►
From YouTube: GraphQL.js Working Group - 2022-04-27
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
B
And
today
is
27th,
so
just
in
case,
if
anybody
let
the
drawing
I'll,
send
it
in
the
chat
yeah.
So,
let's
start
from
from
my
first
point
and
as
usual,
if
you're
watching
this,
like-
I,
I
know
for
sure,
has
already
signed
all
these
papers.
But
if
you're
watching
this
recording
and
want
to
join,
you
need
to
sign
a
membership
agreement.
Participation
guidelines,
the
code
of
conduct.
You
can
find
this
link
in
the
agenda.
You
can
sign
the
sng
video.
B
B
Second
point
is
about
introduction
hi,
my
name
is
one
and
next
to
you.
D
Like
all
these,
in,
like
one
pull
request
like
for
this
year
and
then
it
will
be
easier,
so
at
least
people
can
I'll
I'll
just
do
it
after
this
thing.
So
it's
like
you
know.
B
B
D
B
Yeah,
we
can
do
click
on
so
action
item
from
previous
meeting.
Don't
think
we
have
it
filed,
but
I'm
like
released
version
of
yeah
four
branch
out
update
the
branch.
B
Yeah
good,
we
skipped
official
part,
and
now
we
there
is
a
gender
link
in
the
chat
and
now
we're
discussing
action
items
from
previous
meeting.
B
B
Yes,
I'm
only
and
at
that
point
we
can
like
ask
people
to
test,
because,
like
big
things
that
we
currently
know
of
big
breaking
changes
that
we
expected
is
done,
there
is
like
some
some
other
breaking
changes,
but
it's
not
like.
We
know
for
sure
they're
happening
in
advance.
It's
like
what
we
had
on
on
a
table
doing
and
universally
agreed.
So
I
think
it's
enough
scope
for
alpha
release
other
than
that
yeah
about
a
h.
B
I
figure
out
configuration
with
repo,
so
yeah
and
you
tested
much
like
a
couple
pr.
So
that
thing
is
working
and
you're
in
the
same
group
with
all
other
maintainers
except
yakov.
So
it
should
work
not
only
for
you
but
for.
B
I'm
not
sure
like
like
in
details
how
to
do
it
but
yeah.
I
will
look
into
it
and
like
if
something
happens,
you
can
reach
like
an
time
disappear.
You
can
reach
leo
other
people
from
foundation
and
they
give
you
and
by
maxis
or
I
think,
technical
certain
committees.
There
is
like
10
people
there,
so
somebody
will
respond.
B
To
remove
this
one
yeah
yeah,
just
like
quick
update
on
streaming
different,
I'm
actually
like
previously,
I
was
a
bottleneck,
but
right
now
there
is
like
some
discussion,
big
discussion
format,
so
I'm
not
sure
like
what's
the
status
of
it,
I
need
to
to
speak
with
rob,
but
definitely
a
rope
prepared,
like
I
think,
two
two
separate
pr's
like
we
have
a
stack
of
three
pairs
and
two
of
them
like
not
related
to
stream
different
anyway.
So
I
will
it's
like
one
video
and
I
will
work
on
merging
them
first
and
how.
B
C
B
Result
for
one
it's
good
way
to
experiment,
with
instance
of
like
with
different
things,
about
instance,
of
like
process
of,
for
instance,
so
and
other
things
like
ideally
the
same
thing
we
discussed
like
webpack
and
yeah.
B
I
remember
about
it,
but
I
like
it's
down
in
my
pipeline,
it's
to
create
the
test
case
for
webpack
yeah,
this,
like
instance,
of
but
yeah
in
similar
idea.
Ideas
is
a
bright
explanation
or
the
right
test
and
experiment
like
give
up
people
opportunity
to
create
create
a
solution
that
ticks
all
the
boxes.
B
I'm
like
not
sure
if
it's
exist,
but
since
it's
causing
so
much
pain
like
it
should
be
like
combination
of
tests
and
explanation,
oh
and
like
written
requirements
or
on
what
working
solution
should
have
here.
So
no
progress
on
that
at
the
moment.
But
right
now
it's
like
a
right
stage
to
address
that
and
when
what.
B
B
Yeah
yeah
and
scope
yeah,
I
will
put
them
you
know
to
to
take
it
in
the
bible
and
halfway,
something
that
we
have
to
something
that
we
can
discuss
like,
for
example,
set
of
requirements
or
like
tests
or
and
if
we
all
agree
it's
like
requirement
and
maybe
a
potential
solution
yeah.
So
since
we
have
short
agenda,
I
propose
to
do
to
exit.
B
B
I
think
I
think
it
was
addressed
by
list
somewhere
and
since
I
just
don't
want
to
like
reiterate
arguments,
I
need
to
find
it.
I
I
did
a
quick
search
but
didn't
find
it,
but
like
idea
and.
B
D
B
Graphical
you,
you
can
make
a
point
that
graphical
can
also
have
a
context
like
through
settings
or
through,
like
browser,
for
example,
can
choose
figure
out
why
graphical
can
figure
out
the
language
for
something
and
pass
it
as
a
context
but
yeah?
I
might.
B
So
in
a
sense
you
should
not
show
errors
to
end
users,
it's
more
like
programming
errors
or
exception,
and
you
should
not
show
exceptions
or
programming
errors
to
end
users,
so
I
think
yeah.
We
need
to
clarify
to
that
person
if,
if
he
want
to
like
to
show
errors
from
graphql
js
to
end
users,
it's
like
even
if
right
after
some
discussion,
we
had
contact
says
still
it's
not
the
it's,
not
things
that
we
want
to
do
or
designed
to
do
it.
It
will
require
like
a
bunch
of
things.
B
D
Yeah,
it's
too
big
like
like.
If
we
have
to
provide
context,
that's
too
big
of
a
change.
I
and
I
don't
think
it's
like,
like.
I
don't
feel
like
it's
necessary,
but
I
like,
I
don't
know
because
they
have
like
they
presented
a
use
case.
So
I
can't
be
like
I
mean
we
can
present
them
like
a
better
solution
to
their
use
case.
B
Figure
out
like
like
big
big
thing
about
that,
if
use
case
yeah
yeah,
we
need
to
ask
about
that,
but
if
use
case
is
provide
errors
to
end
user
not
like
to
programmer
but
to
end
user
with
use
cases
out
of
scope,
it's
like
it's,
it's
not
what
graphql
was
designed
to
do
and
the
recommendation
it's
like
it's
not
written
somewhere
of
in
somewhere
official,
but
idea
like.
If
you
have
validation
errors
you
want
to
show
to
user.
You
need
to
do
it
as
like
union
on
response.
D
B
Scours
can
expect
what
people
put
in,
like
regular
in
the
relaxed
validation
in
scholars
and
using
the
same
regress
in
user
message,
even
if
this
person,
like
sure
about
his
own
errors,
it's
like
scour
wipers
developed
by
people
they're
not
designed
to
to
show
like
client
type
errors
like
they're,
not
provide
good
ui
ux.
The
idea
is
to
provide
good
the
dates,
so
our
errors
is
like
for
developers
or
for
operation
personnel
to
find
the
networks,
and
things
like
that
yeah.
C
B
It's
not
what
we
going
to
do
if
use
case.
It's
like
developers
the
way
without
english
language,
knowledge
it
it's
it's
like
yeah,
but
we
need
to
ask
this
person
first.
I
think.
B
Yeah
about
validation,
wow
so
about
first
topic,
any
about
passing
context
in
scholars.
B
And
by
the
way,
I'm
totally
happy.
If
anybody
like
answer,
I
quit
or
see
more
issues
from
from,
like
even
a
personal
standpoint,
it
will
create
community
discussion.
So
if,
if
nothing,
nobody
adds
anything
we
can
switch
to.
Second
one.
B
D
B
So
yeah,
I
need
to
review
it
so
yeah,
thanks
for
bringing
with
that,
like
I'm
yeah.
So
what
what
I
will
write?
Oh
that
person
basically
question
here,
is
like
it's
a
champion
is
required.
It's
a
lot
of
effort
on
both
sides,
crafty
address
and
spec,
and
it's
like
it
should
be
one
person
like
handling
with.
So
if
this
person
wants
to
to
be
champion
for
for
awhile
in
spike
and
graphql
jazz,
it's
not
necessary.
B
One
thing
I
don't
want
to
do
is
like
to
to
just
change
graphql
js
and
not
too
much.
I
think
there
is
like
some
valuable
things
in
graphql.
Jazz
can
be
backported
to
spack.
B
So
that's
that's
why
I
think
it's
like
changes
from
both
sides.
It's
not
like
one
or
two
and
yeah.
It's
a
lot
of
work,
oh
wow,
but
I'm
surprised
yeah,
it's
actually
the
thing
I
I
I
I'd
ask
for
so
yeah.
I
will
ask
this
person:
okay,
yeah
fine
thanks
about,
but
I
think
I
personally
know
him
so
yeah.
B
B
And
I
think
biggest
one,
it's
number
seven,
it's
about
schema
first
approach
to
graphql
js
page
can.
Can
you
provide
the
context
on
on
what.
D
So
having
some,
because
because
I
mean
this
kind
of
approach
with
like
stl
for
us
to
bc,
a
lot
like
many
people
are
using
it
like-
and
this
is
like
this
is
like
one
of
the
like.
If
you
want
to
do,
sdl
graphical
tools
is
one
of
your
biggest
dependency
and
like
this
is
like
a
small
function
which
we
can
provide
as
a
utility
for
people
who
are
trying
to
like
we
can
just
port
over
graphql
tools,
utility
for
adding
resolvers
to
schema
into
graphql
js
itself.
D
B
That's
a
couple
qualification
like
it
says,
deal
first
because
when
you
say
schema
first,
like
any
graphql.
B
So
yeah,
I
think
it's
I
need
to
one
one
thing
I
need
to
do
just
in
case
we're
missing.
Something
is
just
clean
because
leave
a
long
time
here,
post
what,
but
I
don't
remember
clearly
reading
his
motivation
like
yeah.
D
D
B
D
B
Yeah
a
couple
like
historia
trivia,
graphql
tools
was
created
by
paul,
and
I
think
it
was
it.
It
has.
The
same
name
has
the
same
name,
graphql
tools,
and
one
of
the
reason
why
it
was
created
is
exactly
wet.
B
B
B
F
Can
I
can
I
just
jump
in
with
one
quick
thing
yeah,
so
I'm
not
sure
exactly
what
the
default
implementation
is
like.
What
the
name
of
the
function
is
now,
but
from
what
I
recall.
Graphql
tools
originally
added
the
resolvers
in
place,
meaning
it
mutated.
F
The
original
schema,
I'm
not
sure
if
that
was
lee's
original
hesitation,
eventually
as
part
of
a
reorganization
of
code,
there's
a
new
way
of
doing
it
that
uses
the
map
schema
function
that
creates
a
new
schema
and
and
and
the
the
graphql
tools
map
schema
function.
F
That
requires
everything
similar
to
like
the
sort
lexographic
schema
does.
But
the
map
schema
is
like
a
generic
version
and
it
uses
that
to
create
a
new
schema
that
doesn't
mutate
the
original
schema,
and
that
was
originally
thought.
I
think,
if
I'm
remembering
that,
maybe
that
would
be
more
palatable
to
you
know
to
integrate
upstream,
but
then
it
turns
out
that
it's
actually,
quite
you
know,
expensive
or
slow
to
you
know
recreate
the
entire
original
schema.
F
When
all
you
really
want
to
do
is,
you
know,
add
the
resolvers
to
some
extent
it's
a
special
case,
you're
not
really
mutating.
The
original
schema,
you're
sort
of
finishing
finishing
it
off
to
you,
know
finishing
the
creation
of
a
now
an
executable
schema.
So
I'm
not
sure
what
the
default
is
now
we
then
we
added
an
option,
basically
to
you
know
to
mutate.
The
original
schema
like
it's
like
an
ad
resolvers
in
place,
option
or
something
like
that.
F
So
when
you
know
when
we
talk
about,
I
guess
what
we
might
want
to
have
upstream.
I
I'm
pretty
sure,
but
I
do
want
to
ask
which,
which
version
we
would
want.
I'm
pretty
sure
we
would
want
the
quicker
one,
the
special
case,
one
where
it
does
mutate
the
original
schema,
because
that
works
better.
That's
what
people
really
want,
but
it
does.
It
does
carry
that
wrinkle
that
it's
to
some
extent
finishing
off
or
mutating
the
original
schema.
F
That's
one
small
point
and
then
another
small
point
about
terminology,
I'm
going
to
say
for
a
second,
because
I
want
to
hear
what
you
guys
have
to
say
about
that.
B
F
Yeah
yeah
there's
a
second
thing
that
I
wanted
to
add,
but
I
I
think
we
should
talk
when
we
when
we're
asking
lee,
for
example,
you
know
ma,
you
know.
Why
exactly
did
you
want
like
this?
You
know.
Is
it
now
that
sdl
I
mean?
Maybe
we
want
to
focus
a
little
bit
on
which
version,
because
there's
that
wrinkle
about
whether
we
do
we.
C
F
B
B
Because
in
if
we
have
it,
it's
part
of
build
schema,
it's
mean
it's
like
schema
created,
oh
build
an
extent,
schema
schema,
created,
stuff
assigned
it's
the
same
time.
It's
nothing
is
mutated
or
versus.
We
should
have
like
separate
function
to
create
new
schema
with
resolvers
attached.
F
Okay,
I
mean
that
sound.
That
sounds
good,
I
think
you
know
integrating
to
build
schema
would
accomplish
everything
that
everything
that
you
know
people
need.
I
don't
know
if
you
have
a
thought.
D
I
think
yeah
putting
like
yeah,
I
think,
build
schema,
makes
sense
more
than
mutation
mutating,
one
that
I
can
discuss
internally
with
others
think,
and
I
can
get
back
on
that.
One.
F
I
think
you
know
I
didn't
experience
the
problem
directly.
I
can
try
to
search
through
the
archives,
but
definitely
people
at
the
the
guild
have
have.
I
think
arda
in
particular,
I
think,
has
might
might
have
some
actual
times
on
it,
but
I
think
it
wasn't.
It
wasn't
a
theoretical
concern.
I
mean.
B
B
Maybe
it's
not
big
issue
per
se,
but
in
general
design
is
to
separate
with
schema
things
into
separate
functions
like
because,
for
example,
I
don't
know
if
it's
a
issue
or
not,
but
if
you
build
schema
from
introspection,
you
cannot
attach
resolvers.
If
we
add
like
to
build
schema
at
the
extent
schema,
it's
mean,
like
you,.
B
You
create
schema
from
introspection
and
cannot
attach
resources.
Probably
it's
not
something
that
people
want,
but
ideally
we
have
separate
function
like
f
for
a
case
where
you
create
schema.
However,
you
want
and
when
just
want
to
attach
resolvers,
but
if
there
is
like
performance
like
real
performance
issues,
when
we
can
figure
out
like
yeah.
So
one
thing
I
want
to
be
sure
that
you
saw
like
real
thing
with
performance,
because
for
me
it's
like
startup
time
thing.
B
F
F
F
But
but
he
sent
a
pr
to
to
graph
the
old
tools
saying
that
one
of
the
most
expensive
parts
of
schema
creation
is
the
validation
and
since
then,
you
know
when
there
are
options
to
disable
that
and
graphical
gis,
and
he
said
you
know
we
should
pass
those
through.
So
I
I
think
people
who
are
doing
are
worried
about
performance
are
also
skipping.
The
validation
to
make
a
long
story
short.
B
B
I
need
to
think
about
this,
but
yeah
so
two
options
that
we
have
right
now
like
build,
schema
or
separate
function.
F
B
F
The
other
thing
was
to
preempt
your
question
about
nomenclature.
Basically
everything
that
in
the
ad
resolves
to
schema
function,
it's
not
really
resolvers,
as
you
sort
of
pointed
out,
it's
anything
that
is
not
in
the
sdl,
so
means
any
function.
Whether
it
involves
you
know,
scalar,
parsing,
serializing
and
all
that
stuff
is
doable
by
address,
alvarez's
schema
and
unfortunately
the
function,
I
guess
is
purely
is
poorly
named.
I
guess
or
started
off
with
you
know
resolvers,
and
it
really
includes
like
any
function
which
field
resolvers,
hypersolvers
serialize
parts.
F
You
know
whatever
there
is.
So
you
know
that's
just
a
question
of
how
we
might
name
it.
You
know,
but
that's
might
might
be
what
you
were
alluding
to.
B
Instance
of
this
type
of
this
type
of
also
and
the
result
type,
and
I
think
it's
used
like
door
sign
or
something
like
that.
So
I'm
like
one
one
thing
since
its
reference
implementation.
B
Maybe
we
need
to
explore
like
new
format
if
you
want
to
edit,
I'm
not
remember
exactly
like
what
what
was
my
problem
with
like
a
format
but.
F
The
main
the
main
format
is
that
for
objects
and
interfaces
which
then
copies
the
resolvers
into
the
objects,
which
is
another
optimization
there's.
Those
are
a
map
with
the
type
name,
the
field
name,
and
then
you
also
have
the
option.
F
However,
to
with,
I
think,
it's
within
the
fields
and
you
can
override
any
property
of
the
object,
because
again
it
mutates
everything
in
place
and
you
can
use
a
double
underscore
before
that.
Not
a
dollar
sign,
but
you
know
something
similar
and
then
you.
F
Yeah,
so
you
can,
you
can
modify
anything
you
know
so
you
might
want.
You
might
want
to
change
the
format
and
just
so
the
default
is.
If
you
don't
use
anything
like
for
fields,
you
don't
like
name
the
property
you're
defining
it.
It
assumes
that
it's
the
resolve,
property
and
you're,
giving
a
resolver,
but
you
could
change.
B
B
You
can
do
like
typescript
unions,
but
it's
like
yeah,
and
if,
if
we
choose
a
double
underscore,
it
can
be
problematic
for
introspection
fields.
If,
if
people
want
a
tattoo,
I
don't
know
why
but
yeah
I'm
I'm
yeah.
So
what
I'm
saying
is
I'm
like
she
value
for
his
function
and
I
think
we
should
add
it,
but
but
I'm
like
against
copying
copying
what
graphql
tools
do
right
now
without
design
discussion,
so
yeah.
F
Actually,
just
to
correct
the
record,
because
I'm
realizing,
I
made
a
mistake,
although
I'm
not
disappearing,
what
would
you
say,
but
the
double
underscore
was
the
original
version
way
back
in,
like
I
don't
know,
version
three,
I
guess
of
of
graphql
tools
and
then
they
switch.
They
still
support
that,
but
now
now
you're
encouraged
to
pick
the
property
names
based
on
the
the
config
object
for
the
you
know
the
scalar
or
whatnot
or
the
field,
and
those
are
the
white
and
that's
the
better
way
of
doing
it
so
yeah.
B
Yeah
and
another
thing
I
just
remember:
if
we
decide
to
put
this
object
into
build
schema,
we
need
to
also
implement
other
make
executable
schema
type.
Behavior
like,
for
example,
allow
you
to
pass
a
couple
of
documents,
because
if
resolvers
are
assigned
inside
the
function-
and
you
can
pass
only
one
document
and
you're
doing
like
validation
and
you
need
to
do
validation,
because
people
can
mistype
type
name
or
something
you
need
to
to
create
error.
B
But
I'm
like
something
yeah,
yeah,
I'm
and
to
to
to
be
sure
like
I'm,
for
make
executable
schema
experience.
Ideally,
so,
since
it's
reference
implementation,
people
use
hdl,
people
should
be
able
to
it's
like
long
standing
issue.
People
arrest
this
issue
like
three
or
four
years
ago.
B
It's,
not
it's
big
big,
a
lot
of
work
for
like
small
result,
but
if
you
start
with
saying
basically
like
you
can
build
the
reference
implementation
support,
building,
sdl
schemas
from
sdl
without
hacks
executable
schemas
for
hdl
without
hacks.
I
think
it's
like
a
lot
of
work,
but
the
result
is,
is
really
a
pace
for
web
work.
D
D
B
No,
I
mean
like
using
fs3
file.
Example
can
be
not
specific.
Okay,
it's
like
everybody
mowers,
can
understand
what
fs.35
scene.
E
B
It's
just
like
in
case,
like
imagine,
people
asked
about
this
for
like
years
and
we
release
with
thing,
but
it
work
works
only
through
one
file
or
you
need
to
manually
can
cut
files
we
have
equals
function,
can
cut
iced
tea,
but
yeah.
It's
probably
not
what
people
want
people
probably
either
want
manually,
open
this
file
or
use
like
some
order,
or
some
like
third-party
thing
to
to
to
get
array
of
strings
so
yeah.
B
C
B
And
also
ideally,
study
what
make
executable
schema
doing,
because
I
think
basically
like
schema
that
functions
that
you
refer
to
it's
like
specialized
one
but
yeah
I
from
top
of
my
head.
We
need
to
do
validation,
so
if
person
provide
resolver
for
non-existing
type
or
know
an
existing
field,
we
need
to
error.
B
B
If
now
we
can
switch
to
like
now,
we
have
two
options:
one
we
can
finish
or
if
jakov
or
alex
or
use
the
h
have
any
other.
C
C
B
Yeah,
it's
I'm
probably,
but
do
you
refer
white
pr
with
just
a
syntax
stuff
that
we
discussed
like
just
like
parsing
and
printing
and
that's
all
yeah
so
yeah?
If,
if
if
what
without
the
execution
yeah,
I
found
it.
C
So
yeah
and
it's
yeah,
there's
one
with
execution
and
one
without
execution,
and
the
one
without
execution
is
just
the
stuff
that
you
asked
for.
The
apollo
release
is
relying
on
the
one
with
execution,
because
there's
one
function
in
there
that
it
needs,
but
I
can
move
that
function
over
to
the
one
without
execution.
If
yeah.
C
Character
so
there's
we
need
the
one
there's
one
that
takes
into
account
the
schema
and
the
operation
and
figures
out
what
types,
what
the
return.
B
Types
of
things,
yeah
yeah:
we
need
what
yeah.
So
I'm
like
what
is
straightforward
thing.
Basically,
since
you
probably
know
better
than
me,
I
actually
missed
like
one
working
group
this
year.
I
think
about
discussion
about
executing
behavior,
I'm
like
I'm,
I'm
out
of
scope
right,
but
one
thing
I
understand
for
sure:
it's
like
it's
not
cause
topic,
so
I'm
like,
but
syntax.
B
I
think
I
think
more
was
like
discussed
so
yeah
I
will
have
on
my
part
and
size
is
reasonable.
It's
it's
140
lines,
so
I
I
will
review
it.
B
Actually.
I
partly
reviewed
it
yeah,
it's
it's
I'm
bottleneck
forward
and
since
we
start
alpha,
think
it's
even
less
issue,
so
we
can
release
that
stuff
and
the
alpha
tag
and
like
if
something
discovered
later
we
can
measure.
B
B
B
Like
validation,
rules
that
don't
touch
execution,
semantics
yeah
feel
free
to
move
it
or
we
can
do
it
in
stages.
You
can
add
more
and
more,
I'm
okay,
with
both
way.
If
it's
1000
wine,
it's
it's
reviewable
to
assets
mostly
test
cases,
and
it's
mostly
like
you
have
snapshots
in
your
test
cases.
It's
probably
the
major
part
yeah.
So
on
this
one,
I'm
bottleneck.
B
I
will
put
an
analysis
for
this
week
to
review
it,
but
yeah
you're
not
disappearing
right.
So,
if
we
discuss
discover
some
issue
afterwards,
I
will
ping
you
and
ask
to
fix
it
or
change
it.
Yeah.
C
For
sure,
okay,
yep,
okay,
cool
I'll
get
that
fixed
up
today
and
and
yeah
sounds
good.
Thank
you.
B
No,
if
no,
it
was
a
very
productive
meeting.
We
discussed
like
butcher
things,
especially
like
with
sdl
attached
resources
and
coin
control,
no
ability
yeah,
it's
so
happy
to
see
next
week,
and
hopefully
we
discuss
also
discuss
batch
of
issues
offline
and
we'll
have
progress
on
all
the
topics.
Yeah,
bye,
great.