►
From YouTube: AsyncAPI SIG meeting 17 (Feb 18, 2020)
Description
This is the recording for the AsyncAPI Special Interest Group (SIG) meeting #17.
Attendees:
- Fran Méndez
- Francesco Nobilia
- Łukasz Gornicki
- Michal Hudy
Moderation:
- Łukasz Gornicki
Agenda & Notes:
https://github.com/asyncapi/asyncapi/issues/310
Chat Dump:
00:12:42 Fran Méndez: template for commits?
00:15:59 Francesco Nobilia: https://thoughtbot.com/blog/better-commit-messages-with-a-gitmessage-template
00:21:13 Fran Méndez: We should probably adopt the best practices for each of the communities instead of trying to come up with a new approach.
00:21:56 Fran Méndez: :)
00:22:59 Fran Méndez: Don't have an opinion on which one we should use.
00:23:32 Fran Méndez: Gratis!
00:27:21 Fran Méndez: My only concern with Github actions is that we'd probably be betting too much on GitHub and we could get vendor locked in the future. Not a big concern but worth thinking about it.
00:28:59 Fran Méndez: Yup, not a big problem.
00:31:06 Fran Méndez: Admins can still bypass this limitation, right? :troll:
00:31:34 Fran Méndez: I'd disable it
00:31:36 Fran Méndez: hahaha
00:31:42 Fran Méndez: kidding
00:33:24 Fran Méndez: It's actually better
00:33:38 Fran Méndez: We'll experience what others are experiencing
A
A
A
A
A
A
A
B
B
A
A
C
A
Regarding
Java
stuff,
right
yeah,
so
maybe
before
I
will
go
in
details
to
the
agenda
items,
not
sure
if
you
want
to
stay
for
the
Whole
meeting
or
not
some
maybe
did
you
join
to
help.
Do
you
have
specific
questions
you'd
like
to
ask
first
or
you
just
want
to
stay
for
the
Whole
meeting,
and
you
don't
mind
if
I
go
through
the
whole
agenda.
I
would.
C
A
That's
that's
fine,
but
the
dancer
is
current
is
doing
well.
I
mean
word
going
forward
as
you're
gonna
see
on
the
at
the
agenda.
We
have
some
topics
typically
related
to
the
to
the
project,
how
we
handle
working
on
the
weather
code
and
how
to
make
it
standardized.
So
it's
much
better
for
the
community
to
contribute
and
on
board
okay,
perfect.
So
it's
so
it's
pretty
good,
I!
Think
but
yeah.
There
are
Street
three
topics,
but
the
second
one
is
kind
of
let's
say
two-in-one,
so
the
first
one
is
automate
NPM
releases.
A
We
have
releases,
let's
say
a
bit
semi
manual,
so
we,
the
CI
of
course,
is
responsible
for
publishing,
docker
images
and
NPM
packages,
but
to
trigger
the
release.
We
manually
attack
the
repo
with
proper
talk
somewhere
is
used,
and
then
the
CI
picks
up
that
there's
a
talk,
change
to
the
repo
and
triggers
the
publishing
of
docker
image
and
NPM
and
I
wanted
to
briefly
kind
of
discuss
how
we
could
fully
automate,
also
automate
the
NPM
publishing.
So
it's
also
much
easier
to
release.
A
It's
not
just
one
person
that
has
to
do
the
tack
or
make
a
judgement
about
the
tack.
It
can
be
more
people,
because
I
think
it's
going
to
be
much
easier
than
to
crawl
number
of
maintainer
and
one
of
the
at
least
the
solution
that
I
know
that
I've
noticed
I,
never
used
it
on
my
own,
but
I
I
saw
it
in
many
projects
is
that
they
they
kind
of
introduced
a
conventional
comets.
Let
me
open
the
link,
so
basically,
the
the
comets
have
a
given
prefix.
A
A
Or,
for
example,
then,
nothing
is
changed
because
it's
a
a
comet
that
is
not
changing
anything
in
the
package,
but
if
there's
a
there's,
a
fix
or
or
some
different
prefix
that
we
decide
basing
on
the
other
spec,
the
CI
is
doing
a
patch
or
minor
or
major
release,
and
there
are
many
many
known
projects,
for
example
electron.
That
probably
you
know
many
many
people
know
electron.
So
that's,
for
example,
how
they
are
they're
working.
A
So
when
you
go
to
the
history
of
comets,
all
of
them
there
are
they
prefix
and
that's
not
the
only
project
that
is
doing
it
and
that's
the
most
adopted
I
would
say,
solution
that
people
used
to
own
to
automate
publishing
of
the
packages,
of
course,
except
of
prefixes.
We
could
think
about
just
some
specific
labels
on
the
PO
requests,
but
I
think
I
mean,
from
my
point
of
view,
there's
no
point
to
reinvent
the
wheel.
If
the
community
is
already
adopting
the
conventional
comets.
C
C
A
Be
responsible
for
that
and
giving
the
proper
content
for
the
comet
method
yeah
exactly
because
they
as
you
as
you
noticed,
you've,
read
my
story
which
is
I.
Thank
you
so
much.
So
it's
a
as
you
notice,
so
we
during
this
quashing,
you
can
amend
the
the
commit
message
and
the
maintainer
should
know
what
given
change
is
doing
to
the
project.
If
it's
a
fix
or
not,
what
version
should
be
if
it's
mine
or
major
or
patch,
so
he's
making?
This
final
call
amending
the
commit
message
merging
poor
request
and
the
CI
is
triggered.
C
C
A
C
A
So
the
PR
messages
are
not
taking
care
in
the
commit
message.
This
is
something
that
has
reference
used
by
as
a
reference,
but
this
is
not
the
commit
message
when
you
squash
your
pull
request,
then
the
first
committee,
the
police
place,
is
used
as
the
message,
but
you
can,
of
course,
edit
it
in
the
github.
A
A
You
okay,
but
in
general
this
I
mean
like
Francesco
you're,
more
into
Java
right
and
me
how
I
mean
I
know
what
are
you
from
and
I
don't
remember.
We
never
managed
still
finally
introduced
this
automated
NPM
publishing,
but
do
anything
did
you
see
something
some
different
way
of
doing
it?
Maybe
you
know
some
different
way
of
automated
bumping
conversion,
like
maybe
with
how
it's
done
with
Java
and
maven,
for
example,
when
you
want
to
push
it
to
some.
C
A
A
So
it's
an
issue
from
robber
I
think
the
result
that
they
did.
The
first
reason
why
it
was
created
was
a
I
put
request
that
I
was
reviewing
when
I
was
adding
the
automated
publishing
of
dr.
images
where
I
just
touched
the
make
files
and
SH
scripts
and
it
basically,
the
the
idea
from
Robert
here
is
to
again
make
the
project
as
friendly
as
possible
to
as
large
a
number
of
developers
as
possible
from
different
operating
systems,
especially
Windows,
because
we
have,
for
example,
if
it
comes
to
generator.
A
And
we
had
longer
discussion
here,
exchange
here
and
I
mean
at
the
end.
I
agreed
that,
especially
in
in
no
js'.
It's
super
easy
to
get
rid
of
the
Mayock
file,
because
NPM
allows
you
to
have
custom
scripts
defined
in
NPM.
So
instead
of
having
just
additional
Meg
file
to
trigger
docker,
build
it's
better
to
have
a
script
so
later
on,
we
can
do
just
npm
run
docker
build
and
the
the
only
part
in
the
sh
scripts.
A
A
But
again,
it's
really
rare
to
get
I
think
contributors
that
will
contribute
to
both.
So
if
we
can
at
least
make
it
friendly
to
direct
script
developers
and
have
a
usual
approach
that
community
has
to
have
these
targets
inside
the
NPM,
the
package,
Jason
and
I.
Don't
think
it's
a
a
problem.
Of
course
we
cannot
do
it
in
go.
A
A
Yeah,
exactly
so
for
I
think
it's
pretty
common
in
the
previous.
In
my
past
projects,
I
was
always
using
the
had
the
scripts
from
NPM
like
from
package.json,
so
we
could
do
NPM
run
whatever
and
for
ago.
As
you
said,
we're
gonna
use
keep
my
guys
to
tell
us
what's
the
best
yeah,
because
you
guys
I've
got.
It
depends
on
which
company
behind
the
project,
because
for
Google
they
use
a
basil
that
is
pretty
complicated.
A
A
Then
the
other
part
of
the
issue
is
the
adding
some
static
code.
Analysis
and
I
mean
it's
pretty
obvious
that
at
some
kind
of
time
we
have
to
have
those
checks,
especially
that
generator
can
run
on
server.
People
can
run
it
in
different
environments.
We
should
make
sure
that
the
all
the
checks
are
are
in
place.
A
A
A
A
A
A
A
It's
covered
with
Travis,
except
of
two
exceptions,
one
for
the
asking
API
website
and
this
because
we
were
using
ninety
five
for
publishing.
We
also
use
it
for
for
visiting
of
the
website.
So
let's
say
the
whole
CI
is
based
on
not
if
I
didn't
make
sense
to
something
different
and
for
the
generator
recently
I.
When
I
was
doing
the
automation
of
docker
image.
A
Publishing
we
used
github
actions
and
I
would
say
this
is
really
the
direction
you
should
choose.
It's
really
handy
to
use
and
the
amount
of
plugins
the
community
growing
around
github
actions.
It's
just
crazy,
how
much
it
grows
and
how
good
it
is.
So
I
would
say
as
we
go
forward
whenever
we
do
some
changes,
some
improvements.
A
Okay,
so
there
are
also
votes
for
the
key
hop
actions,
because
I
really
nice
and
I
integrated
the
github
by
themself,
but
there
is,
as
far
as
I
remember,
there's
one
problem
with
collections.
That
means
it
is
not
possible
to
arrest
art
jobs
on
manages
to
the
master
branch.
So,
if
anything
fails,
then
you
have
to
create
a
new
pool
new
commit
to
trigger
the
job.
Maybe
it
was
already
fixed,
I,
don't
know
because
I
just
look
at
them,
I
look
at
little
actions.
I,
don't
know
two
months
ago,
so
I
add
another
current
status.
B
A
If
pushing
the
new
versions
of
your
application
to
NPM
or
Saturday
that
ok
ting
pushing
the
new
docker
image
to
the
dock,
already
students,
actually,
that
is
something
fairly
on
that
stage
you
will
need
to
trigger
or,
for
example,
created
the
new
release
or
something
like
that.
But
this
is
something
that
should
have
already
happen
early
ready,
but
you
need
to
remember
about
that
that
there
was
no
option
to
restart
the
job.
A
As
far
the
last
time
I
was
using
gates,
I
had
to
use
the
CLI
to
make
sure
that
the
somehow
encrypt
the
other
secrets
to
save
it
in
Travis-
and
here
it's
just
smoothly
integrated
with
github
UI
on
the
repo
level,
you
just
set
secrets
and
then
use
them
in
the
pipeline.
So
it's
it's
really
pretty
pretty
easy
to
maintain.
A
A
A
A
A
A
No,
we
don't
see
that
okay
or
that
I
mean
it's
our
job
like
you
can
see
the
chat
as
well,
but
when
I
open
it
on
my
screen,
you
don't
see
I'm,
sharing
it
right.
Yes,
okay,
cool,
so
I'm
next
time,
I'm
gonna
mention
that
I'm
reading
the
chat,
so
a
friend
in
the
chat
is
made
some
comments
about
the
topic,
but
there
are
other
jokes:
we
can
ignore
them,
so
one
thing
is
Forks
only
ready
mentioned
once
we
agree.
A
A
A
A
And
working
on
the
same
maintainer
see
if
we
follow
this
approach,
work
on
experience
exactly
the
same
flow
as
others
exactly
now,
the
I
would
just
say
that
do
some
cleanup
and
make
sure
that
we
don't
have
the
keys
and
projects
enabled
in
the
tops,
because
we
already
have
a
lot
of
tops
in
the
repo
we
don't
use
those.
So,
let's
just
disable
them
objections.
A
A
Okay,
only
so
my
suggestion
would
be
to
enforce
scores
and
merging
only
on
the
repo.
So
you
cannot
marriage
April
requests
in
a
classical
way.
That
later
on,
looks
like
on
a
attached
image.
So
the
first
community
that
you
can
see
is
a
result
of
scores
in
marriage
and
the
second
one
is
a
result
of
a
merge
typical
merge
like
classical
merge.
So
again,
even
like
you
could
see
the
the
example
of
electron
and
I
I
would
say
most
of
the
mature
projects.
A
They
follow
the
approach
that
they
have
really
clear:
a
clean
history
of
the
comets,
so
instead
of
marriage,
that
does
not
say
more
too
much.
If
you
don't
click
on
the
link,
you
can
have
a
descriptive,
commit
message
and
again,
it's
also
needed
for
the
conventional
comets
that
we
discussed
at
the
beginning
of
the
meeting.
A
So
without
their
approval.
The
merger
is
not
possible
that
setting
first
of
all,
it's
would
give
a
clear
information
who's.
The
code
owner
so
cause
the
maintainer
of
the
of
the
given
repository
at
the
moment
and
and
its
parts
and
I
think
in
the
in
the
generator,
especially
when
you
have
in
generator.
We
have
several
different
templates
provided
by
different
people.
I
think
it
would
be
also
nice
to
then
to
kind
of
manage
who's
reviewing
which
template,
because
we
know
who
contributed
where
so.
That's
a.
C
A
A
Okay,
so
you're
so
you're,
suggesting
that
we
should
make
sure
that
the
settings
we
set,
because
it's
possible
that
let's
say
I
approve
today,
but
then
the
day
after
you
do
some
additional
commits,
then
my
approval
should
be
dismissed
and
I
should
be
requested
again
to
to
approve
okay,
yeah
I
would
say
that
should
be
by
default.
Enabled
this
clean
of
the
approvals
after
commits
okay.
C
A
A
A
A
C
A
For
sure
you
cannot
assign
it
to
yourself,
I,
don't
think
if
you're,
not
a
member
of
the
organization
you're
not
allowed
to
you,
don't
have
an
option
to
assign
it
to
yourself
I.
Think
only
members
of
the
organization
can
do
it.
So
if
you're
interested
with
picking
up
the
issue
just
start
a
discussion
under
the
issue,
I
mean
pingas
on
it
slack
as
well.
If
you
want
on
the
generator
or
channel,
especially
in
case
of
generator,
but
definitely
do
a
discussion
under
dish
as
well.
B
A
B
Really,
actually,
we
don't
have
yet
a
process
for
onboarding
people,
that's
something
that
we
would
like
to
to
create.
But
yes,
actually
Lucas
is
it's
now
helping
with
this
processes
with
all
this
stuff
right,
so
yeah
and
probably
with
the
collaboration
of
everyone
we
will
be
able
to
make
or
to
create
the
process
for
newcomers,
new
contributors
right
so
yeah,
nothing
really
I,
don't
have
anything
really
relevant
to
say.
A
A
B
A
For
me,
it
was
so
natural
that
I
didn't
even
indicate
it
at
the
beginning,
so
it
will
be
I
mean
for
me.
It
was
so
natural
to
take
over
the
meeting
and
run
the
meeting
that
I
didn't
even
indicate
at
the
beginning,
like
people,
if
you're
confused,
that
you
don't
see
from
but
me
blah
blah
blah
I
didn't.
B
A
Okay,
so
I
would
say
if
there
are
no
more
questions,
comments
thanks.
A
lot
for
joining
guys,
see
you
in
two
weeks,
and
there
will
be
also
after
this
meeting
shortly
after
this
meeting
I'm
going
to
create
a
issue
for
next
meeting.
If
you're
gonna
have
something
for
the
agenda
feel
free
to
add
your
topics,
do
it.