►
From YouTube: CDEvents Working Group (EMEA/AMERICAS) - May 16, 2023
Description
For more Continuous Delivery Foundation content, check out our blog: https://cd.foundation/blog/
A
B
No,
unfortunately,
we
have
a
travel
restriction,
so
I
was
unable
to
go,
unfortunately,
but
for
my
manager
went
so
that
that
was
at
least
one
of
us
went.
That
was
not
too
shabby,
but
yeah
yeah
I've
been
actually
sick,
like
oh
yeah.
These
last,
like
few
weeks
well
I've
been
trying
to
like
get
it
under
control
like
sick,
as
in
like
just
like
more
paranoia
like
I,
had
like
really
high
blood
pressure,
so
I've
been
like
trying.
B
B
Really
good
figuring
that
out,
but
but
yeah
what
about
you?
Did
you
go
end
up
going
to
cdcon
or
no.
C
B
C
B
Yeah
yeah,
but
I'm
gonna
be
there
for
like
two
weeks
so
that
that
should
be
fun.
B
A
B
Went
to
South,
Korea
and
I.
A
B
Tried
live
shrimp
one
time
so
yeah
yeah
I'll
try
anything
once
right,
so
yeah
yeah
but
yeah.
That
was
that
was
really
fun
and
then
is
it.
Is
it
just
gonna,
be
me
and
you
or
was
Andre
also.
C
B
But
let
me
let
me
open
up
the
dock.
Real
quick
kind
of.
B
A
B
Oh,
no,
no
just
or
the
hack
mdio
on
one
that
we
use
for
this
meeting,
meaning
notes.
A
C
B
It's
actually
basically
done
so.
The
reason
why
the
hold
up
is
I
need
to
go
through
internal
teams,
which
I
have
a
scheduled
meeting
for
Thursday
with
two
big
teams
here
at
Apple,
and
the
reason
for
these
meetings
is
basically
to
make
sure
that
our
goals
and
proposals
align
with
their
use
cases,
because
you
know
the
whole
goal
is
to
make
this
work
for
everyone
right
yeah.
So
what
I
was
thinking
is,
since
this
is
a
little.
B
This
is
taking
a
little
bit
longer
to
get
these
meetings
and
you
know
kind
of
get
more
of
the
internal
Apple.
You
know
people
all
on
board
on
the
same
page,
because
you
know
it's
a
fairly
big
company
and
you
know
I
I
want
to
what
I
was
thinking
is
I
want
to
create
a
draft
PR
and
then
maybe
we
can
iterate
their,
maybe
well,
there's
only
a
few
action
items
that
I
have
so
I'll
start
from
from
the
top.
B
B
To
oh
yeah
yeah,
just
a
different,
commit
yeah,
yeah,
okay,
cool
cool,
so
what
I
will
do
then
is
I
will
create
a
draft
PR
after
this
meeting,
so
you
guys
can
at
least
get
an
idea
of
like
what
I'm
you
know
what
we're
thinking
about
and
then
on
Thursday.
If
that
meeting
goes
well,
I
can
then
move
it
out
of
draft,
and
we
can
put
it
in.
You
know
ready
to
be
reviewed
if
that
works.
For
you
perfect
sure,.
D
C
D
Know
there
was
another
another
topic:
I
just
want
to
point
out.
Steve
did
a
talk
at
open
ssf
and
he
pointed
out
that
there
are
20
million
known
Jenkins
work
was.
D
D
C
C
A
D
D
If
we,
what
you
know,
that's
another
topic
that
the
the
artillius
community
really
have
to
chew
on
a
little
harder
in
making
it
easier
for
them
to
add
or
tell
us
to
the
data
the
evidence
store
process,
but
we'll
we'll
see
how
that
goes.
We're
working
on
it,
yeah.
B
Yeah
yeah
I
didn't
want
to
interrupt
their
but
yeah
yeah.
So
this
is
the
lynx
proposal,
so
it
has
quite
a
few
things
in
here
kind
of
some
outlines,
so
the
basic
gist
of
The
Proposal.
It's
basically
what
we
talked
about
kind
of
the
hybrid
approach.
I
do
have
a
comment
in
here
that
is
hard
to
see,
but
but
it
goes
over,
like
semantics
is
the
first
section
which
is
which
is
nice.
You
know
just
having
you
know
something
that
we
all
agree
upon.
B
The
goals
of
this
of
this
proposal,
which
is
you
know,
obviously
being
scalable,
provide
a
quick
way
retrieving
all
relative
links
keep
linked
data
structures,
small
and
simple,
so
that's
kind
of
the
goals
there,
so
I
talked
about
iPhone
links,
I
kind
of
give
kind
of
the
idea.
B
You
know
like
what
it
is.
You
know
people
that,
like
especially
here
at
Apple,
we
don't.
We
don't
have
too
many
familiarity
with
links
in
general.
Nor
do
we
have
any
like
even
with
tracing
people,
don't
really
know
tracing,
so
they
don't
really
have
this
concept
of
like
connecting
events
or
anything
like
that
to
go
over
some
descriptions
of
of
iPhone
links,
kind
of
the
semantics
there
and
the
biggest
thing
here
is
an
Eiffel.
B
It
would
be
a
an
array
of
links
like
you
know,
everything
that
had
happened,
but
in
this
proposal
we
actually
state
that
links
will
always
be
the
direct
parent
of
of
whatever
thing.
So,
if
you
need
to
get
like
a
grandparent,
you
actually
need
to
call
an
API
which
I
talked
about
this
link
service,
which
can
be
a
service,
a
part
of
the
whole
city,
events,
event,
bus
lifecycle
or
or
service,
or
what?
What
not.
B
Exactly
exactly
so
so
it
always
so
when
you
get
in
a
vet,
you're
always
going
to
get
the
immediate
parent
and
the
reason
for
that
is
nine
times
out
of
ten.
That's
what
you
care
about,
because
that's
usually
the
thing.
That's
calling
you
right
or
expecting
you
to
do
something,
but
you
know
we
don't.
We
didn't
want
to
limit
the
the
potential
use
cases.
So
if
you
wanted
to
go
out,
you
know
grandparent,
grandparent
or
you
know,
whatever
ancestor,
you
can
go
ahead
and
retrieve
and
whatnot.
B
So
the
biggest
thing
here
is:
we
talk
about
adding
this
new
proposed
field
here,
which
is
a
link,
and
the
biggest
thing
in
Eiffel
is
that
you
guys
have
this
custom
data
which
I'm
sure
you're
aware
of
which
we
have
the
span
context,
which
I
think
we
can
start
here
and
if
it
works
well,
we
can
potentially
find
a
good
place
in
the
Eiffel
spec
that
allows
for
this,
but
I
think
right
now,
just
keeping
it
in
custom
data
because
we
don't
want
to
marry
like
we
don't
want
to
force,
see
events
like
special
use
cases
onto
onto
Eiffel,
so
I
think
having
a
custom
data
for
now
and
then,
if
there
is
a
bigger
Outreach
needed
for
Eiffel,
we
can
always
move
it
out.
C
Have
some
some
questions?
I,
don't
really
recognize
the
the
links,
the
object
from
iPhone
if
I
go
up
to
the
the
previous,
oh
right
right
there,
which
one
so
it
looks
yeah
this
one.
So
it
looks
there
that
you
have
a
link,
so
if
it's
contain
an
object
of
meta
and
data,
so
what
it
actually
looks
like
in
iPhone
is
that
you
have
metadata
and
links
on
the
same
level
and
there's
no
links.
No
external
links.
Objects
like
this,
so
so
links
is
something
on
the
same
level
as
data
and
and
meta.
B
B
Right
so
this
is
a
triggered
event.
C
C
B
Okay,
okay,
cool
cool:
let
me
update
that
this
needs
to
be
moved
out
and
then
let
me
link
this
and
then
use
it.
Okay,
so
cool
okay,
so
I
just
need
to
move
the
metadata
and
data
and
just
put
the
actual
links
there.
Okay,
that
makes
it
easier.
B
C
C
B
Wanted
yeah
yeah.
We
definitely
want
this
method
because
we
need
to
know
that
it
was
an
activity
triggered
a
bit
and
then
okay,
let
me
go
back
here
then
go
back
to
the
schemas
I
just
need
to
look
at
the
links,
object,
I'm,
pretty.
C
B
B
Maybe
maybe
it
makes
sense
just
to
include
this
then
this
whole
thing
well,
yeah
well,.
B
Yeah
no
I'm,
just
because
like
well.
The
issue
is
we
need
the
meta
like.
We
need
at
least
the
metadata
and
probably
the
custom
data
as
well,
because
the
custom
data
could
have
something
that
that
you
know
the
the
consumer
would
want
or
potentially
could
act
on
right.
C
B
Exactly
exactly
so,
I'm
thinking
is
potentially
because
the
Eiffel
link
could
still
be
the
same
like
at
the
DB
level
right
and
when
you
request
the
service,
when
you
you
know
it
could
just
adapt
this
and
change
this.
To
look
like
this.
Essentially,
you
know
with
whatever
metadata
and
and
what
have
you
so,
maybe
that's
a
potential
option.
B
I'll
have
to
think
about
that
some,
but
that's
definitely
an
interesting,
an
interesting
problem
here.
Yeah.
Let
me
let
me
think
about
that,
but
yeah,
because
that
kind
of
changes
things,
because
we
we
definitely
need
at
least
the
type
it
would
be
good
to
have
the
version
as
well,
just
to
know
actually,
maybe
not
the
version,
maybe
that
we
could
just
have
the
the
type
and
then
also
the
metadata
or
the
custom
data
is
also
really
important.
B
I,
also
like
the
triggers
like
knowing
what
triggered
it
like.
Basically,
we
need
everything.
C
But
then
you
have
it
all
I
mean
and
then
the
next
step.
The
next
event
you
send
you
will
thank.
C
B
Yeah,
but
that's
what
I'm
saying
like
you
only
ever
have
the
immediate
parent
that
that's
making
it
so
like
the
so
you
wouldn't
have
any
number
of
ancestors.
It
would
only
be
the
immediate
pairing,
so
you'd
only
ever
see
see
this
so
so
that
that
might
be
fine,
but
yeah
because,
like
we,
we
definitely
don't
want
to
spend
all
ancestors
like
that
would
just
be
way
too
large.
But,
having
like
the
immediate
impairing
I
think
makes
the
most
sense.
C
Yeah,
that's
the
event
type
which
we
call.
We
call
it
five
as
well
in
City
events.
So
that's
how
it's
called.
C
D
And
what
is
the,
what
is
the
target
that
is
the
the
following
event.
B
C
No,
whenever,
where
we've
done
something
too
I
mean
we
that's
up
to
the
bus
to
Federate
it
wherever
it
can
Federate
right,
okay,
so
on,
we
don't
specify
any
listener
here.
In
events.
C
B
Okay,
I
don't
know
if
we
need
to
go
that
crazy
for
this
spec
just
yet.
B
Yeah
definitely
like
I
want
to
get
so
so
the
idea
is
to
kind
of
lay
the
foundation
and
then,
if
we
need
or
we'll
we'll
need
that
at
some
point,
we'll
need
some
way
of
ensuring
the
the
like,
knowing
where
something
came
from
right,
which
it
would
be
a
signature
but
I
feel
like
yeah.
That
would
be
like
the
next
step
was
like
kind
of
like
security
related,
so
but
yeah.
That
would
be
not
part
of
the
aspect,
but
definitely
a
future
spec.
D
So
the
reason
why
I
bring
these
the
the
topic
of
signatures
of
it's
because
I
am
on
a
probably
a
Folly
of
a
task
to
try
to
get
the
open,
ssf
and
the
CDF
to
think
more
like
each
other
you're.
C
D
D
The
Folks
at
the
open
ssf
are
all
security.
Folks,
really
a
few
of
them
really
have
any
real
devops
experience.
This
conversation
would
go,
some
of
them
would
understand
some
of
it.
Others
would
be
like
I'm,
not
sure
what
we're
talking
about
yeah.
C
D
The
CD
side,
we
never
talk
about
security,
so
the
more
that
we
can
make
that
a
priority
and
pull
that
information
in
the
better
off
will
be
to
get
the
eyes
of
the
open,
SS
app
and
the
CDF
needs
the
open
ssf
and
the
open
ssf
really
needs
the
expertise
coming
from
the
CDF.
So
that's
why
I
bring
this
kind
of
stuff
up?
D
I
I
think
it's
cool
that
we
get
a
base
version
out
there,
but
if
we
can
be
sure
we
keep
talking
about
that
and
I
know
that
we've
talked
about
signatures
and
see
the
events
in
general.
So
just
from
marketing
and
eyes
perspective,
the
more
we
do
around
security,
the
better
we'll
be.
B
B
Exactly
exactly
that's
exactly
what
else
was
just
going
to
mention
so
for
this
spec
we
could
probably
leave
it
out,
but
definitely
for
CD
events
that
can
be
an
action
item
for
another
proposal
on
addressing
you
know,
authenticity
of
the
requests
or
or
an
event.
D
B
C
We
have
had
people
from
from
who
who
wants
to
Wilson.
That
was
the
most
see
me
thing
that
copier
ago
I
don't
remember
his
name
now,
someone
who's,
also
opener
stuff,
I'm,
pretty
sure
yeah
anyway,
so
yeah,
we
have
had
some
discussions,
but
it
was
some
time
ago
now.
So
we
should
really
revive
those
suggestions
for
sure,
yeah,
I
think
so
too,
and.
D
I,
this
is
my
first
introduction
to
Eiffel
Eiffel,
so
I
apologize
for
being
not.
C
C
It's
very
similar
to
City
events.
I
would
say
it's
a
predecessor
to
citizens
in
some
sense.
It's
it's
what
we
use
today
in
Ericsson,
which
is
solid.
It's
we
use
it
while
waiting
proceed.
Events
models
we
recycled
because
it
serves
our
needs
quite
well
cool,
but
there
are
benefits
perceived
events
that
we
want
to
leverage
on,
of
course.
So,
therefore,
we
are
highly
involved
to
make
sure
this.
C
D
C
C
Just
how
it
has
been
I
mean
it
just
has
happened
to
not
stand
up
anywhere
so
I
guess,
but
when
the
community
was
created
a
long
time
ago,
it
was
it's
really
considered.
If
we
should
be
part
of
some
higher
Foundation
and
since
then
we
haven't
really
made
the
effort
either
to
to
try
to
join,
but
yeah
well,
I
have
thought
of
it,
but
I'm
not
sure
if
it's
worth
it
now,
when
we
are
looking
for
Citizens
but
I.
D
Well,
there's
I
mean
it's
an
interesting
proposition
between
the
two.
It
feels
to
me
like
this
communication
protocol,
maybe
a
little
farther
along
in
some
areas,
but
not
as
far
along
in
other
areas
as
CD
events
is
gone.
So.
C
D
Maybe
we
just
did
that
with
Red
Hats
and
porous
with
artillius
yeah.
C
B
If
it
makes
sense-
but
you
know
like
like
you
know
like
what
mute
was
just
saying-
you
know
like
you
know,
this
should
be
like
an
inspiration
and
actually,
after
checking
out
the
example
like
we
need
the
format
to
be
more
like
this,
as
opposed
to
just
a
link
of
ID.
So
we're
gonna
have
to
make
some
changes
to
get
what
we
need.
You
know
for
us
to
be
able
to
use
it
internally
at
Apple
to
look
something
more
similar
to
this.
B
So
if
there
is
overlap,
then
I
I
would
think
you
know
that.
That's
that's
a
good
idea,
but
yeah
like
like
Emil,
was
saying
like
I
I,
think
you
know,
if
they're
trying
to
solve
the
same
problem,
you
know
having
two
competing
unless
one
can
leverage
the
other
which
which
isn't
being
done
today,
I
don't
know
if
it
makes
too
much
sense,
but
but
yeah,
but
yeah
going
back
to
what
you
were
saying
a
meal.
B
So
this
is
kind
of
like
what
I'm
thinking
for
like
the
structure,
because
we
do
need
the
metadata
we
do
need.
The
data,
like
you
know
like
this
triggers
and
whatnot,
and
this
custom
data
as
well,
but
we
can
actually
change
this
type
to
be
the
actual
type
of
the.
C
C
B
Gonna
help
in
the
future,
so
I
do
have
some
images
that
I've
been
working
on
and
I.
Think
images
are
way
more
helpful
than
just
text
so.
D
B
D
D
C
There's
also
some
more
generic
one
more
generic
one
there,
but
it's
long
time
since
it
was
updated,
but
it's
anyway
there
is
some
information
that
could
be
interesting.
If
you
have
time
goodness
I'm
interested
in
reader.
B
Yeah
so,
and
and
like
I
said
Emil,
that's
why,
like
I,
wanted
to
have
these
internal
meetings
with
apple
just
to
make
sure
what
did
because
the
less
data,
the
better
right
yeah
for
good
reason?
But
but
you
know,
like
I,
said
I
can
I
can
pass
along.
Is
there
any
caveat
or
is
there
any
field
that
you
see
right
now
that
you
might
not
think
is
useful,
like
I'm,
like
maybe
Target,
isn't
really
useful.
D
Target's
one
that
confused
me,
that's
why
I
suddenly
said
well
wait!
A
second
is
this
a
component
yeah.
B
C
What
points
out
the
event
that
we
are
linking
to
the
parent?
Really
that's
what
the
parent
is
and
the
rest
is
just
application
from
that
event,
so
I
I'm
still
a
bit
confused
how
we
should
use
this
complex
structure
of
the
event's
body,
but
anyway,
yeah
yeah.
B
B
Yes,
yes,
exactly
so,
and
that's
what
I
was
thinking
just
get
rid
of
this,
because
it's
like
right
here
or
we
can
just
rename
this
to
that
essentially
so
either
one
is
fine.
B
Yeah
yeah,
so
my
thinking
was
most.
You
would
have
always
like
one
parent,
but
my
thinking
was.
You
could
potentially
have
a
service
that
groups
a
bunch
of
requests
together
to
send
it
down
like
as
a
single
event.
But
like
has
you
know
like
like,
for
instance,
let's
say
you
have
something
that
publishes
artifact
a
b
and
c,
and
that
a
b
and
c
gets
triggered
for
a
single
event
to
like
deploy
something
right.
B
That
would
technically
have
three
parents
of
a
b
and
c
of
of
deploy,
deploying
right
and
that's
kind
of
like
where
I
was
thinking
with
that.
C
So
that's
one
thing
and
I
mean
you
know
if
we
quite
often
have
multiple
links
in
our
events
as
well
to
different
types
of
targets,
so
one
could
be
the
parents
another
one
could
be
the
the
context
of
some
other
kind
it
could
be,
and
if
you
send
a
test
event,
you
can
have
the
target
pointing
at
a
artifact.
You
have
to
test
it.
Then
it's
not
the
parents
probably,
but
it's
it's
a
relation
to
another
event
that
was
sent
earlier,
an
artifact
event.
C
A
D
A
D
Hey
guys,
I
have
to
jump
because
I
have
an
articulous
meeting.
I
just
want
to
I'm
sorry
to
interrupt,
but
I
just
want
to
bring
up
one
thing
before:
I
jump.
Congratulations!
Mail
on
getting
the
City
events,
talk,
contributor
award
and
we
have
elections
coming
out
for
the
the
CDF
Toc.
C
D
I
have
the
list
of
everybody
that
I
believe
should
be
in
that
in
that
voter.md
file,
which
gives
us
a
election
Authority
or
a
registered
voter.
What
I'm
thinking
of
doing,
though,
is
sending
out
an
email
to
everybody
and
telling
them
to
check
that
their
name's
on
the
list.
Let
them
know
that
the
election's
coming
up
and
if
they're
not
send
them
the
Forum
to
register.
C
D
That's
kind
of
taken
care
of
I'll
take
care
of
it
because
I
have
that
list
from
the
the
the
CDF
Awards,
so
I'll
just
send
out
to
everybody
and
say
Here's
the
process
make
sure
your
name
is
on
the
list
and
if
not,
please,
register
and
we
want
to
make
sure
we
can
get
CD
events
in
the
TOC.
C
I
did
send
a
heads
up
on
the
inspection
of
a
few
weeks
ago
before
this
water
70
was
created.
Okay,
so
we
can
also
make
a
new
comment
on
the
stack,
Channel
or
yeah.
A
mail
is
fine
as
well
for
both.
C
C
D
B
Cool
all
right,
so
I
I.
C
B
Up
moving
the
custom
data
just
to
parent,
because
I
think
parent
ID
and
Global
ID,
just
at
this
level,
makes
makes
sense.
But
I
I
hear
what
you
were
saying
about
links
having
context
to
various
other
links.
B
I'm,
just
like
so
as
a
part
of
the
cicd
flow
I'm,
trying
to
think.
A
See
what
you're
talking
about
yeah
yeah
I
would,
let
me
see
the
switch
and
then
I'll
stop
sharing
soon.
C
B
C
B
C
C
The
next
thing
that
happens
is
that
the
pipeline
is
triggered
so
an
activity
trigger
event
deny
for
distance,
which
signals
that
now
a
new
pipeline
has
been
started
caused
by
this
Source
change
event.
So
there's
a
course
link.
So
this
is
the
link.
The
Targets
in
this
event
is
then
the
event
ID
of
this
one.
So
that's
the
Target
Field
in
the
link
right.
B
C
C
Which
then
links
to
the
triggered
events
saying
that
now
I
started
something
and
the
thing
I
started
was
this
activity
triggered
thing
that
was
secured
before
and
then
here's
another
event,
another
link
type
called
activity
execution.
So
this
is
not
really
caused
by
that
event
being
sent
it's
rather
caused
by
a
timer
or
a
queue,
or
something
like
that,
and
it's
just
a
reference
that
what
we
have
started
is
what
is
referenced
by
this
activity
execution
and
then
the
next
thing
that
happens
is
that
we
create
a
baseline.
C
So
this
is
a
composition
defined
event,
as
they
call
it
an
apple
which
signals
that
we
have
now
a
new
Baseline
created
based
on
this
new
source
change.
That
came
up
here
so
here
we
see
that
one
of
the
elements
in
this
baseline
or
the
element,
probably
in
this
based
on
the
I,
mean
the
one
of
the
parts
of
the
Baseline
itself-
is
actually
this
Source
change
this
new
source
change.
So
we
have
a
reference
to
the
Distortion
that
was
created
a
while
ago,
and
it
was
created.
I
mean
this.
C
This
composition
was
then
created
under
the
context
of
this
pipeline.
So
there
are
two
different
events:
it's
a
sorry,
two
different
links
there
with
different
types
because
they
mean
different
things.
One
of
the
links
mean
that
this
is
the
source
change
the
element
that
is
part
of
the
Baseline.
They
are
the
link
points
to
the
context
in
which
this
was
was
secreted,
so
those
are
two
different
links
in
the
same
event
and
then
we
create
an
artifact.
That
means
we
built
something
here.
C
We
create
the
artifact
and
we
reference
then
to
the
baseline
or
the
composition
that
was
used
when
building
this
artifact.
So
that's
one
event
link
there
and
then
we
have
the
finished
event
saying
that
now
this
activity
was
finished,
the
first
the
build
step
was
finished,
so
this
is
actually
not
the
full
pipeline.
It's
just
one
of
the
steps
in
the
pipeline
and
again
we
link
to
the
activity
trigger
event
to
say
what
what
was
finished.
C
So
we
don't
include
all
information
about
the
activity
itself
in
this
event,
but
we'd
rather
link
it
to
this
other
event.
So
here
we
are
heavily
dependent
on
on
a
database
collecting
all
these
events,
of
course,
so
we
can
afterwards
see
what
happened
in
what
order,
and
then
we
were
calling
water
as
well
yeah,
yeah
and
and.
C
This
case
test
with
finished
events,
and
then
we
have
something
called
confidence
level
event
which
is
more
or
less
a
maturity
level
saying
that
after
we
have
tested
this
now,
we
have
gained
a
new
confidence
of
this
artifacts.
So
the
subject
of
this
confidence
level
for
this
maturity,
the
subject
of
it
was
this
artifact
that
we
tested,
and
it
was
done
under
the
context
of
this
step
in
the
pipeline
and
going
on
similarly
yeah
whatever.
So
you
get
the
point
yeah
a
clear
example
of
multiple
links.
C
From
from
several
of
these
events
in
the
cicd
context,
yeah
yeah
we
I
mean
we
only
include
the
basic
information,
the
type
of
the
event
or
the
link
and
the
the
target
ID.
B
The
one
thing
that
I
think
that's
a
little
bit
different
here
is
the
city
event
or
the
CD
service
will
only
care
about
the
delivery
of
an
artifact
and
if
they
don't,
if
they
also
care
about
the
test
or
whatever,
and
and
that's
why
I
want
I
wanted
to
go
over
the
the
links
Service
as
well,
they
can
actually
call
the
links
service,
saying
Hey,
I
want
the
test
event
of
This
Global
ID.
B
Give
me
all
the
links
for
test
events,
and
it
would
go
ahead
and
return
those
if
they
cared
about
them.
If
they
need
that
context,
but
nine
times
out
of
ten
a
CD
event
or
CD
service
is
only
going
to
care
about
the
delivery
of
an
artifact
and
then
do
something
off
that
artifact
from
a
CI
system.
So
that's
kind
of
like
what
I'm
thinking.
B
Yeah
definitely
in
this
regard
for
CV
side
of
things
for
CI
it's
a
little
bit
different
like
you
would
only
really
care
about
a
source
change
really
and
then
the
CI
system.
You
know
whatever
triggered
you
know,
that's
that
Source
change
would
send
the
CD
event
saying
hey.
B
There
was
a
source
change
to
let's
say
GitHub
and
it
would
just
have
you
know
a
root
link,
I
guess
to
itself
or
whoever
saying
that
this
event
or
this
change
got
submitted
and
now
build
build
this
artifact
test
it,
but
I,
don't
think
it
would
need
I,
don't
think
the
CI
event
or
sorry.
B
The
CI
system
would
necessarily
care
about
all
the
links
between
these
things,
mostly
because
was
when,
when
it's
ingesting
the
like
the
artifact,
for
example
the
artifact
or,
let's
say
the
test-
let's
say
it
finishes
a
test.
That's
still
within
the
system,
and
you
know
it
submits
that
submits
a
test.
Successful
the
parent
ID
is
still
going
to
be
itself
right,
so
I
don't
think
it
needs
much
more
information.
Like
it's
just
saying,
hey
this
came
from
us.
We
know
you
know
where
all
that
information
is
already
right.
B
We
already
have
all
that
information,
so
so
yeah
I
I,
like
like
I,
said
I,
just
don't
see
a
the
need
for
anything
other
than
the
parent
in
this
case,
even
if
you're,
a
CI
system
or
CD
system
and
I
could
be
wrong
here
and
that's
and
that's
why,
like
a
lot
of
a
lot
of
the
talks,
are
the
meanings
that
I
had
last
week
or
no
two
weeks
ago,
Office
document
was
only
CD
people.
B
This
week
it's
CI
people,
so
so
I'm
gonna
get
a
different
set
of
use
cases
there,
but
but
the
this
is.
This
is
kind
of
why
I'm
leaning
towards
maybe
we
only
need
the
parent
and
and
not
this
not
like
a
context.
For
example,
yeah
I,
don't
think
we
need.
We
need
this
graph
structure
because.
C
B
C
I
mean-
and
we
have
multiple
purposes
for
it,
I
mean
one
is
for
driving
the
flow
itself.
So
as
you
call
it
a
cioc
lead
system
that
would
know
that
now
this
is
an
artifact
and
we
should
test
it
and
then
we
test
it
and
then
we
should
publish
it
and
set
some
confidence
on
it
or
whatever,
but
that's
just
the
CI
system
itself,
the
the
engine,
the
CI
engine
or
the
pipeline
engine,
but
we
have
a
lot
of
other
consumers
in
this
area
as
well.
C
B
B
Each
other-
yes,
yes,
okay,
I,
think
that's
where
the
disconnect
is
okay,
so
I
understand
now,
okay,
so
the
the
reason
so
we
didn't
get
into
the
link
service
because
it's
low,
it's,
we
didn't
get
that
to
the
section
of
that
of
that
proposal,
but
everything
you're
asking
for
is
part
of
the
link
service.
So
basically,
this
link
service
has
all
this
information.
It's
basically
the
Eiffel
links
in
the
database.
That's
basically
what
it
is
I
mean
we'll
probably
need
to
do
a
little
bit
of
changing,
but
so
yeah
I
think
I.
B
C
But
in
order
to
populate
the
link
service,
then
you
wouldn't
need
all
the
or
to
duplicate
all
this
data,
because
the
link
service
would
would
listen
to
all
events
right.
Yes,
so
you
wouldn't
need
to
duplicate
all
data
from
the
parent,
because
the
link
service
has
already
here
heard
about
the
parent
event
when
it
came
on
the
bus,
yeah.
B
B
The
the
only
reason
why
the
parent
is
there
is
because
a
lot
of
the
times
Services
that's
what
they
care
about.
That's
why
the
parent
is
in
in
the
actual
in
the
payload.
In
this
case,
and
rather
than
going
and
calling
because,
like
you
know,
when
you
get
an
event,
you're
gonna
be
like
hey
I
need
this
parent,
like
all
almost
all
the
time
right,
so
we're
just
gonna.
B
Send
that
to
you
always
that's
kind
of
the
thinking
there
and
if
you
need
more,
you
can
go
ahead
and
reach
out
to
the
link
service
to
give
you
more
information,
if
you
need
it,
but
a
lot
of
the
time
you
just
you're,
mostly
interested
in
the
parent
but,
like
you
said
the
UI
of
showing
everything
you
would
be
more
interested
in,
not
the
parent
but
in
this
case
the
link
service,
which
returns
everything
connected.
C
But
if
we,
for
example,
it
would
be
a
a
consumer,
then
that
listens
to
this
artifact
published
event.
In
the
end,
they're
saying
that
this
thought
that
was
published,
then
it
will
I
mean
the
parents.
To
this
I
mean
the
artifact
created
event
is
far
away
from
this.
It
was
a
long
time
ago.
It
was
created,
then
a
little
activities
that
happened
in
between
a
lot
of
things.
C
So
the
parents
to
this
artifact
published
event
is
probably
not
the
artifact
created
event,
but
rather
a
another
activity
or
a
test
case
or
something
a
test
case
execution
or
some
confidence
level
or
whatever.
Or
would
you
consider
that
when
sending
this
artifact
published
event,
then
you
would
always
set
the
parent
to
a
certain
type
of
event?
That
is
the
the
artifact
event
in
this.
B
Case
yeah,
my
thinking
is
like
you
would
group,
so
that's
that's.
How
I
was
thinking
about
I
didn't
outline
that
and
the
proposal,
but
I
probably
I,
definitely
should
it
would
always
brute
by
the
type
that
you're
a
part
of
like,
for
instance,
a
test
execution.
You
know
its
parent
would
always
be
a
source
code
change
or
or
a
build,
for
example,.
B
Yeah
build
exactly
so
yeah,
so
that
was
my
thinking,
because
I
I
think
it
it's
hard
to
make
sense
of.
You
know
like
like
what
you
were
saying,
like
an
artifact
published
being
a
test
event
like
what
can
you
do
with
that
right?
So
my
thinking
was
like
it
would
always
be
a
part
of
the
artifact
creation
or
build,
or
whatever.
B
Yeah
yeah,
so
so
yeah
yeah
I'll
I'll,
definitely
outline
that.
That's
definitely
that's
a
really
good
call
out,
because
I
was
thinking
about
that
and
just
kind
of
made
that
assumption.
But
I
did
but
yeah
I
didn't
really
put
in
words.
B
C
Yeah,
so
the
consumer
of
like
published,
doesn't
need
to
to
to
find
out
the
ultra
creative
event
in
a
database,
but
it
could
just
get
it
by
default.
I
was
just
thinking
one
more
thing
that
about
the
the
context
here,
but
we
often
use
this
context
links
you
know
to
some
activity
related
events
which
says
you've
watched
pipeline
step.
It
was
done,
but
in
your
proposal
that
maybe
it
shouldn't
really
be
a
in
that
case.
It's
not
really
a
context
or
sorry,
a
parent
link.
C
It's
rather
a
some
local
ID
of
the
pipeline
step
that
you
could
introduce
in
in
some
of
the.
In
all
the
events.
In
a
certain
step,
you
can
group
those
together
if
you.
B
So
I
I
just
want
to
make
sure
I'm
understanding
what
you
mean
by
a
local
ID
just
to
make
sure
like
because
I
heard,
like
everything
you
said,
but
I
I
I
just
want
to
make
sure
I'm
I'm
assuming
the
same
thing
so
you're
saying
like
let's
say
we
have
stage
a
Stage
B
stage,
C
right
stage,
a
would
be
the
local
ID.
Or
is
this
like
a
unique
ID
or.
C
Like
what
it's
sort
of
stage,
so
so,
when
okay,
okay,
when
a
bill
happens
in
stage
eight
for
example,
then
the
all
of
all
events
related
to
that
build.
Of
course,
it's
an
ultra
created
event,
but
there
could
be
other
events,
maybe
as
well.
They
should
all
include
that
that
ID
of
the
stage
itself
stage
eight
okay
and
then
maybe
some
tests
are
happening
in
Stage
B
on
that
artifact,
but
they
would
then
be
referencing
some
other
context.
C
I
mean
another
stage,
yeah
yeah
and
then
maybe
the
publishing
or
the
deployment
or
the
delivery
or
whatever
it
could
be.
A
third
in
stage
see
them
and
that's
yeah,
so
so
that
we
can
see
what
what
activities
or
what?
What
yeah,
what
artifacts
events
more
or
less?
What
we're
happening
in
different
steps
of
the
pipeline.
B
Yeah
yeah,
that
that
makes
sense,
okay,
cool
I
can
definitely
update
that
I
think
that's
definitely
a
good
call
out,
because
that
at
least
we'll
give
you
some
context
of
the
structure
of
the
pipeline
as
well.
So
that's
definitely
really
helpful.
Okay,
cool!
That's
a
really
good
call
out
so
I'll
definitely
update
that.
B
Yeah,
no,
no
I,
really
like
that.
Okay,
cool
yeah,
that's
really
good
feedback.
So
let
me
update
that
and
then
I
will.
Let
you
guys
know
on
well.
I'll
still
submit
I
I
have
a
bunch
of
comments
down
on
the
proposal,
but
I'll
submit
the
draft
to
at
least
show.
B
B
Okay,
cool
cool
so
go
ahead
and
do
that
and
then
I'll
also
add
some
other
like
images
just
so
you
know
people
that
are
not
familiar
or
you
know
you
know
to
get
a
better
understanding
of
like
the
flow
of
things
you
know
like.
Like
you
said
you
know,
once
I
saw
this
picture
I
think
we
immediately
both
understood
like
what
we
were
talking
about
so
yeah
picture.
B
You
know
a
picture
says
a
thousand
words
right,
so
I'll
I'll
include
some
pictures
to
kind
of
help,
explain
some
things
and
then
we'll
go
from
there,
but
yeah.
This
was
really
good,
really
great
yeah,
yeah,
yeah,
okay,
cool
I
think
we're
pretty
I
think
we're
almost
done.
Yeah.
C
A
C
So
we
didn't
talk
about
the
roadmap,
but
I
mean
that's
also
just
links
anyway.
So
okay,
Google
for
the
agenda,
if
we
started
I,
didn't
know
55
minutes
into
the
meeting,
so
the
action
items
were
empty.
There
were
no
action
items
and
trace
the
order
to
mentioned
this
about
the
COC
project,
representation
Representatives,
so
just
keep
it
in
in
the
minutes.
Even
though
Tracy
talked
about
it
there
not
too
much
to
mention
that,
and
then
we
have
the
zero
series
which
was
released.
A
C
B
Support
yeah,
yeah
and,
like
I,
said
I
think
once
the
draft
is
out
and
like
I
said
once
I
have
this
last
internal
meeting
with
with
the
two
major
teams
here
at
Apple
that
are
interested
in
this
and
they're
quite
large
teams.
If
we
can
get
them
on
board,
it's
it'll
be
a
huge
win
for
CD
events
in
in
general,
so
that's
kind
of
the
plan
and
yeah
so
I
think
I.
Think
this
I
think
we're
moving
along
great.
B
Like
I
said
a
lot
a
little
bit
slower
than
I
would
have
liked.
But
you
know,
let's
just
do
the
nature
of
completely.
C
Then
we
just
have
the
link
straight
for
the
Java
SDK
in
golang
and
Python
and
Jenkins
plugin
I.
Don't
think
you
have
any
information
on
those.
No
I,
don't
probably
probably
skip
all
that
I.
Don't
have
any
other
replacement,
what
else
either
so?
Okay,
you
can
skip
that
from
the
agenda.
So
that's
what
was
in
the
agenda
really
so
I
wanted
to
discuss
this
as
a
main
focus,
and
now
we
did
so
that's
perfect.
All.
B
Right
awesome,
awesome,
cool
so
also
I'll,
give
a
summary
also
to
Andrea.
I
know
he's
out,
but
you
know
I
think
I'm
sure
he's
also
really
curious
on
on.
C
B
Update
him
and
then
we
can
go
from
there.
Yeah.
B
Yeah
definitely
all
right:
yeah
I'll
send
one
later
today
and
yeah
meal
I'll
be
talking
a
little
bit
later.
Then
right,
yeah,
perfect.
All
right
have
a
good
one.
All.