►
From YouTube: Argo Contributor Experience Office Hour 15th Oct 2020
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
A
I
think
we
can
get
started,
but
can
you
hear
me
well?
Can
I
get
the
confirmation?
Yes,
yes,
awesome!
Thank
you!
Yes,
okay,
so
I'm
alex
I'm
a
software
engineer
on
argo
cd
and
today
is
our
contributors
meeting.
So
the
goal
of
the
meeting
is
to,
I
guess,
answer,
questions
and
discuss
issues
and
we
had
some
agenda
for
this
meeting.
A
So
let
me
share
my
screen.
I'm
going
to
show
agenda
here
yeah,
so
I
think
during
last
meeting
I
promised
to
prepare
a
small
presentation
about
argo,
cd,
ui
development
and
I
hope
it
will
take
like
10-15
minutes
and
I
guess
there
will
be
questions
and
we
can
take
them
offline
and
we
had
several.
You
know
proposed
topics,
so
we're
going
to
talk
about
pushing
images.
A
Container
registry
writing
tests
and
then
the
feature
request
and
the
pr
from
peter
and
one
more
question
about
creating
documentation
for
using
different
clusters.
I
guess
during
development,
okay-
and
I
hope
I
will
my
part-
won't
take
more
than
10
minutes.
So
hopefully
we
can
discuss.
A
Okay,
so
since
I'm
first
let
me
get
started,
I
prepared
a
small
presentation
which
is
kind
of
just
a
script
of
what
I'm
going
to
talk
about,
and
I'm
going
to
publish
that
presentation.
So
it
can
be
used
later.
You
know,
if
necessary,.
A
Okay,
so
the
topic
is
argo,
cd,
ui
development
and
first
I
wanted
to
just
go
through
the
basic
basics:
the
stack
which
we
use
for
developing
the
ui.
So
we
use
kind
of
two
languages
I
would
say
typescript
and
sas,
which
is
you
know,
frenzy's,
javascript
and
css.
Fancy
versions
of
these
two
languages
and
we
heavily
use
react
so
pretty
much.
You
know
everything
is
based
on
react
and
we
use
two
components.
React
router,
react,
form
and
kind
of
heads
up
that.
A
I
think
we
both
the
versions
of
these
components
are
pretty
old
and
so
that's
important,
because
the
documentation
might
be
different
for
the
latest
versions,
but
it
is
what
it
is.
This
is
what
we
use
right
now
and
there
are
a
couple
more
components
which
worth
mentioning.
One
is
rxjs
and
I
will
talk
a
little
bit
later.
You
know
more
about
it
and
foundation
sites,
we
use
just
a
css
part
of
it
and
we
basically
we
use
some.
A
So
the
most
important
piece
which
we
use
is
a
grid
framework
and
that's:
okay,
that's
the
most
important
components
and
before
we
talk
about
these
components
more,
I
wanted
to
quickly
go
through
the
directory
structure
which
we
have
in
the
application
and
that's
important
to
know
just
you
know,
so
you
can
understand
how
to
navigate
in
the
code
and
because
argo
ui
is
not
a
giant
application.
So
it's
fairly
simple.
I
was
able
to
fit
the
whole
directory
structure,
the
top
level
into
a
screen
which
is
very
convenient
and
yeah.
A
So
here
we
have
kind
of
three
most
important
pieces:
the
entry
point
of
the
application
bunch
of
modules,
and
I
call
modules
I
call
module
a
separate
kind
of
part
of
an
application
which
is
not
related
to
other
parts.
So,
for
example,
we
have
a
module
applications
and
it
refers
to
argo
city
application.
A
We
have
model
help
which
just
consists
of
a
single
screen
help
screen.
We
have
login,
which
includes
you,
know,
admin,
login,
ss,
login
and
settings
which
has
repositories
clusters
and
so
on.
So
the
application
entry
point.
A
You
know
javascript.
Basically,
you
would
you
the
end
user
would
have
to
load
too
much
javascript
when
navigates
to
applications
and
to
solve
that
problem.
We
have
last
piece
here
which
is
shared
so
here
we
just
have
a
shared
components
which
can
be
reused
across
all
four
application
modules,
and
maybe
I
should
pause
here
like.
Does
it
make
sense
so
far?
Any
questions.
C
Yeah,
I
think,
but
somebody
found
it
and
who
was,
in
the
other,
meet
told
us.
A
D
A
A
Next
kind
of
I
think
important
question.
I
guess
it's
related
to
any
react
application:
how
to
manage
state
so
and
argo
ui.
It
went
for
transition,
we
started
from
using
redux
and
then,
after,
like
a
month
of
dealing
with
redux,
we
realized
it's
just
too
complex
for
argo
cd.
Basically,
rbc
ui
is
not
facebook
and
we
don't
have
so
much
of
shared
ui
state
and
then,
after
you
know,
moving
back
and
forth.
A
We
just
landed
on
just
plain
react:
state
management,
api
and
the
recent
react
upgrade
introduced
hooks
which
really
make
things
better.
So,
if
you're
working
on
the
new
component,
it
is
really
encouraged
by
encouraging
you
to
use
hooks
if,
if
possible,
as
opposed
to
set
state
api
plus,
we
have
two
components.
A
A
And
the
here
is
the
like:
here
is
the
use
case.
Basically,
every
page
gets
some
input
from
user
and
that
input
is
usually
in
the
in
the
url.
So
in
this
example,
so
this
code
snippet
gets
the
name
of
a
user
account
from
the
url
and
then
it
loads
data
asynchronously.
A
And
what
could
happen
is
that
user
might
choose
change
the
url
and
we
supposed
to
kind
of
stop
loading,
the
stop
previous
request
and
then
issue
the
new
request
and
load.
You
know,
asynchronously
data
for
different
account
or
you
know,
request,
might
fail
and
we're
supposed
to
show
error
message
and
retire
button
or
a
request
might
just
take
too
much
time
and
we
should
show
progressing.
A
A
So
the
first
line
is,
you
know,
component
declaration,
so
here
we're
saying
we're
using
data
loader
and
we
provide
two
parameters
here.
One
is
input
and
input
is
typically
input
of
your
function,
which
loads
the
data
and
the
idea
of
you
know
having
the
parameters
separately
is
so
that
data
loader
knows
when
input
changes.
So
if
it
detects
that
input
property
has
changed
it
will,
you
know,
execute
appropriate
actions,
it
could
interrupt
in
flight
loading
request
or
it
can
and
it
can,
you
know,
create
a
new
request
so
and
the
second
parameter
is
the
request.
A
In
this
case
we
get
a
load,
returns
a
promise
and
promise
eventually
either
fail.
Our
returns
account
data
and
then
finally,
you
know
the
children
parameter
of
data
loader
can,
you
know,
will
receive
an
account
and
render
something
yeah.
So
it's
basically-
and
I
know
it's
not
you
know
my
description
is
maybe
not
fully
kind
of
complete,
but
the
good
thing
is
that
there
is
open
source
library,
which
is
pretty
much
does
exactly
the
same
and
that
library
is
actively
developed.
A
It's
called,
I
think,
react
sync
and
it
just
I
mean
this
library
was
created
after
we
created
data
loader
so,
and
I
think
the
ideal
solution
would
be
to
eventually
migrate
to
that
library,
because
there
is
whole
community
there
is
nice
documentation
and
library,
keep
evolving
yeah.
So
that's
kind
of
for
your
information.
If
you,
if
something
is
unclear
here,
I'm
pretty
sure
documentation
the
documentation
here
can
help.
A
C
A
The
idea
is
that
we
were
forced
to
create
that
component
because
it
just
literally
on
each
and
every
page
like
three
four
times
you
have
to
load
data
synchronously,
and
then
you
have
exact
same
set
of
issues
like
how
to
communicate
to
user
that
something
went
wrong
or
just
taking
too
much
time,
so
yeah
cool.
A
So
that's
that
part
takes
care
of
loading
and
managing
state
during
asynchronous
data
losing
and
then
next
very
common
use
case.
Is
you
know
in
input,
state
management
or
like
form,
state
management,
and
we
use
two
libraries-
it's
not
really
visible
from
here,
but
this
is
the
link
and
it
leads
to
a
react
form
component
and
that
component
it
had
like
quite
a
history,
so
it
was
created
first
and
then.
A
B
A
You
would
have
to
go
to
version
2
and
we
already
start
talking
about
migration
to
the
new
version
or
some
new
for
for
management
framework
and
the
second
component,
which
we
use
a
lot,
is
form
field.
Oh,
there
is
a
typo,
so
let's
missing
form
field.
This
is
something
we
created
just
to
you
know
encapsulate
some
some
of
the
repeated
code
which
we
have
to
write
a
lot
in
inargo
ui,
and
here
is
some
some
additional
information,
so
basically
not
additional
information.
A
The
details
about
this
component,
so,
first
of
all,
we
have
component
form
and
what
it
gives
us.
It
provides
us
for
free,
a
container
for
a
form
state
and
form
state
is
just
a
bunch
of
key
value
pairs
and
it
does
for
us
validation,
so
it's
possible
to
kind
of
configure
a
validation
function
and
that
function.
If
it
returns
an
error,
then
a
form
component,
it
will
prevent
form
submission.
A
So
you
can
be
sure
that
if
there
is
an
error,
if
your
function
returns,
an
error
user
won't
be
able
to
click
submit
button,
and
next
it
kind
of
encapsulates
submission.
So
if
validation
is
successful
and
somehow
submission
got
triggered,
this
function
will
be
executed
and
it
will
does
whatever
you
want
it
to
do.
So,
in
this
case
it's
a
login.
So
basically
it's
a
login
form,
as
you
can
imagine.
It
has
couple
fields
and
that
validation
is
just
to
make
sure
that
users
did
enter
username
and
password
and
submission
is
execute.
A
A
A
This
component
first,
it
needs
to
link
to
the
form
api
from
the
react
form
framework
and
the
most
important
pieces
here
is
that
it
needs
to
know
which
key
of
the
form
state
it's
supposed
to
modify
and
and
how
it
should
modify
it.
So
here
we're
saying:
hey
use,
text
input
and
whatever
is
happened
in
the
text
input
it
should
be
stored
in
username,
key
of
the
form
state
and
additionally,
it
has
a
label.
It's
like
a
basically
a
shortcut
for
you.
A
It
will
render
label
component
on
the
left
side
of
the
input,
which
is
you
know.
A
A
Okay
and
basically,
I
think
it's
this
part
is
easier
than
state
management
using
data
loader,
because
it's
just
you
definitely
can
kind
of
understand,
what's
happening
here.
If
you
try
to
create
one
form-
and
I
guess
the
most
important
thing
is
just
to
be
aware
of
these,
two
components
exist
and
we
use
it
for
form
management.
A
A
A
State
defined
by
you
know,
backend
and
second
piece
is
user
preferences
which
are
stored
in
browser,
local
storage
so
and
it's
kind
of
basically,
whenever
either
of
these
changes,
we
need
to
re-render
the
page
and
that's
how
you
can
do
it
using
reactive
extensions.
So
you
can
say
please
combine
two
events,
two
event
sources
and
I
just
want
to
do
something
every
time
when
one
of
them
changes-
and
I
want
to
use
the
latest
version
you
know-
of
the
either
of
them
so
and
the
data
loader
supports
it.
A
So
it's
possible
to
kind
of
feed
that
event
stream
to
data
loader
and
it
will
execute
the
render
function
every
time
when
new
event
is
emitted
here,
yeah
and
at
least
based
on
my
experience.
I
know
it's
hard
to
digest
it
so,
and
I
kind
of
please
be
aware
it
exists
and
if
you
want
to
learn
more
about
it,
just
google
reactive
extensions.
A
It
has
a
lot
of
good
stuff
there
and
we
use
that
in
our
go
ui.
It
helped
us
a
lot
and
the
reason
is,
you
know,
argo
ui,
I
think
we
kind
can.
We
were
forced
to
use
reactive
extensions
even
more
because
we
have
pages
like
application,
details,
page
application
list
page
and
these
pages
kind
of.
Don't
you
know
user,
don't
have
to
reload
the
page
to
get
the
latest
state,
it's
kind
of
live,
and
this
is
where
you
know,
rxjs
library
helps
us
out.
D
Okay,
does
it
make
sense
at
all
yeah
alex?
I
just
had
a
small
question
so
when
you
say
a
stream
of
events
from
the
arco
cd
api,
so
what
information
does
this
argo
cd
api?
Normally?
Is
it.
D
A
Think
we
have
only
two
apis,
yes
and
that's
kind
of
ipi
is
pretty
much
just
a
long
pulling
and
it
uses
service
and
events
protocol.
So
the
idea
is
that
ui
can
consent
request
once
to
backend
and
that
request
returns
application.
A
You
know
json
representation,
okay
and
it
basically
returns
a
new
json
every
time
when
application
changes.
So
it's
it's
long
putting
but
browser
has
native
support
for
it.
Browser
kind
of
under
understands
it's
the
long
pulling
request
and
it
uses
service
and
events
protocol.
So
the
key
thing
is
that
browser
won't
try
to
close
that
request
too
long,
because
it
knows
it's
long
pulling,
and
I
said
we
have
two
events
so
one
event:
sorry,
two
apis
one
api
is
for
application,
metadata
and
second
api
for
application
resources
in
regis.
D
Okay,
alex
so
like,
I
was
working
on
a
task
where
actually
I'm
supposed
to
bring
up
a
ui
banner
that
indicate
information
about
releases
and
stuff.
So
I
should
be
reading
the
information
from
an
api
right
so
which.
A
D
A
I'm
yeah,
I
feel
it's
going
to
be
a
new
api,
but
and
I'm
pretty
sure
it
won't
be
streaming
api.
So
basically,
I
don't
think
it
makes
sense
to
you
know,
live
show
the
banner.
It
would
be
nice,
but
the
streaming
api
is
expensive
is
the
reason.
Is
a
browser
allows
up
to
six
open
connections
to
one
domain
name
and
basically
every
streaming.
Ipa
request
kind
of
comes
into
the
sixth
limit
and
we
already
used
two,
so
we
had
to
so
it
can
it's.
A
A
A
So,
as
I
said,
we
use
sas,
which
is
just
it's
like
a
syntax
sugar
for
css.
Basically,
it's
a
css
like
language
which
produces
css
and
it
gives
you
a
bunch
of
shortcuts
like,
for
example,
here
you
can
generate
classes
and
reduce
duplication
of
you
know,
text
in
class
names
by
using
this
ampersand,
so
this
code
will
produce
actually
three
classes
and
all
of
them
are
going
to
be
so.
The
top
one
is
going
to
be
application
conditions
class
and
these
two
will
be
prefixed
with
application
condition
yeah.
A
Hopefully
that
makes
sense-
and
I
feel
like
sas-
is
it's
kind
of
such
a
simple
library.
You
don't
even
have
to
learn
it
to
start
using,
because
every
css
is
a
valid
size
code
and
as
soon
as
you
need
a
new
feature.
What
I
usually
do,
I
just
google
or
navigate
directly
to
such
documentation
and
look
if
it
supports
kind
of
you
know
some
shortcut
to
produce.
A
Css
so
that
feature
which
I'm
kind
of
showing
you
right
now,
it's
called
nested
classes
and
there
are
several
more
like
variables.
So
basically,
you
can
have
a
variable
and
avoid,
and
that
helps
you
to
avoid
copy
pasting.
Let's
say
same
color.
Actually.
There
is
example
of
that.
Yes,
so
we
use
variable
right
here
and
there
is
also
a
mix
mixing.
A
Just
check
says
documentation
and
you
will
find
something,
and
the
second
is
bram
and
that's
kind
of
important
too
so
bam
is
just
a
convention
and
that
convention
explains
how
to
name
classes
and
the
reason
is
like
css:
don't
really
have
any
tools,
for
you
know,
name
spacing
and
that's
kind
of
our
answer
to
just
avoid
naming
collisions
so
and
it
saves
you
from
you
know,
thinking
too
much
about
how
to
name
your
class.
A
A
The
top
level
piece
of
every
class
name
is
name
of
your
component,
so
application
conditions
is
a
good
example.
So
this
is
a
component
which
can
be
used
by
you
know
itself.
Basically,
it
doesn't
have
to
be
a
part
of
everything
else
you
can
just
you
can
just
put
it
on
the
kind
of
top
level
page
and
it
can
render
you
some
data,
but
application
components
has
elements
like
conditions,
so
application
conditions
has
elements,
and
one
of
such
elements
is
a
condition
and
condition
is
a
it's.
A
I
guess
so
it's
not
so
easy
to
explain
and
there
is
a
whole
web
page,
which
kind
of
explains
the
philosophy
behind
a
block
element
modifier
and
it's
useful
to
know
because
we
use
it
basically,
each
and
every
css
file
in
our
cd
ui
follows
that
condition,
and
so
it
follows
that
name
and
convention.
So
good.
If
you
read
about
it
and
you
know
so,
you
can
just
follow
it
or
you
know.
If
you
want
to
learn
more,
why
it's
useful!
You
know
read
about
it
in
this
url.
G
A
H
I
can
just
say
a
few
words
about
the
ui
as
well,
so,
as
alex
mentioned
before,
we
use
pretty
outdated
components
and
and
outdated
versions
of
the
library,
and
I
think
the
ui
could
really.
It
could
really
use
some
love.
So
if,
if
there's
anyone
familiar
with
all
the
javascript
and
react
stuff
or
wants
to
become
familiar
with
it,
so
I'd
encourage
everyone
to
to
just
give
the
ui
a
little
bit
of
love
and
bring
it
up
to
dates
that
I
think
that
would
be
very
much
appreciated.
H
I
B
I
Kind
of
an
intermediate
step
in
order
to
like
get
everything
more
updated,
it's
a
large
poll
request.
So
if
you
want
me
to
break
that
up,
let
me
know
into
like
multiple,
smaller
ones,
but
every
commit
in
that
should
stand
on
its
own.
A
I
feel
like
the
good
thing
is
that
we
don't
change
ui
that
often
so
it's
I
think
it's
even
fine
to
keep
this
pr.
You
know
like
everything
is
in
the
same
year,
and
basically
because
there
are
not
so
many
viewer
changes,
I
think
it
won't.
You
won't
have
to
merge
it
frequently
and,
as
you
mentioned,
I
think
everyone
who
has
ui
experience.
It
would
be
really
helpful.
If
you
hope,
is
this
new
review
of
the
pr-
and
I
will
try
to
review
it
too
and
but
you
know
more
eyes
is
better.
A
Yeah-
and
I
think
jan
mentioned,
that
you
know,
ui
is
because
we
have
kind
of
code
scanning
and
dependency
scanning,
and
these
tools
reported
a
lot
of
vulnerabilities
just
because
components
which
we
use.
They
have
vulnerabilities
and
we
have
to
update
them
constantly.
We
try
to
update
them
recently,
but
it's
not
done
yet.
There
is,
there
are
still
things
which
can
be
improved,
yeah,
it's
it's.
H
I
Yes,
sir,
so
I
actually
have
done
ui
for
quite
a
long
time,
so,
in
addition
to
enabling
all
the
typing
stuff,
which
will
help
a
lot,
the
next
thing
I'd
like
to
do
is
also
then
start
contributing
some
tests,
because
right
now
there
are
basically
zero
tests,
meaning
every
single
change
you
have
to
test
manually,
which
is
also
why.
B
I
A
J
Sorry,
I
have
a
question.
I
noticed
today
that
in
the
dependencies
for
argo
cd,
the
argo
ui
dependency,
just
it's
just
a
git
repository.
So
is
there
any
versioning
with
our
ui
or
do
we
always
just
pick
up
whatever's
latest
on
master
or
what.
A
Yeah,
I
should
have
mentioned
that
I
I
forgot
to
mention
I'll,
go
ui
so
yeah.
A
K
A
H
Offer
from
industry
what
tim
just
said,
I
think
it's
very
valid
because
whenever
so
now,
with
the
with
a
huge
change
coming
up,
er
updating
all
the
dependencies,
this
will
not
only
affect
our
cd
but
also
argo
workflows.
I
guess.
A
Yeah
it's,
I
guess.
The
choices
we
have
is
either
try
to
introduce
release,
process
or
get
ready
to
work
closely
with
workflow
team,
because
basically
every
change
we
make
like
last
time.
I
made
a
change
in
that
library
and
before
merging
the
change
I
created
pr,
which
upgrades
workflow
and
uses
my
fork
just
to
make
sure
test
process
and
then,
as
soon
as
I
introduce
the
test
passing,
then
I
used
argo
ui
changes
after
the
review,
of
course,
and
then
changed
my
pr
to
use
latest
version
of
ui.
A
So
basically,
this
is
the
current
state,
we're
kind
of
just
trying
to
use
latest
code,
and
I
guess
I
don't
know
like
what
else
people
do
in
this
type
of
situation.
It's
like.
We
have
a
need
to
share
code
and
I
guess
maybe
different
ways
to
use:
go
module,
also
git
modules,
but.
A
I
All
right,
so
I
was
just
going
to
say
so
the
way
that
I've
done
this
in
all
of
the
ui
projects
that
I
kind
of
maintain
historically,
is
for
something
like
argo
ui.
I
would
have
a
separate
release
process,
probably
automated,
using
like
semantic
release
or
something
like
that,
but
then
in
projects
that
depend
on
it.
I
also
then
generally
have
either
renovate
or
greenkeeper
set
up
because,
as
you
pointed
out,
it's
somewhat
obnoxious
to
have
to
go
and
update
the
dependency
every
single
time
that
you
make
a
change
so
as
much
as
possible.
I
I
try
to
automate
that
workflow,
but
part
of
the
benefits
of
doing
that
are
right
now,
by
pulling
in
code
just
from
like
master
it's
a
bit
of
a
leaky.
I
don't
want
to
say
abstraction
but
like,
for
example,
you
can't
start
using
new,
say
typescript
versions
in
argo
ui
with
also
you
know,
without
ensuring
that
both
of
your
dependencies
or
dependent
projects,
you
know,
can
actually
work
against
those,
whereas
if
we
actually
do
build
and
releases
in
argo
ui,
then
what
you're
shipping
is
just
javascript
and
source
maps
and
typings.
I
A
That
really
sounds
good
and
yeah.
I
think,
like
historically
just
argo,
never
had
ui
developer
and
too
many
ui
developers.
I
guess
the
only
one
was
me
and
I'm
not
really
now
I'm
working
on
back
end
and
the
process
which
we
set
up,
it
kind
of
works
for
a
small
team,
but
I
agree
it
doesn't
really
scale.
So
any
kind
of
these
improvements
very
welcome,
and
especially
now
we're
going
to
have
more
people
who
wants
to
work
on
ui
so
that
yeah,
that's.
That
would
be
super
useful.
J
Yeah
with
the
way,
with
the
way
it's
set
up
right
now,
since
it's
still
a
typescript
project
in
the
like
that
that
argo
cd
is
downloading
every
time
you
build
argo
cd,
you
also
have
to
build
our
go
ui
so
like
daniel
saying
you
could
separate
it
into
javascript
and
then
type
declarations,
and
that
would
that
would
that
that
is
the
best
way
to
ship
a
typescript
project
yeah
at.
A
J
Yeah
it's
a
decent
amount
of
work,
though
the
one
other
thing
I
was
gonna
suggest
that's
easy
is.
In
the
past,
we've
had
projects
that
pull
from
other
projects
master
branches,
but
that
kind
of
leads
to
some
riskiness.
Every
time
you
merge
a
big
pull
request
into
master
like
the
one.
J
C
I
think
in
in,
in
general,
I've
tried
to
avoid
the
development
branch
compared
to
everyone,
making
a
branch
make
me
appear
and
pushing
to
master,
especially
because
after
like
10
people
have
pushed
the
development
branch,
there
is
again
ahead
that
the
development
branch
is
in
good
shape.
So
if
you
can
have
another
solution,.
A
A
Proposing
is
main
branch
plus
plus
stable
branch,
so
basically
the
main
or
master
is
your
is
where
your
development
happens,
and
then
you
kind
of
periodically
push
into
stable
merch
into
stable
yeah.
But
it's
the
same.
It's
like
what
it
gives
you.
I
think
you
can
cherry
pick.
You
know
hotfixes
into
stable
yeah
and
you
don't
block.
C
Yeah
a
lower
barrier
for
entry.
I
think
also
I'm
not
sure,
if
that's
what
I
should
call
it,
but
if
I
mean
if
the
consequence
of
that
is
that
the
the
non-stable
branch
has
a
you
know
lower
barrier
for
entry,
then
the
purpose
gets
defeated,
a
bit,
which
means
you're
no
longer
continuous,
but
if
we
can
yeah,
if
we
can
ensure
that
you
know
everything
we
commit
to
what
whatever
that
master
or
development
branch
do,
we
call
it
is
production
ready
and
that
merge
to
stable
is
more
like
having
a
tag
or
release.
C
Okay,
okay,
yeah.
I
think
that
then
I
think
yeah
I
mean
I,
I
think,
as
a
quick
follow-up
on
this
alex
or
tim.
I'm
not
sure
if
you
have
an
issue
on
this,
let's
actually
make
an
issue
put
it
on
a
milestone
and
that
task
for
the
milestone
need
not
be
to
get
the
whole
thing
done,
but
at
least
to
get
the
proposal
and
have
a
discussion
learn
something
so.
L
So
what
what
is
the
reason
for
having
like
a
stable
versus
a
dev
branch,
if
you
have
something
you
want
to
you
know
doing
release
you
would
have,
will
have
a
like
a
rc
you're
gonna
yeah
yeah.
Can
you
rc
tag
for
your
for
your
release?.
L
No,
I'm
not
I'm
not
big
fan
of
having
so
many
branches.
You
know,
I
think,
most
most
open
source.
You
know
they
have
to
have
the
master
and
then
that's
pretty
stable
for
for
most
of
the
source
projects.
I.
A
Guess
the
problem
with
trying
to
always
you
know,
make
it
easier
to
make
changes
in
master
because
now
you're
kind
of
forced
to
test
all
your
users
before
you
imagine
myself,
unless
you're
100
sure
that
change
is
safe,
yeah,
that's
the
only
problem,
and
if
you
have
two
branches,
then
you
kind
of
it's
easier
to
imagine
to
master
and
but
then
whoever
is
trying
to
merge
master
into
the
stable,
have
to
test
group
of
changes.
J
C
J
Yeah,
I
still
think
what
daniel
said
is
super
worth
doing
and
it
sounds
like
we
have
a
typescript.
You
know
better
in
looking
at
it,
so
I
think
that's
a
really
good
start.
C
F
C
A
Okay,
then
next
topic
we
wanted
to
talk
about
is
pushing
images
to
github
container
registry.
It
was
proposed
by
daniel.
I
I
So
I
know
that
bit
me
on
some
of
my
own
clusters,
where
I
started
getting
rate
limiting,
since
we
do
have
options
and
github
now
has
their
own
container
registry.
I
was
wondering
what
people
thought
about
either
migrating
from
docker
hub
to
get
a
container
registry
or
just
pushing
to
both
that
way.
It's
not
a
breaking
change
for
anyone
and
then
maybe
we
could
like
update
the
deployment
manifest
to
point
at
whichever.
B
I
A
I
had
the
ticket
open
already,
you
know,
publish
basically
we
thought
about
that
and
I
guess
we
didn't
take
into
account
really
anything
and
we
only
thought
it's
you
know
they
only
going
to
the
only
you
know
problem
is
that
they
might
delete
image
if
it's
not
used,
and
rightly,
and
that's
why
jc
did
some
evaluation
and
he
looked
at
it
and
he
was
like
you
know
we
didn't
have
enough.
C
I
A
Yeah
and
I
I
stopped
sharing
my
screen,
so
I
can
open
the
document,
but
basically
we
track.
You
know
how
many
times
argo
cd
image
gets
downloaded
and
it's
downloaded
extremely
frequently.
I'm
going
to
show
it
one
more
time,
so
it
could
be
a
real
problem.
I
didn't
know
there
is
a
rate
limit,
so
this
is
a
graph.
A
B
B
I
Out,
okay,
so
the
way
that
it
works
is
if
you're
unauthenticated
it's
per
ip
address
and
your
rate
limited
to
six.
Excuse
me
100
polls
per
six
hours,
if
you're
authenticated
it
rate
limits
based
off
of
you
know
your
your
account
identity
and
your
rate
limited
to
200
polls
every
six
hours.
I
believe
right,
but.
I
A
A
Okay,
I
think
isak.
I
guess
you
had
a
question
about
writing
tests
for
application
controller.
C
It's
like,
I
think,
left
saying
that
he
would
go
in
and
look
at
some
recording,
so
we
could
put
it
back.
Okay,
meeting.
A
A
I
think
we
were
left,
I
think
yeah.
I
guess
that's
why
he
left
the
message
here.
So
we
can
talk
about
okay,
we'll
move
both
topics
to
the
next
meeting
then,
and
okay,
and
next
topic
is
from
regina
aging
documentation
about
development
setup
with
different
clusters.
E
Yeah,
so
I
added
this,
I
thought
that
it
might
be
helpful
for
especially
new
users
when
they're
getting
set
up
doing
like
the
setup
steps
for
development
with
argo
cd
to
have
different
steps
for
using
like
setting
up
environments
with
different
clusters
like
k3d
clusters
or
openshift
clusters.
E
There's
currently
like
not
any
documentation
on
this
that
I
could
find.
I
thought
it'd
be
helpful.
I
wasn't
sure
if
there's
a
reason
that
it's
not
included
already,
I
just
wanted
to
get
people's
ideas
on
it
if
they
think
it
would
help.
G
E
I
mean
that's
slightly
fine,
I
I
started
a
couple.
You
know
installation
instructions
with
like
at
least
k3d
and
open
shift,
so
I
could
probably
set
up
a
pull
request
to
add
some
documentation
for
that.
A
I
personally,
I
feel
like
we
use
a
lot
of
users,
use
developers,
use
mini,
cube
and
k3g.
I
use
mostly
for
you
know
whenever
I
need
to
do
something
which-
and
I
don't
want
to
destroy
my
mini
cup
setup-
I
quickly
create
k3g
cluster,
you
know
do
whatever
I
need
and
it
doesn't
tear
it
down
and
I
feel
like
maybe
because
of
lessons.
I
have
not
migrated
to
k3g
completely,
because
it's
kind
of
yeah.
A
And
the
documentation
can
help
to.
You
know,
share
knowledge
because,
maybe
like
I,
I
keep
using
myself
as
an
example,
but
I
use
mini
q
because
it
works
and
I
maybe
don't
know
enough
of
benefits
of
hit
k3
sk3g.
So
if
there
is,
if
there
was
a
documentation,
maybe
it
would
motivate
me
to
move
on
and
you
know
use
a
better.
L
Choice,
it's
also
be
useful
to
have
like
a
linux
and
mac.
I
think
the
the
mileage
is
different.
The
best
depends
on
like
some
people
have
somehow
some
difficult
time
with
k,
k
3d
using
mac,
but
it's
really
easy
on
the
next
documentation,
which
would
be
helpful
to
have
you
know
different
different
depth
machines
yeah.
I
think.
C
H
A
Okay-
and
we
have
three
minutes
left-
I
yeah-
I
don't
think
we
can.
A
A
M
Yeah,
I
just
want
to
quickly
mention
I
don't
know
if
everybody's
gone
through
the
issues
also
read
jonathan's
comments
because,
like
I
read
them,
and
they
make
a
lot
of
sense
and
give
you
context
of
how
it's
currently
being
done
with
the
values
field
and
how
that's
being
implemented
in
code.
So
I
I
get
how
it
makes
sense
in
code
and
how
it
works.
I
was
just
wondering
from
a
ux
perspective,
if
that's
still
worth
addressing,
to
make
it
easier
for
a
user
to
provide
parameters.
A
I
know
that
from
you
know,
like
users
wanted,
I
just
I
I
was
able
to
you
know
quickly
understand
what
is
it
about
because
it
was
us
in
the
past,
but
maybe
we
will
need
to
you
know
I
will
jump
into
the
threat
and
try
to
so.
Basically,
I
know
for
sure
user
wants
it,
and
if
there
are
additional
questions
I
will
try
to
answer
in
in
the
ticket.
M
Okay,
I
can
start
a
slack
thread
for
this.
Then
okay,.
N
N
N
Which
will
remove
all
the
all
the
typescript,
all
the
yarn
or
the
webpack
from
that
coreco
base?
So
it's
just
golang
and
you
know
some
manifests
and
it'd
be
great
to
get
people's
help
on
redefining
some
of
those
wireframes
and
screenshots
for
using
events
in
the
user
interface,
because
we
don't
have
many
people
with
a
very
great.
You
know
the
maximum
depth
of
knowledge
of
the
different
use
cases
that
people
are
trying
to
solve
with
that.
So
people
want
to
get
involved
with
that.
Obviously
join
the
argo
devs
channel
to
learn
more.
A
Thank
you,
does
it
mean
you
will,
does
it
make
it
even
more
important
to
work
on
argo,
ui
and
your
release
process
or
like
some
way
to
smooth
contributions?
I.
N
Don't
we
don't
find
any?
I
mean
the.
If,
if
you
ask
me
our
number
one
issue
with
the
argo,
a
ui
code
base
was
the
one
that
you
fixed
recently
and
where,
where
there
was
a
particular
library
that
did
not
really
support
a
react,
state
updates
correctly,
which
meant
you
couldn't
build
like
a
variety
of
components
and
now
that's
been
fixed.
We
can
build
kind
of
like
kind
of
smarter
components
that
are
more
isolated
in
the
code
base.
N
The
second
issue
really
is
around
security
vulnerabilities
in
that
code
base
it
it
scores
very
poorly
on
the
snick
reports
and
part
of
that
is,
you
know
my
philosophy
on
security.
Vulnerabilities
is
the
great
best
way
to
deal
with
scooter.
Vulnerability
is
remove
the
vulnerable
code,
get
rid
of
the
vulnerable
dependency,
but
there's
quite
a
lot
of
dependencies
in
that
code
base
that
need
to
be
looked
at,
and
it's
quite
obviously
quite
a
big
piece
of
work.
To
do
that.
That's
that's
you
know.
The
release
process
to
be
honest,
is
absolutely
fine.