►
From YouTube: CDEvents Working Group - 2022-11-09T11:00:05Z
Description
For more Continuous Delivery Foundation content, check out our blog: https://cd.foundation/blog/
A
A
E
A
I
was
hoping:
Andrea
had
prepared
this
meeting
and
was
supposed
to
to
be
here,
but
apparently
he
got
some
some
inevitability,
so
you
couldn't
join
I,
don't
know
why
so
I
haven't
prepared
anything
with
this
now
I
haven't
even
looked
into
the
agenda,
so
I
guess
we
can.
Just
if
you
have
anything
you
would
like
to
bring
up
please
do.
A
A
A
Sorry,
sorry,
I'm,
I'm,
sorry,
I'm,
mistaken
yeah.
G
E
So
we
have
done
the
POC
and
the
recent
in
the
Spinnaker
Summit,
and
that
happened
in
the
trade.
So
we
have.
D
E
Presentation
there
to
show
like
how
Spinnaker
will
be
integrating
with
CD
events,
no.
C
E
G
E
Yeah,
so
that's
the
update
and
we
had
a
chance
to
meet
many
of
the
Spinnaker
folks
there
and
and
talk
about
this
POC
and
and
I
had
some
couple
of
comments
also
like
how
we
can
integrate.
Actually
this
Pinnacle
into
the
series,
our
actual
CD
events,
implementations
right.
A
A
Quite
all
right,
I
guess
what
else
so
yeah
I
just
started
to
say
that
I
was
hoping.
Andrea
should
be
here.
I
I
want
to
share
the
meeting,
but
it's
not
so
I
guess
we
need
to
do
it
in
some
kind
of
cooperation
at
least,
but
one
idea
other
debugger
is
that
we
could
talk
about
guitar
actions
more.
If
you,
if
you're
interested
in
discussing
that
further,
is
there
anything
there
that
you
would
like
to
dig
deeper
into.
D
Sure
so
I
started
working
on
a
boc
for
GitHub
actions
with
Brad's
help.
So
yesterday
Brad
shared
the
POC
that
he
usually
works
with
with
me
for
CD
events
and
based
off
of
some
of
that.
I
got
an
idea
on
how
I
would
create
an
adapter
for
GitHub
actions,
CD
CD,
even
for
GitHub
actions,
and
so.
D
A
little
bit
of
a
disclaimer
I
started
work
on
the
POC
like
an
hour
ago,
so
this
is
definitely
not
not
the
best
where
I'm,
where
I'm
blocked
at
is
I'm,
not
a
patch
I'm,
not
actually
able
to
forming
the
CD
event.
But
it's
I'm
not
able
to
send
the
CD
event.
D
I
can
for
now
go
ahead
and
show
you
what
I'm
exactly
doing
I'll
just
share
the
screen.
If
that's
okay.
A
D
F
Just
something
I
remembered
yeah:
if
someone
makes
a
pull
request
and
is
this
after
they
merge
the
request
it
will
go
or
when
they
raise
the
pull
request.
D
D
Yeah,
so
so
I
I
already
covered
that
part
on
like
what
events
would
come
in
in
the
last
call.
I
will
show
the
talk
to
you
where
it
is
okay,.
E
D
I
have
it
here:
I
can
just
go
over
it
once
before
so
So,
based
on
the
last
meeting.
These
are
the
event
mappings
that
that
were
finalized
for
this
POC.
D
So
whenever
a
pull
request
is
opened,
a
change
created
event
will
be
sent
whenever
the
pull
request
is
edited
or
actually
any
other
events
from
these
events
happen.
Updated
event
is
sent.
Then
if
a
pull
request
is
closed
but
but
not
most,
then
the
change
abandoned
event
is
sent
if
a
pull
request
is
closed
and
merged,
because
GitHub
code
request
will
just
send
a
close
and
we'll
have
to
check,
dig
deeper
and
check
if
it
is
merged
on.
So,
if
that
has
happened,
we
will
send
a
change
most
event.
D
If
the
event
is,
if
the
pull
request
is
open,
we
send
a
created
event
again
same
as
the
opened
event,
and
if
the
pull
request
is
synchronized,
we
send
an
updated
event
again.
So
this
is,
this
is
for
the
GitHub
pull
request
event
type.
There.
D
Type
called
the
pull
request,
review
event
type
and
the
event
that
is,
that
makes
sense
for
us
to
map.
In
this
cases,
I'm
gonna
pull
request
review
is
submitted.
We
can
create
an
event
called
change
reviewed
event
and
send
that
across.
D
Yeah,
so
I'll
just
go
back
to
the
demo
here,
so
it's
not
really
a
demon,
so
I
have
taken
the
payload
from
there
and
then
put
in
this
file
here
called
event.json.
So
I
can
work
with
this
one
locally
I
am
push
how
I'm
working
with
this
in
GitHub
actions
itself
is
in
GitHub
actions,
I'm
taking
the
event
payload
and
putting
it
in
a
environment,
video,
so
I'm
dumping
it
in
the
environment.
Variable
here
called
GitHub
context
and.
D
Go
go
file
as
a
script,
so
I'm
doing
go,
run
and
then
running
it
like
this.
So
it
is
easier
for
me
to
test
also
locally
and
for
local
testing.
What
I'm
doing
is
I'm
running
this
command
and
putting
all
of
the
event
data
in
this
environment
variable
and
then
reading
it
back
from
here
in
the
script,
so
how
I'm
testing
it
is.
D
If
there
is
a
event
of
event,
name
is
pull
request
and
the
action
is
open
when
I
say,
send
a
new
change
created
event
currently
I'm
not
able
to
really
make
this
work.
So
if
I,
if
I
run
this
here,
I
get
a
invalid
memory
address,
so
I
get
a
sick
for
sick,
sick
I
I'm,
not
sure
why
this
is
happening.
It's
been
a
long
time
since
I've
done
goals.
Also
I
think
it
might
be
something
related
to
this
Cloud
event.
Map
cloud
event
over
here.
D
Also,
I
am
not
really
failing
this
anywhere,
so
it
might
be
silently
feeling
I
I,
don't
know
if
something's
going
wrong
so
where
I'm
sending
this
is
this
link
over
here?
It's
basically
I'm
running
k-native
and
I'm
running
the
cloud
cloud
events,
player
and
I'm
hoping
to
get
this
CD
event
on
the
cloud
events
player.
So
it's
running
over
here.
H
Very
technical
thing:
you
have
a
store
before
the
CE
I'm,
not
sure,
if
you're
supposed
to
send
a
reference
or
if
you're
signing
object,
then
my
gov
knowledge
not
the
best,
but
you
can
check
that
one.
H
F
H
F
Actually,
just
updated
the
readme,
because
the
the
snippet
was
wrong
as
well,
and
I
actually
put
the
star
on
as
well.
D
F
Do
you
plan
to
for
the
actions
you
plan
to
put
a
Docker
file
in
with
the
action
eventually
or
you
just
use
the
script.
D
D
I'm
I'm,
starting
to
feel
that
the
docker
file
might
be
an
Overkill,
because
if
it's
a
simple
one
page
or
go
script,
maybe
it
could
be
just
use
like
this
is
what
I'm
thinking,
but
nothing
that
we'll
have
to
pull
the
modules
for
for
CD
events.
A
Docker
file
might
help
so.
D
So
I'll
set
up
go
over
here
and
then
I'm
running
it
like
this.
So
I
did
a
simple
check
in
the
beginning
to
just
pass
the
Json
and
it
worked
pretty
well.
So
the
only
I
think
hurdle
that
will
come
in
running
this
without
a
Docker
file
would
be
modules,
but
that
also
should
be
pretty
straightforward,
but
every
time
it
runs,
then
it
will
have
collect,
pulled
the
CD
CD
event,
Library
I,
don't
know
which
is
which
is
better
or
worse
like.
D
C
F
Focus,
the
the
user
won't
even
see
it
right
because
they're
just
calling
the
the
action.
Yes
if
they
have
to
install,
go
and
all
that,
like
that's
pretty
tedious.
F
D
D
G
A
F
Yeah
I
I've
had
experience
working
with
production,
so
I
was
just
giving
him
like
some
pointers
on
what
I've
done
in
the
past
yeah,
so
I'm
just
helping
them.
It's
a
fun
project,
yeah.
What.
F
So
you
can
use
JavaScript
my
effort,
yeah
JavaScript,
I
I,
like
personally,
to
use
Docker
file
and
Bash.
So
for
the
entry
point
of
the
docker
file,
you
just
have
a
bash
script
and
then
that
can
run
and
and
the
user
can
specify
what
version
of
things
that
they
want
as
well.
So
it
works
quite
nicely
instead
of
baking
it
into
the
image.
F
G
D
D
F
I
can
help
you
there,
because
the
the
Jenkins
X
team
wants
to
make
an
adapter
for
City
events
as
well.
So
maybe
that
could
be
good
to
do
both
those
projects
at
the
same
time,
and
then
you
can
use
their
like
adapter.
They
build
to
sort
of
practice
with,
for
both.
F
Yeah,
maybe
we
can
ask
detect
on
you.
You
probably
know
yourself
that
you
might
even
help
me
the
techton
adapter
that
they
made
for
City
events.
I
assume
it
would
be
pretty
much
the
same
for
Jenkins
X
as
well.
D
But
I'm
not
sure
if
they
have
a
that
testing
instance
for
the
community
I'm.
F
B
And
I
I
spoke
to
the
Jenkins
X
team
yesterday
and
encouraged
them
to
get
re-engaged
on
on
this
piece
of
work.
So
it's
good
timing
to
to
refresh
that
and
and
get
them
working
on
it
perfect.
F
A
F
A
F
B
X
has
a
has
a
feature
called
lighthouse,
which
is
how
how
it
currently
responds
to
all
of
the
events
that
are
coming
in
from
GitHub
or
whichever
source
code
control
system
you're
using
and
I
think.
The
intention
was
to
extend
that
Lighthouse
functionality
to
to
accept
native
events
so
where
those
events
are
coming
from
is
an
open
question
at
the
moment.
So
it's
possible
that
there
would
be
a
broker
container
that
that
you
could
instantiate
on
either
side,
but
the
the
formal
interface
where
Jenkins
X
would
be
through
lighthouse.
G
A
Yeah,
but
that
was
what
I
was
thinking
if
it
should
be
a
broker
container
in
its
own,
which
multiple
producers
and
consumers
could
connect
to.
So
there
would
be
one
Central
broker
or
a
federator
of
these
events.
A
E
E
A
A
E
E
Yeah
so
now
Spinnaker
is
the
one
like
another.
It's
a
it's
in
another
namespacing
kubernetes
controller
yeah.
So
it
can
be
like
something
else
can
come
with
different
namespace
in
the
same
same
cluster
as
well
again,.
D
Yes,
yeah
I
was
I
had
an
idea
where,
once
this
POC
is
at
a
point
where
it
can
be
used
or
where
repositories,
maybe
it
will
be
good
to
employ
it
in
the
Jenkins
repo
Jenkins
X
repo
and
probably
have
like
one
place
where
we
can
like
collect
all
these
events
and
probably
do
like
whatever
experiments
like
Dura
or
whatever,
or
anyone.
D
But
I
think
it
will
be
a
good
place
to
start
getting
like
real-time
CD
runs
so
or
regarding
pull
requests,
changes
and
stuff
and
I'm
guessing
pactone
already
must
be
doing
it.
Oh,
but
but
I,
but
I
feel
like
this.
Could
this
could
like
be
for
like
just
experimentation
with
some
metrics,
probably
which
will
be
GitHub
specific.
F
And
a
question
at
what
point
would
we
bring
this
repo
or
project
into
the
City
events
or
for
GitHub,
so
we
can
make
like
issues
and
collaborate
together,
Etc.
A
Well,
whenever
I
would
say,
I
mean
we
can
set
that
up
more
or
less
right
away.
Of
course,
we
can
do
that
more
or
less
directly.
F
Yeah
I
think
it's
good
for
like
doing
a
bit
of
planning
around
making
some
issues
and
delegating
to
work,
etc.
A
I
could
do
that
for
you.
What
should
the
name
be
then
we
haven't
yet
added
any
books.
I
think.
Should
this
be
considered
a
book
or
is
it
more
of
a.
F
You
can
change
it,
but
yeah.
It's
not
ideal.
What
about
just
as
simple
as
like
see
the
events,
GitHub
action
or
something.
A
A
F
At
least
we
would
probably
have
to
is
there
like
in
the
org
developer
group,
that
we
could,
because
we're
not
part
of
the
org
as
well
yeah.
A
Yeah
we
haven't
really
set
it
up,
I
guess,
but
we
I.
Can
you
write
to
us?
Let's
see,
no,
none
of
you
are
there
right?
Okay!
A
Yes,
there
is
a
member
level,
though,
if
you
just
placed
your
GitHub
ideas,
I
can
add
it
as
members.
There
sure
in
the
chapter
in
the
foreign.
D
H
A
F
Unless
we
did,
if
we
call
it
actions
and
then
we
have
one
forget
lab
one
for
because
once
we
make
the
first
one,
then
we'll
GitHub,
it's
going
to
be
easy
to
do
it
with
gitlab,
and
it
depends.
If
you
want
it
in
the
same
Repro
though,
or
if
you
want
different
reports
for
each.
F
I
get
I
think
it's
called
a
runner
I,
don't
think
it's
got
an
action.
Yeah.
F
H
A
C
Sure,
let
me
just
share
something
sure.
C
So
I've
started.
B
Creating
some
some
issues
to
to
give
some
talking
points,
and
this
is
what
I'm
proposing
as
our
audience
strategy.
B
So
what
I'm
doing
here
is
looking
at
you
know
what
are
the
personas
of
of
the
audience
that
we
expect
this
material
to
be
addressing
and
I've
identified
a
number
of
different
categories
that
I
think
we're
we're
likely
to
see
people
coming
in
from
from
these
areas
of
Interest
and
then
off
the
back
of
this
I'm
documenting
a
basic
strategy
for
how
we
communicate
with
those
audiences.
B
Are
the
key
one
is:
is
about
understanding
how
we
want
to
be
able
to
publish
into
the
website,
because,
obviously,
what
we're
doing
is
developing
spec
revisions
that
are
independent
of
the
website
content
and
and
then
exposing
those
spec
versions
in
a
controlled
fashion
through
the
website.
B
So
that
means
that
we
we
need
to
be
consistent
about
how
we're
doing
that
and
have
a
a
sort
of
General
process
for
managing
the
release
cycle
of
that,
and
also
addressing
any
presentation
and
style
issues
that
that
might
come
along
with
that.
So
that
we're
we're
getting
a
consistent,
look
and
feel
on
the
website.
But
we're
not
creating
a
lot
of
overhead
that
that's
adding
to
the
difficulty
of
maintaining
the
the
specs.
B
A
C
A
If
there
should
be
some
specific
release
handling
together
with
the
release
of
an
a
new
event,
Edition
or
a
spec
spectrivation,
something
on
top
of
the
ordinary
pull
request,
reviews
video
for
for
the
spec,
of
course,.
B
Yeah,
so
it
would
be
good
to
kind
of
have
that
conversation
over
the
next
few
days,
if,
if,
if
people
have
some
time,
just
to
add
add
some
comments
to
this,
because.
C
It
was
I.
B
I
really
need
to
understand
what
you
want
from
from
this,
so
that
I
can
make
sure
that
I'm
I'm
addressing
any
challenges
that
that
might
occur
from
you,
know,
presentation
and
rendering
perspective,
and
things
like
that.
A
Yeah
I
guess
one
one.
First
input
would
be
that
it's
impossible
to
select
the
revision
of
the
spec
graphic
on
the
website
through
a
drop
down
menu
or
something
I
guess
and
have
the
correct
version
of
the
spec
shown,
depending
on
what
what
revision
you
selected
and
then
we
also
briefly
touched
upon
if
we
should
somehow
run
the
documentation
from
the
sdks
and
maybe
from
other
things
as
well
in
the
organization.
A
That's
certainly
a
stretch
for
this
work
that
you're
doing,
but
we
should
maybe
have
it
in
our
minds
in
a
way
when
thinking
about
how
to
visualize
the
different
revisions
for
such
things
as
well.
If
those
should
go
around
finding
out
in
some
way
so
selecting
a
revision
of
the
specs
should
then
also
affect
maybe
other
parts
of
the
same
website
which
could
contain
other.
B
Yeah
yeah,
so
there
could
be
some
interesting
challenges
in
in
in
doing
that
in
so
we
need
to
structure
it
so
that
it's
compatible
with
the
you
know.
The
underlying
doxy
platform,
yeah.
A
B
A
We
do
tag
the
source
files
with
the
version
number
and
we
obviously
create
the
GitHub
release
on
it.
But
there
there's
no
real
written
process
for
it,
I
think
I.
Guess
we
have
to
do
the
obvious
things
and
we
don't
I,
don't
think
we
have
a
well.
A
Oh,
that's
a
good
question:
I,
don't
know!
Actually
I
mean
I,
haven't
done
it,
it's
it's
Andrea
I
will
stand
it.
We've
done
it
twice
now,
one
for
this
zero
one,
zero
version,
and
once
for
this
year
one
one
version:
I
guess:
that's
Andrea
would
go
to
the
city
event
stopped.
Every
Point
update
some
tag
there
and
then
regenerate
it.
But
I
I
can't
say
I,
don't
know.
A
To
always
look
at
to
get
the
latest
from
the
spec
now
that
could
also
be
one
way.
Maybe
that
is
okay,
because
yeah
this
was
actually
I
mean
the
the
setup
of
the
website.
The
rendering
of
the
documentation
was
done,
of
course,
before
we
made
the
first
release
and
then
the
Run
of
tags
really
to
do
to
look
for
so
maybe
it's
just
a
nightly
job
that
takes
the
latest.
That
could
actually
be
the
case
today.
You
might
be
right
so
then
it's
not
that
all
a
finished
process.
B
Yeah
and
obviously
you
know
that
any
solution
could
be
deemed
the
correct
one
at
this
stage,
but
it
would
be
good
to
to
actually
have
a
firm
decision
on
what
the
desired
behavior
is,
because
that
that
could
have
a
big
influence
on
how
we
need
to
structure
some
of
the
site.
A
Yeah
I
guess
you
would
you
would
have
quite
three
hands
down
here.
I
remind
you
if
you
have
ideas
on
what
would
benefit
from
the
document
generation
there,
but
I
would
guess
that,
of
course,
one
way
is
to
have
a
configuration
file
in
the
city
events.devel
report
for
the
documentation
that
states
what
tag
should
we
now
get
from
the
Spectrum
when
we
render
this.
C
B
I
think
that's
a
that's
a
very
good
question
and
probably
key
to
how
much
trouble
we're
likely
to
create
for
ourselves
yeah
it
it
a
lot
of
it
will
depend
on
whether
the
existing
process
has
considered
this
already
or
if
it's
just
pulling
the
the
latest
version.
B
A
I
can
I
can
certainly
say
it
with
99.9
certainty
that
it's
not
considering
that
at
all.
Yet
this
is
too
too
new
for
that,
but
and
obviously
it
sounds
probably
easier
to
just.
We
render
everything.
So
we
know
that
we
have
the
correct
things
of
everything
every
time
and
that
might
work
the
first
say
10
or
20
revisions,
but
then
we
might
end
up
in
very
slow
publishing
on
the
website.
B
Yeah
so
I
think
there's
definitely
an
outstanding
question
there
as
to
what
the
desired
behavior
is-
and
this
is
one
of
the
challenges
that
really
hit
the
Jenkins
X
team.
G
B
Because,
with
with
with
Jenkins
X,
you
heard
a
similar
problem,
where
you
know
you
had
a
a
basic
API
for
for
the
the
tool
set
and
then
a
lot
of
explanatory
documentation
that
was
independent
from
the
the
application
itself.
B
But
every
time
there
was
a
significant
change
to
the
API,
all
of
your
other
documentation
went
out
of
sync
and
you
also
ended
up
in
a
situation
where
Google
was
mainly
indexing,
previous
versions
of
documentation,
and
so
when
you're,
trying
to
search
for
information,
you
were
getting
stale
information
rather
than
the
the
latest.
A
C
B
G
G
A
B
G
B
The
kubernetes
case,
what
they
tend
to
do
is
just
put
lots
and
lots
of
footnotes
on
on
the
narrative
trying
to
warn
you
of
incompatibilities
and
now
that
gets
very
messy
and
you're
talking
about
having
to
have
a
full-time
documentation,
Nation
Nation
Nations,
of
all
of
those
issues.
B
So
so
this
is
one
of
one
of
those
areas
where
the
early
strategy
needs
to
reflect.
You
know
how
much
time
and
investment
is
going
to
be
available
to
maintaining
the
documentation.
A
B
A
Right,
yeah,
I
think
what
we're
quite
you
know
in
a
quite
good
situation
in
a
way
because
we
have
our
primer
and
our
readme
and
all
those
things
in
the
same
repository
as
the
spec
itself.
Currently,
what
we
don't
have
there
is,
of
course,
obviously
the
actual
website
text,
which
is
specific
for
the
website.
A
So
whenever
that
comes
out
of
sync
with
a
spec,
if
you
do
some
groundbreaking
changes,
which
then
makes
the
the
documentation
on
the
websites
invalid
or
need
to
be
updated
for
some
reason,
then
we
will
have
that
problem,
I
guess
anyway,
but
for
now
we
have
most
of
the
text
in
the
same
repo
and
that's
one
benefit,
then
I
guess
I,
don't
know,
I
mean
before
we
reach
the
1.0
version
of
CD
events.
It
doesn't
make
too
much
sense
to
have
a
very,
very
easy
access
of
old
revisions
of
the
spec.
A
Obviously,
when
we
Implement
our
proof
of
Concepts
and
different
things,
we
need
to
be
to
be
able
to
find
the
old
versions
as
well
if
we
are
compatible
with
this
0.1
or
0.203
version
or
whatever.
But
it's
not
as
important
as
when
we
have
reached
1.0
and
and
after
that
so
I'm.
Thinking
now
on
your
assignment
territory,
we
we
requested
to
have
a
versioning
of
of
the
website
right
in
place.
C
A
I
guess
the
list
that
could
be
done,
then,
is
that
is
to
just
show
what
what
version
of
the
spec
are.
We
now
looking
at
what?
What
is
this
latest
version
that
we
see
now?
Is
it
version
0.1
or
zero
two?
What
is
it
and
that
could,
of
course,
be
retrieved
from
the
file
in
the
spec
repo
or
a
tag?
Maybe
yeah.
B
A
B
Yes,
yeah
and
then
then
you
can,
you
can
make
those
links
visible
within
the
navigation
menu
so
that
you
can
choose
to
go
to
V1
or
V2
or
at
that
level,
that
that's.
That
would
certainly
suffice
for
being
able
to
surface
the
the
spec
documentation
through
the
website
with
a
historical
View.
B
B
B
Otherwise
you
get
into
trying
to
surface
multiple
versions
of
the
content
of
the
website,
which
multiplies
your
maintenance
overhead.
A
Yeah
yeah
and
we
don't
want
to
spend
most
of
our
time
on
the
actual
website
documentation.
We
want
to,
of
course,
work
on
the
protocol
and
the
Box
primarily
but
yeah.
It
needs
to
be
get
to
go
around
and
on.
Of
course,
we
the
website
information
is
up
to
date,
always
whether
the
least
minimum
effort
required,
of
course
or
whatever
but
I,
mean.
A
Now,
since
we
do
this
big
restructuring
I
guess
we
want
to
redo
it
again
for
the
1.0
version
in
I,
don't
know,
say
a
year
or
something
so
it
would
be
good
if
we
have
a
good
idea
on
how
to
actually
handle
this
also
for
some
time
after
1.0
now,
otherwise,
I
could
say
that
it's
it's
good
enough
to
just
show
the
latest,
but
as
long
as
we
show
what's
what
exact
same
semantic
version
is
used
is
shown
here.
A
That
I
would
like
to
see
that
there
is
a
way
to
to
handle
it
through
differentiating
different
versions
through
URL
parameters,
maybe
them
or
something
similar.
But
then
again
we
need
to
say
sure
this
decide
should
we
remember
all
versions
every
time
we
render
the
documentation
or
should
we
just
render
the
last
one
and
then
somehow
keep
the
old
ones,
but
then
we
need
to
make
sure
that
the
links
still
work
from
the
old
one.
I
don't
know
between
wherever.
If
you
replace
some
files,
what
happens?
C
B
If
you,
if
you,
if
you're
doing
it
dynamically,
then
well
I,
suppose
it's
gonna,
it's
gonna
run
as
a
independent
process.
Isn't
it
so
it's
yeah
you
are.
You
are
going
to
hit
that
situation
where,
as
you
get
more
versions,
you
are
going
to
slow
down
the
overall
rendering
process,
but
you
won't
be
re-importing
content
against
the
historical
spec
versions.
So
in
theory
you
could
go
through
and
manually
prune
out
ones
that
you
you
felt
were
deprecated
and
that
would
allow
you
to
keep
on
top
of
it.
A
C
B
Because
of
the
way
this
is
deploying
as
a
CI
CD
process,
it
is
effectively
going
to
render
the
whole
website
every
time
you.
You
merge
a
pull
request,
but
that
will
that
that
will
be
very
rapid
with
only
a
handful
of
versions
in
place
and
if
it
starts
to
become
a
problem,
you
can
manually
prune
old
versions
out
of
the
website,
repo
and-
and
so
you
know,
you
can
deprecate
versions
and
remove
them
and
then
you'll
speed
things
up
again.
A
Yeah
I
guess
there
needs
to
be
a
process
for
that
as
well.
I
guess
there
could
be
some
kind
of
should
we
call
it
this
allow
best
file
or
something
that
says
these
versions
should
not
anymore
be
rendered,
or
these
are
too
old,
just
render
the
ones
which
are
maximum
one
year
old
or
two
years
old
or
something.
B
I
mean
I
think
this.
This
is
probably
an
infrequent
enough
process
that
you
can
do
it
manually.
C
B
Because
you're,
if
you're,
if
you're
import,
if
if
your
publication
of
a
new
specification
is
always
just
taking
the
the
latest
release
version
and
adding
that
to
the
repo
in
a
new
namespace,
then
what
that
will
enable
you
to
do
is
just
you'll
have
an
accumulating
set
of
namespaces
in
in
the
the
website.
Repo
yeah.
B
A
A
A
The
the
website
trip
as
far
as
I
know:
I,
don't
think
it
is
I
might
be
wrong,
but
I
don't
think
anyway,
let's
see
we're
out
of
time
already
so
the
rest
of
you
in
the
meeting.
A
A
But
it's
perfect
that
we
try
to
solve
this
now
when
we
are
in
such
an
early
stage,
so
we
don't
get
hit
by
it
later
on.
If
we
can
solve
it
now,
okay,
so
I
guess
we
need
to
stop
there,
but
they
also
reach
out
area
on
slack.
If
you
need
more
attention
on
it,
if
we
should
be
there.