►
From YouTube: AsyncAPI SIG meeting 23 (May 12, 2020)
Description
This is the recording for the AsyncAPI Special Interest Group (SIG) meeting #23.
Attendees:
- Fran Mendez
- Łukasz Gornicki
- Michal Hudy
- Paul Taylor
- Vijay Chandras
Moderation:
- Fran Mendez
- Łukasz Gornicki
Agenda & Notes:
https://github.com/asyncapi/asyncapi/issues/378
B
Some
hello
in
a
unusual
time,
it's
8
a.m.
UTC,
which
is
not
that
early
for
for
Europe,
because
it's
I
mean
in
Poland
it's
10
a.m.
so
it's
pretty
fine,
but
for
States
it's
probably
impossible
to
join
so
I
hope
you
will
guys
enjoy
the
recording.
Whoever
is
watching
us
in
few
hours
before
we
jump
into
agenda.
We
have
new
faces,
I
hope
it's,
because
we
enabled
those
people
to
join
us
because
of
the
different
time
zone
of
the
meeting.
So
Paul
Heyman
a.
C
B
Awesome
so
for
for
sure,
for
few
meetings,
we're
gonna
try
out
this
shift
of
T
of
the
time
of
the
meeting
and,
let's
see
if
we
regularly
have
visitors
at
10
a.m.
I'm
super
happy
to
also
meet
at
10
a.m.
instead
of
6
at
6
p.m.
and
the
VGA
you
joined
us
I.
Think
it's
this
for
me.
So
first
time,
ie
is
your
name
in
the
community
yeah.
D
I've
just
joined
in
maybe
two
weeks
ago
into
the
slack
Channel
and
I'm
closely
following
a
sync
API
development
and
there's
a
few
presentations
that
happen
last
month
or
so,
and
one
of
the
solution
designer
for
what
did
the
presentation
took
me
to
have
a
look
about
async
API
and
how
even
driven
architecture
or
how
to
define
standards?
And
things
like
that.
So
yeah
I'm
fairly
new
to
this
yeah
I'm
based
in
metal,
australia,
melbourne
yeah.
B
B
It's
on
I'm
gonna,
I'm
gonna
jump
to
screen
sharing
to
show
the
agenda
for
today
and
let
me
know
whenever
because
I
think
I'm
I'm
having
some
delay
with
the
audio,
so
kind
of
pink
me
somewhere.
Somehow,
if
you
get
me
lost
in
on
the
wire
somewhere,
so
yeah,
let
me
start
sharing
screen
but
front.
You
have
to
get
give
me
back
again
the
the
rights
to
do.
B
B
For
me,
it's
it's
I'm
based
in
Poland,
so
for
me
it's
10
a.m.
yeah.
B
B
B
So
I
would
say:
let's
unless
some
of
someone
from
the
group
is
here
for
a
specific
topic
and
just
want
to
focus
on
this
one
and
and
leave
the
meeting,
then
let
me
know
then
we're
gonna
start
with
with
the
topic
that
is
more
convenient
for
you.
But
if
you're
here
for
the
entire
meeting,
let's
just
go
on
by
one.
B
No
I'm,
fine,
whatever
okay
cool,
so
the
regarding
the
disabled,
disable
hooks,
so
don't
have
today
on
the
meeting
many
contributors
to
the
generator
so
I'm,
just
gonna,
be
pretty
quick
and
yeah
I
hope
you
watch
the
recording
and
and
later
on
onslaught.
Let
me
know
what
you
think
so:
I'm
I'm,
currently
working
on
the
on
a
feature
or
refactoring
of
the
hooks
functionality,
and
just
let
me
show
you
something
one
of
the
at
the
moment.
B
One
of
the
functionalities
we
have
in
the
in
the
CLI
and
in
the
library
is
that
you
can
disable
particular
hook
in
the
generator,
but
it's
not
hook
itself
like
a
single
function
that
will
be
disabled
with
this
flag.
But
it's
about
disabling
a
of
the
hooks
of
a
given
name,
because
before
when
we
were
talking
about
hooks,
we
only
had
one
type,
one
name
of
the
hook
which
was
generate
after
so
these
were
a
set
of
functions
that
you
could
write
on
your
own
for
your
code.
B
Generator
template
and
those
functions
were
always
triggered
by
the
generator
at
the
end
of
the
generation.
What
we
introduced
new
new
hooks
like
generate
before
and
others
already
so
to
rinkeby,
factoring
I
kind
of
noticed
that
there's
a
we
need
to
shift
a
bit
naming
in
the
hooks
so
because
we
have
generated
before
generated
after
a
trust.
So
we
should
rather
call
those
like
typed
type,
and
the
name
is
something
that
we
will.
B
Most
of
us
will
have
to
provide
in
the
hooks
like
for
the
function
names,
so
the
hook
name
will
be
at
the
end
function.
Name-
and
my
question
was
here-
I
think
that's
already
salmon
already
in
the
comment
said
that
he's
also
agreeing
to
do
it
in
a
separate
issue,
but
would
be
awesome
if
later
on,
after
working
the
video,
you
could
share
your
opinion
here.
If
we
should
maybe
deprecated
this
disabled
hook
flag
and
just
have
like
the
deprecated,
the
functionality
of
it.
B
So
instead
of
disabling
the
hole
type
of
the
hook,
maybe
it's
would
be
much
better
free
for
the
author
of
the
templates
to
disable
only
a
specific
hook
by
the
name,
so
only
a
specific
function
instead
of
the
set
of
set
of
functions
or
maybe
we
should
support
both
so
being
able
to
disable
a
given
set
of
functions
or
or
trust
a
particular
function.
You
would
have
just
two
different
flags
for
it.
A
I
will
rename
this
flag
to
become
disabled,
to
become
disabled
books
in
plural
and
and
then
it
will
disable
all
the
hooks
of
a
specific
type
right.
So
it
will
be
disabled,
hooks,
hook,
type
right
and
and
then
created
another
another
option
with
the
same
name,
disabled
hook
with
the
current
name
that
will
take
in
a
hoop
name,
you
know
and
it
will
disable
the
hoop
by
the
name.
You
know
so
we
have
both.
You
know
and
I'm,
not
sure
if
I
explain
myself
correctly,
but
yeah.
B
B
A
B
A
D
B
B
Okay
back
to
the
so
let
me
check
participants
list
because
I
got
a
message
from
from
me
how
from
the
keema
project
that,
if
I'm
able
to
delay
the
topic
of
a
thing,
I
react
at
the
end,
so
I
don't
see
him
in
the
waiting
room.
Yet
so
maybe,
let's
just
skip
to
topic
from
poll
and
then
yeah
we'll
go
back
to
us
and
API
react
component
and
if
how
it's
not
here,
I'm
just
gonna
share
what
what
he
discussed
with
me
so
Paul.
C
Sure
so,
as
you
know,
I'm
always
thinking
about
unhappy
path.
I
really
found
happy.
Guy
I
tell
you,
but
I
wouldn't
have
thought
and
given
to
you,
know
doing
the
consumer
to
know
what
the
dead-letter
Hume
had
been
defined
for
a
particular
topic
or
event.
Obviously,
if
the
consumer
is
trying
and
keep
strict
ordering,
we've
now
put
the
fifo
flag
internally
for
SQS.
B
So
I
know
that
it
was
brought
up
before
so,
but
I'm,
the
first
Fran
you
want
to
jump
in
or
I
should
write
the
answer.
Yeah.
A
A
So
certain
brokers
have
dead
little
kiddos
like
rather
than
queue,
for
instance
for
mqtt
I
mean,
but
maybe
mosquito
doesn't
have
I'm
not
sure
again.
I,
probably
they
do.
But
let's
say
that
it's
a
simple:
it's
a
super
simple
and
QT
broker
and
it
doesn't
have
that
acute
or
dead
letter
topics.
Let's
say.
A
So
in
this
case
the
protocol
is
included
a
and,
and
it
doesn't
have
that
letter
Q
in
the
case
of
RabbitMQ,
is
in
QT.
It
has
matter
Q,
so
we
cannot
base
this
on
the
protocol.
So
then
we
cannot
put
this
in
from
this
information
in
the
bindings
right
again.
I'll
have
to
check,
but
for
sure
this
is
something
that
that.
A
A
C
Well,
it's
kind
of
interesting
right
because,
as
you
say,
not
not
every
protocol
gonna
support
it
even
protocols
which
don't
natively.
So
let's
say
Casca
right.
You've
gonna
depend
on
me,
but
you
can
still
have
another
topic
which
acts
as
a
dead
letter
Q.
So
I
think
that's
the
really
interesting
part,
because
then
you
start
to
enrich
what
the
base
protocol
supports.
Yeah
I
agree:
it's
it
probably
doesn't
fit
in
a
bind.
It
doesn't
fit
in
the
protocol.
I'm
happy
to
look
at
okay,
an
extension
if
that
helps,
but
if
it
doesn't
fit
anymore.
C
A
A
B
From
from
my
site,
I
would
just
ask
that
also
consider
maybe
creating
a
official,
also
issue
in
in
US
and
API
repo,
because,
especially
in
this
topic,
I
know
that
there
were
some
discussions,
but
most
of
the
discussions
about
the
letter,
Q
they're
happening
in
slack,
where
we
yeah
we're
using
the
free
version
of
slack.
So
after
some
time
messages
are
just
deleted,
we
don't
yet
backup
them,
and
here
we
have
also
discussion
zone
so
basically
have
no
record
of
any
previous
discussions
anywhere.
B
So
maybe
also
consider
adding
issue
on
sing
a
p.I
rapport,
because
then
more
people
will
also
notice.
Not
everyone
are
on
slack.
Not
everyone
are
on
our
meetings
and
I
noticed
that
actually,
as
an
API
repos,
followed
by
many
people
and
they
kind
of
engage
in
every
issue
there
so
yeah
take
it
also
into
consideration.
B
A
B
B
A
B
B
So
today's
meeting
is
to
clarify
this
code
ownership
that
it's
now
changed
into
the
approach
that
they
are
not
the
only
code
owners
in
the
project
anymore
and
it's
a
regular
Aston
KPI
project
that
will
be
developed
together
and
me.
How
stage
is
yours
if
you
would
like
to
add
something?
Oh
I
think
you'd
like
to
say,
so.
We
really
think
that
we
should
under
the
project
because
they
listen
to
it.
For
example,
it
didn't
have
that
much
time
to
work
on
that.
B
E
B
And
we,
we
still
will
work
on
that
when
we
really
need
something.
So
like
that,
we
open
the
project,
we
leather-like
to
open
it
for
everyone
and
clearly
open
yeah,
and
it's
important
to
notes
that
you're.
Not
abandoning
it,
because,
like
ma
che,
that
was
core
contributor
he's
still
there
in
the
code
owners
and
was
the
code
code
review
of
the
public
west.