
B
B
It
could
be
as
complicated
or
as
gnarly
as
you
want,
but
you're
you're,
just
saying
like
these
are
the
the
bits
that
that
people
are
going
to
recognize
or
people
are.
People
are
going
to
recognize
that
they
want
to
change
these
specific
Parts.
Why
here's
the
way
that
you
do
that
without
having
to
you
know,
do
the
mental
gymnastics
around
like
how
to
how
to
do
the
the
alpha
stuff
and
that.
C
So
that
means
we
can
pull
this
out
and
then
have
our
own,
whatever
implementation,
whether
there's
Alphas
or
not,
but
we
make
it
so
that
it
there's
just
a
background.
Color
variable
or
you
know,
theme
you
know
like
you
know:
foreground
color,
whatever
you
know
the
different,
whatever
naming
conventions
we
want
I.
You
know:
I've
worked
on
many
Design
Systems
that
have
different.
You
know
conventions,
but
we
can
just
pick
one
set
of
variables.
We're
also
moved
to
move.
C
We've
also
been
moved
to
Radix,
so
it's
possible
that
we
can
build
off
of
some
of
radix's
CSS
variable
goodness
as
well.
If
I
don't
know,
if
you
noticed
that
but
yeah,
that's
that's
something
that
we
might
be
able
to
do
as
well,
but
but
for
the
most
part
we
can
start
with
colors,
and
then
we
can
make
sure
that
that
the
typefaces
there's
a
variable
for
the
type,
the
font
family
right
is
there
and
fonts
typefaces
and
all
like,
because
that
could
be
a
next
step.
C
B
C
A
C
Because
it's
hard,
yeah
and
I
think
this.
This
is
something
we
can
improve
on
yeah
I've
heard
of
some
one
user,
I
remember
saying
that
they
they
would
just
set
the
the
HL,
the
H
SLA,
both
variants
to
the
same
color
to
get
around
it
I'm
like
that,
that's
not
how
it
should
be
yeah
yeah!
We
can
do
better,
you
know
yeah
and
then,
if
we
want
to
keep
the
alpha
channels,
we
can
but
we
Supply
a
whole
variable.
That's
hls
like
for
background
color
variable
it's.
C
If
you
want
to
like
being
like
it's,
it's
flexible
enough
that
you
can
add
that
complexity,
but
you
can
also
just
have
simple
like
you
know
who
who
knows
like
some
of
our
users
are
people
who
make
and
they
just
make
server
Frameworks
in
other
languages,
they're,
not
web
developers,
and
they
maybe
they
just
want
hex
values
or
simple
rgba.
They
don't
they're,
not
interested
in
this
complexity,
but
they
want
us
a
custom
theme
for
their.
C
You
know
server
and
elixir
or
Scala
or
python
or
whatever
you
know,
pythonists
are
more
into
design,
but
you
know
it's:
just
each
Community
has
its
own
needs
and
it
does,
but
most
of
those
communities
don't
need
Alpha
Channel
layers
like
in
their
theme,
so
yeah
I
think
this
is
something
we
can
improve
on
and
then
you
check
out
all
the
documents.
If
we
need
to
I.
B
Think
it's
wishful
thinking
to
expect
that
we
can
offer
like
a
simple
version
and
a
complicated
version.
All
like
complicated
meaning
uses.
Alpha
right
in
in
this
context,
uses
Alpha
as
complicated
at
the
same
time,
and
the
a
better
option
is
to
flip
the
way
that
we're
doing
it
right
now,
where
Alpha
is
the
default
and
you
have
to
like
go
hard
to
override
all
of.
C
B
Yeah,
let
me
finish
the
thought,
the
we
flip
the
way
they
were
doing
it
now
and
we
have
a
much
simpler
way
to
just
Define
a
default
theme
using
semantic
values
like
like
we're
talking
about
just
like
you're
talking.
B
B
The
minority
of
people
that
want
a
much
more
complicated
setup
in
terms
of
the
The
Styling,
it's
those
those
installations,
should
be
grabbing
class
names
and
doing
hard
overrides
and
that
world
should
be.
You
know
what
what
what
we're
finding
people
are
doing
right
now.
Is
they
just
want
to
override
some
like
kind
of
simple
Styles
and
they're
having
to
reach
out
and
grab
class
names
and.
C
C
Are
it
should
be
right?
The
simple
approach,
the
90
approach?
That's
that's!
Really.
What
we
need
to
do
with
this
Mass
user,
Land,
open
source
is,
is
make
it
easy
to
support
the
90
of
use
cases
and
then
the
other
10.
They
can
figure
that
out
and
that's
not
supported,
but
hey.
We've
got
some
class
names
for
you.
If
you
want
to
to
do
that,
you
know.
A
B
C
So
yeah
I
think
I
think
we
can.
We
can
dial
it
in
I.
Think
there's
also
a
way
that
someone
who's
like
if
we
had
these
like
simpler
variables,
where,
instead
of
instead
of
inter
interoperating
the
variables
inside
and
hard
coding
the
hsla,
if
we
just
had
a
static
variable
that
the
background
color
is
or
whatever,
for
example,.
A
C
A
B
My
synthetic
already
we
just
we
just
need
to
build
out
the
even
we
need
to
abstract
that
even
further
I
think
like
at
the
theme
prop
level
yeah,
and
then
we
need
to
make
sure
that
Thomas
and
the
stellate
like
we
need
to
do
all
that
replacement
and
get
it
in
as
a
draft
PR
and
then
make
sure
that
the
only
installation
I
know.
That's
that's
heavily
using
the
like
the
default
Theme
by
choice
is
distillate
is
delayed
because
that's
just
it's,
they
did.
C
But
that
also
means
maybe
a
breaking
change
or
area.
We
could
make
it.
We
could
slate
it
for
4-0,
like
just
here's,
some
breaking
changes
coming
down
the
pipeline,
and
this
actually
brings
up
another
question
or
another
thought
that
I
have
working
with
graphical
react.
More
is
how
do
you
feel
about
graphical
react?
Even
having
Styles
class
names
is
another
issue,
but
just
having
its
own
Styles
versus
all
the
Styles
and
graphical
react.
Do
you
think
that
that
maybe
they
belong
inside
graphical.
C
Because
in
graphical
react,
that's
where
we
specify
all
the
CSS
for
some
of
these
components.
But
what?
If
the
components
like
were
like
Radix
Primitives,
where
there's
they're
totally
unstyled
or
minimized,
reset
styling
and
then
in
graphical?
That's
where
we
have
all
the
more
complex
CSS
Styles.
Thus.
B
C
B
Well,
there's
a
couple
of
things:
number
one
neurotics
is
dead,
Long,
Live
Radix,
and
so
it's
a
problem
that
we're
just
now
getting
into
it.
A
B
Dear
the
second
thing,
and,
more
importantly,
because.
B
B
I
was
wondering
I,
just
wonder
if
I
just
wonder
if
I
mean
there's
plenty
of
activity
in
the
in
the
Radix,
primitive
Primitives
repository
they're,
just
better
Alternatives
that
are
popping
up.
I
mean
this
is
the
this
is
the
ebb
and
the
flow
of
of
UI.
This
is
you.
B
This
but
anyway
that
so
that's
the
first
point.
The
second
point
is
that,
just
to
directly
address
the
question
of
whether
or
not
we
can
pull
Styles
outside
of
graphical
react,
even
if
we
had
a
separate
package
that
was
just
a
core
library
of
of
Primitives
I
think
graphical
react
would
still
need
styles
for
a
lot
of,
like
you
know
the.
B
If
you
know,
if
you
know
like
Atomic
CSS,
like
for
more
of
like
the
molecular
and
organism
components,
there
would
still
need
to
be.
Yes,.
A
B
Also
exports,
all
of
we
use
stitches.
Currently
it
exports
all
of
the
styling
right.
So
everything
comes
out
of
the
UI
Library
wherever
you're
using
it,
whether
you're
using
it
in
the
Pathfinder
package
or
in
our
dashboard.
C
Yeah,
this
is
how
we
originally
did
it
in
2019
well
Lara
from
Facebook,
in
order
to
collaborated
on
a
set
of
design
components
in
a
storybook
that
were
just
primitive,
like
molecular
components
that
had
no
state
or
anything
and
were
designed
to
be
used
to
build
graphical
to
and
or
I
think
yeah
or
one
I.
Think.
Even
this
is
what
we
were
going
for
at
that
point
and
that's
what
we
did
and
we
it
was
all
in
the
same
project,
but
they
were
in
different.
C
There
was
like
a
components
directory
you
might
remember
seeing
this
a
while
ago,
but
it
was
like
it
was.
It
was
set
up
for
that
like
for,
for
that
with
that
convention
and
I
was
kind
of
I.
Just
was
yeah
I
guess
that
was
part
of
the
problem
of
the
call
like
this
just
having
a
graphical
react
bucket
for
everything.
That's
react
because
you
want
to
have
your
your
stock,
your
you
know
molecular
Atomic
components
separated
out
from
your
state
aware.
C
You
know
components
that
are
that
have
hooks
and
everything
tied
in,
and
so
we
got
close
to
that,
but
I
think
there's
still
some
domain
separation.
That
could
happen
so
we
could
have
like
graphical
UI,
you
know
and
then
graphical
react
separately
and
then
graphical
would
compose
those
together,
for
example,
or
react,
components
or
I.
Don't
know
something:
a
graphical
react
to
dash
components
and
graphical
react,
Dash
state,
for
example.
C
That
could
also
reduce
the
complexity
of
graphical
react
and
make
it
easier
to
document
how
to
use
it
and
maybe
I,
but
I
I.
Think
I.
Don't
think
we
need
to
worry
about
that
too
much
because
it
seems
like
with
the
plug-in
API
people
are,
are
happy
in
the
new
design.
B
I
think
you're
getting
the
right
track.
Enhancing
the
plug-in
API
is
like
the
best
thing
that
we
can
do
right
away
and
until
the
level
of
remember
like
early
last
year
and
even
the
year
before,
there's
loads
of
enthusiasm
for
like
I,
don't
use,
react,
I,
wanna,
I,
wanna,
recreate
graphical
with
view
or
angular
or
whatever,
until
that,
until
that
enthusiasm
comes
back
and
people
are
asking
for
that
again,
there's
really
very
little
value
in.
B
A
B
Of
stuff
in
there
that
we
can
just
pull
out
so
I
think
we
can
get
a
twofer
with
this
with.
If
we,
you
know
like
you're,
doing
moving
the
focus
over
to
the
the
plug-in
architecture
and
how
all
that
works.
We're
going
to
get
a
lot
of
General,
cleanup
and
and
better
understanding
of
the
repository.
Just
by
doing
that.
C
You
know
and
and
more
yeah
implicit
design
like
domain
separation
and
I
think
the
same
thing
will
happen
if
we
have
like
a
a
theme
example
like,
for
example,
I'm
thinking
of
making
graphical
webpack
become
like,
like
I
I,
have
it
now
set
up
so
that
it
deploys
with
netlify
again
at
slash
webpack.
So
if
you
go
to
graphical
tests
webpack,
then
you
get
a
you
know
a
plug-in,
consuming
build
and
then
that's
where
we
can
dump
all
the
kitchen
sink
level.
C
Examples:
I'm
thinking,
that's
where
we
could
even
do
a
custom
theme
where
we
say:
here's
a
you
know
a
Windows
XP
theme,
for
example,
but
you
know
you
know
like
old,
when
amp
or
whatever
I
don't
know,
there's
all
there's
things
we
can
do
there
for
sure
and
and
then
that's
where
we
might
find.
You
know
that,
oh
well,
it
seems
like
we
don't
have
a
variable
for
Border
radiuses
or
something,
and
some
people
want
a
more
less.
You
know
bubbly
designed.
C
So
you
know,
there's
there's
all
kinds
of
things
that
will
probably
discover
by
going
down
that
road,
but
we'll
also
learn
just
from
getting
feedback.
Here's
another
thing:
we
could
open
a
discussion
about
the
themes
and
just
just
asking
hey:
does
anyone
have
open
source
themes
that
they
want
to
share
with
us
and
let's,
let's
see
what
you're
having
to
do
with
class
overrides
and
see
how
we
can
make
that
a
little
easier
see
if,
like
we
see
there's
you
know
six
open
source
projects
doing
the
same
override
of
the
same
class?
C
B
C
C
Yeah
they
sailor
helps
maintain
one
of
the
other
python
graphql
implementations,
I,
can't
remember,
which
but
yeah
a
lot
of
good
conversations.
We
had
there
yeah
there's
some
teams
I'm,
not
mentioning
that
I
can't
remember
everyone,
but
yeah.
There
was
a
lot
of
really
cool
people.
There
really
smart
conversations
happening
and
it's
all
and
all
that
so.
C
But
yeah
so
yeah
we
can
get
some
nice
improvements
added
to
the
plug-in
API.
We
can
pull
out
history
and
Doc
Explorer
as
their
own
plugins
that
are
installed
by
default
for
the
classic
setup,
but
maybe
allow
users
to
override
that
so
that
like
if
they
provide
plugins,
it's
only
those
plugins,
for
example,
there's
different
ways
to
do
it,
because
right
now,
I'm
imagining
we
would
just
append
and
we
would
have
our
default
plugins.
C
If
we
were
to
separate
those
out
and
then
it
would
be
a
pen
I
think
we
actually
have
a
prop
now
for
disabling
default.
Tabs
they're
disabling
default
plugins,
but
yeah
I
need
to
look
up
that
because
there's
been
a
lot
of
things
that
have
changed
on
the
graphical
side,
while
I've
been
more
focused
on
other
things
in
LSP,
so
but
yeah
speaking
of
if
anyone's
watching,
to
learn
about
LSP
changes.
C
There's
some
exciting
things
happening
there
as
well
and
I'm,
adding,
but
first
I'm
adding
a
lot
of
QA,
I'm,
adding
and
then
test
for
vs
code,
graphql
and
more
test
coverage
for
the
graphql
LSP,
so
that
we
can
get
ready
to
start
refactoring
and
add
a
lot
more
features
so
to
make
it
more
performant
so
talking
with
the
guild
about
how
to
build
our
own
loaders
in
graphql
config
that
use
vs
codes
built
in
file
system
API
so
that
our
extension
will
work
on
the
web
as
well.
C
Yeah
a
little
shout
out
for
anyone
listening
for
that,
but
yeah
today
seems
to
be
focused
on
graphical
and
plugins,
so
yeah
I
think
with
those
two
things:
better
plug-in
development
experience,
more
plug-in
domain
separation
and
unlocking.
You
know
the
path
to
that's
already
there,
but
making
it
very
clear
that
people
can
ship
their
own
plugins.
C
However,
they
want
as
long
you
know
and
here's
some
conventions
you
can
follow
with
this
toolkit
and
then
you
know
and
like
the
the
transform
HTML
plug-in
and
everything
or
hook
like
our
own
version
of
that
and
yeah.
Speaking
of
if
anyone's
curious.
Well,
I
mean
you
already
linked
to
yours,
but
yeah
transform
index,
HTML,
specifically
yeah,
so
I
think
we'll
be
able
to
do
a
lot.
There
make
sure
the
theme
is
fully
customizable.
C
Maybe
even
show
like
a
simple
like,
maybe
even
in
the
graphical
webpack
examples
show
a
simple
here's
an
example
one-off
plug-in
just
for
the
example.
That's
like
a
schema
printer
or
something
simple,
just
to
show
people,
here's
how
to
add
your
own
plugin
without
adding
it
to
the
graphical
mono
repo
that
you
can
just
you
can
just
instantiate
graphical
and
have
your
additional
plugins
as
other
modules.
In
the
same
workspace
or
repo,
it
says
it
doesn't:
have
you
don't
even
have
to
publish
your
plugin
to
use
it
I
think
Fitness.
C
B
C
And
as
we
expand
on
it,
it
should
remain
as
such.
There
is
yeah
I,
just
I'm
starting.
This
pattern,
like
I,
did
with
graphical
toolkit
as
well
of
adding
just
a
a
docs
folder
to
each
workspace
that
we'll
use
to
generate
a
website
eventually-
and
here
is
where
I've
added
the
plug-in
Docs.
C
Sdk
from
there
add
a
few
other
things,
but
then
beside
that
is
we
could
have
a
theme,
dot
MD
and
that's
where
you
get
a
good,
some
good
detail
on
how
to
do
your
custom
theme
so
yeah
and
a
link
to
say
if
you
find
you're
having
to
override
you
know
using
CSS
classes
for
a
lot
of
things
open
an
issue
here.
That's
tagged
with
graphical
theme,
you
know,
or
whatever
yeah,
but
yeah
cool
yeah.
C
C
A
validator
step
for
pre-publish
to
make
sure
all
your
exports
are
lined
up
and
that
they
have
that
they
follow
the
the
naming
convention
for
plugins
I.
Think
eventually,
maybe
we
could
have
like
an
init
method
and
every
plug-in
module
exports,
an
init
method.
The
only
problem
is
when
you're
doing
the
Imports
or
the
plugins
and
module
mode.
Then
you're
doing
import
init
as
init
graphical
Explorer,
plug-in
yeah.
B
C
Know
for
each
import
so
like
I,
now
I've
what
I've
done
is
I'm
following
a
convention
where
it's,
if
it's
graphical
plugin,
then
the
the
method
to
instantiate
it
is
the
dot.
You
know
so
Explorer
plug-in
code,
exporter,
plugin
yeah,
so
that
so
we
could
even
like
go
so
far
as
to
like,
even
just
like,
with
their
regex
with
the
validator
say.
Does
your
index.dts
have
a
method?
That's
nice.
C
Okay,
cool
but
yeah.
A
B
C
C
C
A
B
A
C
Work
I
was
doing
so
much
yeah
and
he
had
so
much
on
his
plate
and
I
feel
bad
dumping
plug-ins
onto
his
plate
like
Hey
Thomas,
please
we
could
just
do
a
simple
plug-in
API
and
he's
like.
Oh
no
on
top
of
everything
else,
he
was
doing
I
think
it's
amazing
what
he
did,
but
yeah
I
think
we
can
help
simplify
things
and
help
like
come
back
around
and
say:
hey.
C
Let's
make
this
easier,
hey,
let's
make
sure
everyone
knows
like
I,
think
we
just
need
to
like
Market
the
ability
to
ship
graphical
plug-ins
and
then
add
a
couple
more
cases
like
being
able
to
have
tabs
for
results
and
tabs
for
variables.
I,
wanna,
I'm
I.
C
Have
this
plan
of
making
a
Json
schema
form
plug-in
for
variables,
for
example,
that
dynamically
generates
a
Json
schema
form
when
your
variables
change
and
it
sounds
complicated,
but
we're
already
using
Json
schema
for
Monaco
graphql
I,
already
wrote
a
whole
utility
that
generates
a
Json
schema
for
variables
from
a
graphql
operation,
AST
or
string.
So
we
already
have
the
ability
to
do
that.
C
We
just
need
to
pass
that
Json
schema
to
something
else
and
just
have
a
plug-in
type
variables
or
variables
editor,
whatever
we
want
to
call
it
that
lets,
you
add
a
different
mode
of
editing
variables
and
then
we
can
do
others.
I
would
love
to
do
Json
5,
just
so
it's
all
well
known
I
would
love
to
do
this
with
Json
5.
C
But
if
you
look
at
Monaco
Json
and
vs
code
Json
language
service,
you
might
see
why
that's
still
a
little
far
out
they've
shut
down
I'm
not
shut
down,
but
they
basically
said
out
of
scope,
any
discussions
about
Json,
5
and
but
there's
still
a
possibility
of
adding
a
mode
for
Monaco
for
Json
5..
It's
just
that
right
now.
C
The
easiest
thing
to
do
with
Json
fi
or
with
Monaco
Json
is
Microsoft's
own
version
of
cjson,
which
you
might
be
familiar
with
from
TS
config,
where
you
can
have
comments
where
you
can
have
trailing
commas
and
that's
about
it.
It's
like
it's
like
yeah.
If,
okay
I'm,
not
gonna,
make
a
political
joke.
C
Something
yeah,
so
we
can
say
here's
how
to
make
this
work,
but
there's
people
who
want
to
use
actual
Json
5,
who
are
already
using
Json
5
in
their
implementations,
and
while
it's
tempting
to
think
oh
well,
I
could
just
I
could
just
do
Json,
5,
parse
and
then
Json
stringify
and
then
use
it
with
Json
language
tooling.
That
would
be
great,
but
then
you're
everything's
a
few
characters
off,
because
you're,
adding
double
quotes.
C
I've
tried
this
and
all
these
things
so
there's
like
almost
but
really
yeah
cjson
is
really
all
we
have
in
that,
and
it
just
takes
it's
in
the
v
example.
It
just
shows
how
you
just
enable
a
few
features
for
Monaco
Json
and
that
allows
you
to
to
yeah
be
able
to
have
a
more
forgiving
Json
experience,
but
that's
because
of
their
own
Json
Parts
here
and
almost
every
like
you'll
see
as
well.
There's
a
there's.
Now
a
code
mirror
Json
5
for
code
mirror
six.
C
C
Mode
for
Json
5.-
and
this
is
something
we
could
we
could.
We
don't
have
to
use
I
mean
I,
don't
think
it's.
It's
not
Json
schema
ready
yet,
for
example,
so,
but
me
and
Sam
from
Altair
and
stripe
are,
are
working
on
a
Json
schema,
informed
mode
for
for
code,
mirror
six
I've
recently
started
getting
into
code
mirrors
things
I
really
like
it.
I
use
I've
been
using
it
for
projects,
and
it's
nice,
it's
not
as
feature
full,
yet
as
Monaco
it
takes.
C
It
takes
a
lot
to
get
an
Editor
to
be
as
featureful
as
Monaco,
but
it's
super
nice
for,
like
a
lightweight
documentation,
type
implementations
and
things
like
you
know
like
little
inline
editors
for
your
docs
and
sandpack.
Another
shout
out
to
sandpack
is
a
great
example
that
code
sandbox
is
made
of
like
creating
like,
like
basically
an
embeddable
JS
fiddle
and
more
much
more
customizable
than
that
and
technically,
you
can
even
add
other
Editor
runtimes
to
it.
But
sandpack
is
really
cool
for
again,
like
in
documentation
type
API
docs,
where
it's
like.
C
Let's
document
this
query
field
or
this
domain
of
the
application,
and
you
have
like
a
simple
example
or
something
or
it's
like
for
you-
know
tutorial
or
something
where
you're
showing
use
cases.
But
all
you
want
is
simple:
query
variables
results.
C
That's
like
this
as
well.
It's
not
as
flexible
as
sandpack,
but
it's
just
a
simple
little
component
called
mini
graphical
and
I.
Think
it's
been
improved
over
the
years,
but
yeah,
it's
still
just
using
our
code,
mirror
so,
but
yeah
yeah,
there's.
C
The
thing
that
people
use
to
make
code-
screenshots,
do
you
know
that's
a
code
mirror
graphql
consumer
there's,
so
many
I
was
just
I
was
using
as
I
was
reading
their
source
code
to
figure
out
how
to
do
a
a
code
like
a
highlight
code,
highlighting
validation,
snapshot
tool
like
Percy,
but
for
code
highlighting
and
I
got
so
close,
but
our
our
text,
mate
grammars
for
vs
code
graphql
are
just
too
complex
for
any
web
runtime
other
than
vs
code
because
they
do
this
whole
thing.
C
It's
a
whole
conversation,
but
you
were
going
to
have
to
use
like
a
more
simple
snapshotting
tool,
but
we
are
going
to
introduce
code
highlighting
snapshots
for
vs
code,
graphical
Syntax
for
anyone
listening.
So
we
will
have
regression
free
or
as
regression
free
as
possible
snapshot
tests
for
graphql
strings
and
for
graphql
documents
and
for
embedded
documents
so
but
yeah
so
we'll
have.
C
The
repo
will
have
like
examples
in
JavaScript,
typescript,
View
python,
PHP
Scala,
everything
that
we
already
have
syntax
highlighting
for
so
that's
a
whole
cycle,
but
yeah
yeah,
there's
many
Monaco
graphical
users
out
there
too
I'm
starting
to
see
in
stealth
mode
using
it.
You
know
like
in
proprietary
applications,
I
think
but
I
think
just
based
on
the
download
rate
compared
to
our
cm6
graphql.
That
Sam
made
already
has
way
more
downloads
per
month
than
when
I
go.
B
I've
been
I've
been
I've
been
around,
you
know,
we
use
it
in
Pathfinder,
I
I
haven't
talked
to
anyone
else.
It's
like
actively
using
it.
It's
just.
C
Oh
huge
important,
Improvement,
Dimitri
added
they
so
they've.
They
added
new
exports
to
Monaco
graphql
that
allow
that
allow
us
to
not
accidentally
import
all
the
languages,
because
we
do
this
thing
where
we
import
the
Json
language
mode.
You
saw
that.
C
Yeah
so
there's
a
new
there's,
a
new
export
path
that
allows
us
it's
called
like
Ed's
core
or
something
and
it
lets
you
export,
truly
just
core
features
of
the
editor
and
also
we
can
export
just
the
language,
the
Json
language
mode
instead
of
CSS
HTML
and
all
that.
So
not
only
do
we
do
that,
but
he
added
a
test
Assurance.
Where
now
we
have
a
test
that
runs
that
does
aveep
build
and
counts.
The
amount
of
Monaco
modules
that
are
in
the
Dimitri
went
all
out
on
this.
C
We
merged
it,
not
the
mono
like
it's
in
the
Monaco
PR,
but
then
he
backwarded
into
something
on
main
branch
that
just
that
just
fixes
the
the
module
exports.
Oh.
C
C
So
he
he
he
does.
He
builds
the
Monaco
graphql
wheat
example
and
tests
it
so
Monaco
graphical
has
a
test.
That's
just
that.
Where
V
he
has
execca
call
like
play.
Yarn
workspace,
graphical
or
Monaco
graphql
react
Veet
build
and
then
it
like
manually
inspects
the
the
the
modules
to.
C
A
really
cool
job
with
that
so
yeah
I'm
very
happy
that
he
did
that.
So
it's
it
works
quite
nicely.
It
it'll
be
like
if
we
upgrade
Monaco,
editor
or
Dev,
you
know
our
you
know:
Dev
dependency
version,
it
might
change
slightly,
but
it's
a
good
way
of
assuring
that
we
don't
accidentally.
C
You
know,
start
importing
more
stuff
again
that
we
don't
need
and
there's
also
a
recipe
there
where
so
like.
Where
we
call
you
know,
Json
the
we
call
up
the
Json
mode
internally
in
Monaco
graphql
to
set
the
schema
and
configuration.
We
don't
even
have
to
do
that.
I
only
added
that
as
a
shortcut
right,
because
there's
a
there's
another
recipe
I
want
to
show
that's
just
what
we're
doing
internally,
you
can
do
separately.
C
So
you
can
just
do
query
editor
on
file
change,
whatever
the
the
Handler
for
file
change,
events
debounce
it
and
then
you
you
yourself,
import
the
Json
mode
and
call
dysfunction.
It's
like
I
I
added
that
as
an
I
made
it
an
internal
detail,
so
people
would
have
to
set
that
up
and
to
reduce
the
complexity
of
configuring
Monaco
Json,
but
it
only
it
only
is
that
that
behavior
is
only
called
up.
If
you
set
Json
diagnostic
settings,
I
think
is
what
I
called
it.
C
So
if
you
don't
configure
that,
then
you
can,
you
can
spec
you
can
roll
that
behavior
on
your
own
I.
Think
I'm
going
to
add
something
about
this
to
the
readme.
To
show
how
this
works.
It's
hard
to
explain,
just
speaking
like
I'd,
have
to
show
you
the
code
for
it,
but.
B
C
C
Like
you
would
have
to
call
graphical
language
service
utils,
what
did
I
call
it
get
variables,
Json
schema
from
query
or
I
think
it's
just
get
variables,
Json
schema,
and
so
you
would
have
to
call
that
on
the
editor
value.
So
it'd
be
like
on
query,
editor
change,
past
the
the
value
into
that
into
that
utility,
get
the
Json
FEMA
and
then
call
Json
mode,
diagnostic
settings
or
or
whatever,
and
pass
in
the
schema
and.
A
C
C
A
C
Then
that
would
allow
people
to
customize
the
Json
Behavior
as
they
see
fit.
If
they
wanted
to
I
know
it
wasn't
popular
to
mention
even
on
the
Json
5
issue
of
a
mono,
because
Monaco
yamo
is
perfectly
well
and
fine
and
supports
Json
schema,
but
yaml
is
not
as
easy
to
work
with
as
Json
5.
I
understand.
But
if
we
wanted
to
support
other
Json
schema
enabled
situations
we
can
so
I
was
I
was
that's.
C
Why
I'm
like
eager
to
do
this
with
like
say,
react
hook,
form
Json,
SEMA
or
something
like
that,
because
yeah
all
we
have
to
do
is
handle
on
editor
change,
get
that
Json
schema
pass
it
somewhere.
We
can,
it
doesn't
have
to
be
dates
on
it
can
be
whatever
is
Json
schema
aware,
it's
a
it's
a
powerful
thing
to
a
friend
for
us.
It.
C
B
Ricky,
let
me
see
if
I
understand
this
correctly,
so
in
the
Prototype
that
I
built
using
Monica.
This
was
a
lot,
so
this
was
last
year
there
once
you
have
an
instance
of
the
Monaco
graphql
API,
there's
a
method
called
set
diagnostic
settings
and
that's
where
you
can
choose
to
validate
the
variables
Json,
where
you
pass
in
the.
C
C
B
C
So
I
I
made
it
like
and
and
I
brought
this
up,
because
this
is
part
of
why
all
the
languages
got
imported
into
the
Monaco.
Graphql
bundle
like
whenever
you
bundle
Monaco
graphql,
because
we're
doing
import
languages
from
Monaco
Monaco
editor
that
there
wasn't
really
at
the
time
another
way
to
import
the
language,
the
Json
mode
in
ESN
that
didn't
accidentally
import
all
the
other
modes
that
we
definitely
don't
need
not
to
Discount,
HTML,
CSS
or
typescript
modes,
they're
great.
C
But
our
users
don't
need
them,
usually
so,
and
this
overrides
that,
in
a
way
that
even
messes,
with
like
the
bundler
plug-ins
for
Monaco,
like
beaten
Monica,
so
his
his
solution
fix
that
for
now,
so
that
now
it
only
Imports
languages.json,
which
is
all
we
need
internally,
to
call
this
logic
but
I'm
just
pointing
this
out,
because
we
could
also
just
pull
that
logic
out
and
just
say
if
you
want
Json
variables,
completion
just
set
it
up
in
this
way.
Manually
and
I
know
it
sounds
I'm
making
it
sound
complicated.
C
But
it's
it's
like
only
20
lines
of
code.
Once
you
have
a
query,
editor
instance,
it's
it's
just
20
lines
of
code.
You
know
the
debalance
is
like
an
implementation
detail,
although
you
definitely
don't
want
to
call
this
every
on
every
type
event.
But
yeah
like
it's
yeah
I'll
I'll
have
a
I'll,
create
a
Monaco,
graphql
doc,
slash
advanced.md,
that
that
just
has
some
additional
details
for
this
or
something
awesome.
C
C
Also
doc
site
I
think
could
be
coming
sooner
than
later,
so
Dimitri
is
also
a
co-maintainer
of
the
nextra
ecosystem
and
I.
Don't
know
if
you're
familiar
with
nextra,
but
it's
this
nice
very
nice
abstraction
layer
on
top
of
next
JS.
That's
like
gives
you
kind
of
like
a
nice
little
blog
dock
site,
theming
system,
it's
a
nice,
it's
it's!
It
gives
you
more
flexibility
than
like,
say
docosaurus.
C
If
you
want
to
create
a
custom
theme,
for
example
and
I,
think
it's
a
really
exciting
thing
so
yeah
it
looks
like
the
guild
has
been
switching
to
this
for
a
lot
of
their
stuff.
I
switched
to
it
for
my
personal
blog,
although
I
want
to
customize
the
theme
now
but
I'm,
using
a
pre-existing
theme
and
it's
pretty.
C
C
It's
it's
there's
a
it's
a
whole
I
guess
they
just
got
some
gig
with
the
French
government
to
redo,
like
thousands
of
their
Ducks
mates.
Oh
that's
great,
so
I'm
happy
for
Dimitri,
but
this
also
might
mean
they
will
need
to
give
him
some
help
on
the
Monaco
effort,
because
he's
gonna
be
he's,
got
a
lot
of
exciting
things
happening.
B
The
only
thing
left
with
I
had
a
quick
review
of
The
Monaco
effort.
The
last
piece
that
we
don't
have
yet
is
the
connection
to
the
doc
Explorer
and
we're
gonna
have
to
rethink
that
because,
with
the
hovers
in
code
mirror
you
can
put
a
button
in
there
yeah.
C
Well,
we
have
markdown
in
Monaco,
so
we
can
add
links
with
hashes
for
example.
So
let
me
show
you
how
to
do
that,
because
they're
layers
of
abstraction,
it's
possible
actually
I,
think
we'd
have
to
go
all
the
way
down.
Yeah
like.
C
Is
it
is
possible
with
markdown
like
you,
can
override
the
markdown
the
way
we're
doing
markdown
now
for
the
hover
and
stuff?
Is
that
we're
just
showing
like
we
actually
already
are
generating
markdown
on
the
Fly
for
the
symbols
like
the
query
fields
and
stuff?
So
you
know
how
the
query
fields
are
highlighted.
That
is
markdown
that
I
added,
so
all
like
and
I
added
I
think
I,
even
added
something
for
linking
it.
If
you
want,
but
we
can,
we
just
need
to
add
a
wrapper
to
say,
link
this.
C
You
know,
like
a
configuration
that
says
you
know,
link
base,
like
maybe
like
provide
a
function
that
generates
the
hash
for
each
symbol
or
each
field
selection
when
you're
hovering
it
I
know
it
sounds
complicated,
but
you're
going
to
be
amazed
at
how
easy
it
is
to
do,
and
it
is
definitely
possible.
You'll,
find
issues
in
Monaco,
editor
repo,
where
people
just
are
like
this
doesn't
seem
possible,
but
it
is
it
is.
It
is
because
they
have
their
own
markdown
renderer.
A
C
Trying
this
is
all
happening
in
graphql
language,
service,
interface
methods,
so
those
interface
methods
now
have
a
used
markdown
config,
or
something
like
that.
That
then
generates
a
markdown
version
and
we're
using
this
in
vs
code
as
well,
because
vs
code
supports
this,
but
we
would
just
need
to
add
to
the
graphql
language
service
methods
the
ability
to
you
can
see.
You
can
see
it
in
like
the
worker,
where
we're
instantiating
the
methods
and
I
pass
use,
markdown
true
in
some
places,
I
think
and
that
that's
where
we
trigger
the
behavior.
C
B
Yeah,
but
so
okay,
so
I'm
looking
at
it
now,
I
could
be
wrong
about
this,
but
that
markdown
string
there's
so
there's
an
interface
for
that
generates.
Markdown.
It
has
an
option
to
support
HTML,
but
that
HTML
option
only
allows
a
specific
subset
of.
C
Elements,
but
what
I'm
playing
what
I'm
trying
to
point
out
is
we
don't
even
need
HTML,
we
can
just
use,
we
can
just
generate
markdown
with
with
anchor
links
and
we
just
use
the
anchor
links.
B
C
That
would
that
would
then
trigger
some,
a
behavior
that
the
those
when
you
click
that
link
that
we
would
we
could
intercept
the
have
a
listener
for
clicking
those
elements
like
so
as
it
comes
up,
there's
a
we
attach
a
listener
outside
I
know
it
seems
like
a
hacked,
but
it's
totally
possible.
I've
I
spent
a
lot
of
time.
Thinking
about
this
in
2021,
when
I
added
all
the
markdown
support
and
there's
definitely
a
way.
C
Yeah
yeah
I
think
we
would
just
attach
a
listener
and
then
destroy
it
when
the
popover
hides
or
something,
and
we
can
control
that
in
Mark
in
in
Monaco
editor
main
process.
B
That
hash
is
just
effectively
like
the
AST,
the
keys
of
the
tree
as
you
go
down
something.
C
Like
that
yeah
so
like
we're
already
generating
in
markdown
and
highlighting
I,
just
I
just
surround
it
in
graphql
tags.
If
you
can
see
like
it's
like
it's
like
triple
tick,
graphql
tags
that
I
surround
it
with
and
it
looks
beautiful.
But
all
we
need
to
do
is
add
a
link
to
that
and
then
have
it
say
like
yeah
like
let's
say
like
posts
ID
or
something
like
that.
C
You
know
where
it's
like
post
dot
ID
translates
to
post
underscore
ID,
so
we
just
have
a
simple
normalizer
and
then
we
can
make
that
a
configuration
detail
of
saying
when
you're
setting
up
Monaco
graphql,
you
say
enable
you
know,
documentation,
links
and
then
here's
the
function
to
generate
the
documentation,
links
yeah.
C
I
I
need
to
jump
in
on
that
I'm
glad
I
explained
it
here,
because
otherwise
I
would
have
taken
like
30
GitHub
PR
comments
to
explain
this,
but
what
I
can
also
do
is
just
add
it
to
Monaco
graphql
upstream
or
as
a
PR
to
his
PR,
maybe
and
at
least
get
that
part
going
of
adding
that
yeah.
B
C
C
Here's
something
else,
I
want
to
point
out,
that's
going
to
be
very
important
for,
for
this
too
I
think
so.
Change
sets
recommends
for
pre-releases
to
use
a
branch
so,
instead
of
like
they
recommend
to
not
use
your
main
branch
for
for
pre-release
when
you
go
to
yard,
you
know
when
you
call
change
that
enter
pre
or
change,
that
pre-enter
I
think
is
the
command
and
that
allows
you
to
enter
a
pre-release
State.
The
documentation
tells
you
to
not
do
that
against
main
branch.
C
C
I
think
we
can
do
it
now
that
we're
our
change
sets
are
clean
and
then
we
can
just
create
a
branch
right
off
Main
right
now
called
Monaco
and
then
Target
his
his
PR
can
Target
that,
and
then
we
merge
into
that
and
and
then
call
yarn
change,
sets
free
off
that
branch.
And
then
what
we'll
get
is
as
many
as
two
version
packages
PR's
at
one
time.
C
There
will
be
one
for
Main
Branch
if
we
have
hot
fixes
and
other
like
things
for
vs
code,
graphql
and
LSP
that
are
unrelated
and
then
a
special
pre-release
cycle
that
we
can
do
where
the
change
log
will
stack
up
for
against
this
Monaco
branch.
And
then
we
can
do
that'll
be
much
easier
to
issue
a
set
of
pre-releases
and
then,
when
we
merge
it,
the
change
logs
get
merged
in
nicely.
C
That
way,
we
could
still
do
like
a
little
hot
fix
to
graphical
or
whatever
things
that
are
unrelated
to
Monaco
and
then
yeah
and
I
will
help
with
the
merge.
If
it's,
if
there's
any
messiness
I
know
we
don't
like
these
these.
Hopefully
this
will
be
like
a
month
long
at
most
feature
Branch
I
hate
to
use
the
word,
but
let's
call
it
a
pre-release
Branch
yeah,
it's
like
not
calling
it
a
meeting,
we'll
call
it
a
a
gathering.
C
Don't
want
to
use
the
cursed
word
yeah
so,
but
it's
I
think
this
is
a
great
case
for
it,
because
then
it
then
I'm
not
like
hey
I,
want
to
do
this.
Lsp
release
can
I
release
this
pre-release
and,
like
oh,
it's
not
ready.
Yet.
C
You
know
like
this
will
allow
us
to
unbundle
that
a
bit
and
let's
say,
okay,
we're
cutting
pre-releases
and
then
we
can
ask
Lee
to
advertise
it
on
the
graphical,
Twitter
and
elsewhere
to
say:
hey,
there's,
pre-releases
of
graphical
four
or
people
will
see
it
too,
on
npm
and
stuff
that
there's
a
four
pre-release
happening
and
we
can
create
excitement
and
get
people
to
try
it
out
and
start
bringing
up
and
just
and
let
them
know
that
we're
open
to
to
you
know
bugs
maybe
like
like
once
we're
like.
C
We
could
start
in
like
Alpha
and
then
once
we
get
to
Beta
we're
open
to
bug
reports
or
something
we
can
set
up
some
kind
of
a
approach
that
we
want
to
use
with
that.
So
we
can
say,
here's
Alpha
try
it
out,
but
we're
not
even
ready
for
bug
reports.
Now.
Here's
beta
or
here's
release
candidate
now
we're
ready
for
bugs
you
know
and
then
and
then
we
can
solve
a
lot
of
things
before
we
hit
40
stable.
B
Yeah
I
I'm
I'm
really
unsure
exactly
what
the
load
is
gonna
be
like
when
you
know
with
the
Monaco
switch.
Who
knows
how
many
people
are,
you
know
tying
into
the
reads
and
the
rights
to
the
editor
right
now
in
like
interesting
ways
like
we're
talking
about
with
these
plugins,
where
they're
taking
the
the
value
of
the
the
you
know,
the
editor
string,
the
document
string
and
they're
munging
it
and
then
sending
it
back
in
like
who
knows
like
how
that's
gonna
work.
A
B
With
with
Monaco
world,
so
it'll
I
don't
know
what
that
load
is
going
to
be
like,
but
I
I,
clearly
I'm,
hoping
that
it's
it's
minimal
right.
We're
not
I.
C
B
We're
also
doing,
let
me
verify
this,
but
I,
don't
believe
that
the.
C
Fact
that
we've
already
been
testing
it
Monaco
graphql
against
next.js
with
react
means
we've
cut
out
a
lot
of
other
potential
integration
bugs
as
well.
You
know,
and-
and
you
have
that
fix
for
strict
mode-
that
in
fact
now
they
think
about
it.
Ryan
Matthew
had
about
the
same
fix
where
there's
an
is
editor
loaded
state
for
each
editor.
B
Through
this,
we're
not
doing
we're
not
pushing
the
edit
excuse
me
we're
not
pushing
edits
into
the
end
of
the
Monaco.
Editor
instance
right,
we're
just
overriding
it.
A
C
C
That
way
it's
it's
flexible
and
you
could
even
set
on
different
tabs
that
aren't
even
currently
loaded
into
the
editor,
for
example,
or
things
or
things
like
that
and
I.
Think
yeah,
but
I
mean
I,
don't
think
it
makes
a
big
difference
if
the
current
model,
that's
loaded,
you
do
editor
set
value,
it
would
have
the
same
impact
but
yeah
I.
Think
what
I'm
hoping
is
that
what
we
continue
doing?
C
What
we
should
be
doing
is
when
someone
passes
in
when
we
have
like
a
use
effect
hook
on,
say
the
query
string
that,
if
that
changes,
that
we
call
editor
set
value
right,
isn't
that
what
we
should
do,
that's
what
that's?
Basically,
what
we're
doing
with
code
mirror
at
least
what
we
used
to
do
with
code
mirror.
B
C
B
A
B
Push
edit
operations
or
you
have
to
use,
execute
edits
on
the
editor
itself.
C
C
You
have
to
give
it
the
location
of
what
you're,
overwriting
and
everything.
B
Yeah,
but
if
we're
so,
if
we're
only
doing,
we
don't
do
in
graphical
2.0
and
planned
for
three
and
four
we're
not
doing
any
inserts
into
the
editor
from
outside
we're
not
inserting
from
outside
right
there's.
No,
once
the
visual
operation
Builder
hits
that's
the
first
time
we'll
be
trying
to
insert
into
the
editor
right,
because
right
now,
the
editors
you
just
you
have
to
manually
type
into
the
editors
yeah.
B
Push
edit
operations
without
any
complexity,
because
you
can
push
a
single
edit
offer
edit
operation,
that's
the
entire
range.
Yes,.
C
B
C
C
Do
right,
yeah,
like
that's
something
we'll
get
into
because
I'd
like
to
add
like
oh:
what's
it,
what
do
they
use
the
term
that
they
use
for
when
you
get
like
the
little
light
bulbs
in
vs
code,
I
forget
what
the
heck
that's
called,
there's
a
name
that
they
use
in
LSP
language
for
it,
but
code
helpers.
I
can't
remember
either
way
we
should
be
able
to
do
that.
So
that's
my
idea
is
to
replace
mer.
C
B
B
A
C
B
C
C
C
B
Yeah
all
right
great
meeting
we'll
catch
up
next
time
and
see
you
on
Discord
yeah.