►
From YouTube: Code Review Weekly Workshop - Apr 28, 2023
Description
In this session we talk about finding issues in a code review and pair up on a review.
A
Computer
okay!
Well
thanks
for
hopping
on
the
code
review,
weekly
Workshop
and
maybe
we'll
talk
about
some
things
regarding
code
review
and
maybe
we'll
appear
on
a
code
review
I,
don't
think
I
really
have
anything
to
talk
about.
Okay.
Regarding
code
review,
I
have
been
unavailable
for
code
reviews
because
they've
been
working
on
other
things.
A
B
D
A
Yeah,
that's
great
well,
I
have
done
there.
I
have
been
pinged
on
a
couple
of
code
reviews
that
were
more
or
less
interesting
that
maybe
we
could
I
could
present
as
just
hey
here's
how
I
approached
reviewing
it.
Here's
some
things
that
came
from
it,
but
before
we
do
that,
do
you
have
any
code
reviews
that
maybe
you'd
like
to
do
together
during
this
time.
A
A
Okay,
all
right!
I'll
share
my
screen.
Let
me
close
some
things
out
so
last
night,
my
brother
and
I.
We
hung
out
Oliver
Zoom,
slash,
Discord
and
he's
also
a
programmer,
so
we're
having
fun
talking
about
programming,
but
then
we
haven't
like
played
any
games
together
in
a
long
time.
It's
like.
Oh,
we
gotta
play
something
and
I
played
in
playing
like
Doom
one
a
lot
by
myself.
C
A
Yeah
like
so
it's
like
something's
loading,
I
started
Doom,
one
on
Nightmare
and
I.
Try
to
go
through
it
I'm
so
I'm
like
not
great
at
it,
but
it's
like
okay
I
get
I've
gotten
better
at
it
over
the
time,
but
I've
always
known,
like
hey,
there's
certain
mods
that
let
you
do
this
multiplayer
and
stuff.
So
we
got
so
that's
what
we
did.
We
set
up
the
mods
for
Doom,
one
to
play,
multiplayer
where
we
could
Co-op
together
and
then
do
the
old
school
death
match
and
man.
It
was
so
much
fun.
B
A
A
A
It's
very
very
easy
to
turn
on
a
feature
flag
and
we
want
to
make
sure
we've
captured
what
we
need
to
do
before
we
turn
it
on
before.
We
turn
it
on.
So
when
things
are
behind
a
feature
flag,
I'm,
very,
like
comfortable
with
oh
there's,
maybe
a
user-facing
issue,
that's
unrelated,
but
we
don't
have
to
take
care
of
it
here.
But
let's
create
an
issue
such
that.
A
So
all
this
behind
the
feature
flag,
but
then
it
also
kind
of
depends
on
projects
that
have
kubernetes
agents.
What
and
stuff
it's
like.
Okay,
this
is
some
setup
I'm
not
prepared
to
do
as
a
code
reviewer.
So
when
that
happens,
sometimes
I'll
I'm
still
trying
sometimes
I'll
test
things
locally,
and
if
I
can't
get
the
happy
path,
because
I
don't
have
some
setup,
then
I'll
either
just
rely
on
the
authors
environment
and
ask
them
questions
or
I'll
actually
comment
out
of
the
code.
A
Certain
checks
that
let
me
keep
going
forward.
So
if,
if
it's
trying
to
check
that
I
have
some
setting
that
or
some
integration
that
I
don't
have
I've
done
before
it
doesn't
happen,
often
but
I'll
just
comment
out
of
the
code
that
lets
me
keep
testing
whatever
UI
part
I'm
trying
to
figure
out
that
makes
sense
yeah.
So
let's
want
to
share
here
was
story
about
one
little
change
regarding
okay,
let
me
see
if
I
can
find
the
old
version
of
the
dev.
A
But
the
issue
is
that
we
need
we
need
to
reference
the
root
groups,
cluster
agents,
not
just
my
immediate
group,
so
I
could
be
like
three
nested
groups
in
and,
as
in
my
project
and
I
need
to
be
able
to
reference
the
root
group,
cluster
agents
and
there's
no
graphql
field.
That's
going
to!
Let
us
do
it
in
just
one
query:.
A
A
It
helps
to
put
things
in
some
abstract
box
of,
like
all
right,
I
I
see
how
we're
interfacing
with
this,
but
I,
don't
really
totally
know
all
about
it,
something
to
do
with
I,
so
now
I
actually
do
kind
of
know
about
it.
It's
something
to
do
with
the
communication
to
kubernetes
instance,
so
that
we
could
spin
up
workspaces.
A
That's
my
understanding
of
it.
Well,
so
previously
we
had
one
query:
now
we
have
two
queries
and
what
has
what
was
previously
happening
was
we
would
had.
We
had
one
Apollo,
smart
query
and
then
another
Apollo,
smart
query,
and
so
this
was
the
intent
was
when
the
first
get
project
details,
query
returned.
A
We
would
see
if
it
has
a
group
and
if
it
doesn't
so,
if,
if
it
doesn't,
then
we
can
emit
our
results
right
away,
but
if
it
does,
we
need
to
set
the
group
path
and
re-query,
which
is
going
to
then
trigger
reactively.
This
smart
query
and
we
set
the
group
path
by
calculating
what
the
root
group
would
be.
But
then
we
set
this
smart
query
to
run
and
then
once
we
get
the
cluster
agents,
then
we
can
emit
results.
A
A
A
I
could
see
some
scenarios
now
that
something
isn't
just
cohesive
and
something's
broken
up
into
some
steps.
You
can
run
into
some
interesting
situations,
particularly
like
a
race
condition,
especially
if,
like
in
during
the
race
condition,
while
things
are
changing,
one
of
them
errors
out
and
then
like
the
results,
could
be
from
a
previous
request
or
something
like
that
Enrique.
So
my
main
thoughts
when
I
run
into
something
like
this
is
like.
A
How
did
we
feel
about
this
and
how
we
thought
about
other
Alternatives
or
something
we
can
do
now
or
is
there
a
plan
for
it?
Like
that's
kind
of
my,
my
main
goal
is
like
okay,
we're
behind
a
feature
flag
to
see
what
we're
trying
to
do,
but,
like
maybe
there's
something
we
can
improve
here.
So
yeah
Enrique
was
concerned
and
he
even
sounds
like
well.
We
were,
he
was
talking
about
it.
A
A
A
Yeah,
if
one's
pinging
another
one
like
they,
they
can
get
out
of
sync
and
I
was
thinking
of
a
race
condition
and
one
of
them
failing
and
that
creating
some
sort
of
scenario,
but
I
knew
that
this
is
more
like
a
code
smell
than
an
actual
bug,
because
I
don't
think
you'd
be
able
to
trigger
this
using
just
the
UI,
but
I'm
not
entirely
sure.
The
conversation
continued,
though,
because
as
I
looked
at
it
I
then
identified.
Oh
no.
A
We
actually
have
another
legit
bug
with
the
two-step
smart
query,
where,
if
the
user
selects
a
project
that
has
a
group
we
admit
as
expected,
but
then
the
project
full
path
changes
to
a
project
which
does
not
have
a
group.
We
don't
ever
clear
this
old
cluster
agents
we
just
pulled,
and
so
they're,
including
the
Met
results.
A
A
But
we
have
a
method
now,
which
will
fetch
cluster
agents
based
on
the
group
path.
If
there
is
no
group,
we
just
immediately
return
empty
or
if
we
do
have
it
so
that
this
is
just
one
one
function
that
gets
triggered
reactively,
which
will
always
behave
the
same
of
we
get
our
response.
We
might
need
to
make
a
second
response,
but
then
we'll
admit
the
result,
and
this
way
we
also
don't
have
to
manage
the
state
which
was
I
think
causing
issues.
A
E
A
C
C
C
A
This
different
recently
well
I,
don't
think
it's
been
I
I
think
it's
been
around
for
a
little
while
yeah
I
think
it's
been
around
for
a
while,
but
we
also
have
the
bot
the
same
bot.
That
leaves
this
message
will
also
start
a
pipe,
but
maybe
it
doesn't
start
the
pipeline
on
the
community
contributions
so
yeah
these
predictive
jobs
run
on
unapproved
Mrs,
but
then
we
just
run
the
full
jobs
when
the
Mr
is
approved,
I
think
that's.
What's
going
on.
A
A
A
But
in
the
wild
this
string
can
actually
items
in
the
string
can
actually
contain
spaces
themselves.
So
I
felt,
like
we
kind
of
need
to
do
a
split
on
something
else,
wasn't
entirely
sure
what
so
I
had
a
crazy
random
idea
of
like
we
could
do
something
where
we're
actually
just
doing
like
us.
Somehow
Auto
populating
multiple
items.
A
E
A
I
can
I
can
sense
that
the
like
separators
are
hard
and
what,
if
you
have
it
like
what?
If
a
user
is
creating
this
entry
point
thing
which
is
supposed
to
be
like
a
command
line
argument,
and
so
actually
I
don't
want
to
enter
another
line.
But
maybe
I
want
to
pass
a
string
that
actually
has
a
comma
to
it.
I
see
regularly
like
anytime
you're
using
a
separator.
A
A
B
A
D
A
A
A
A
E
E
A
A
B
B
A
B
A
A
B
A
B
A
A
D
E
A
C
D
B
B
B
A
B
B
D
A
A
A
E
C
A
Through
other
things,
especially
since
I
and
that's
a
good
question,
especially
since
I'm
one
of
my
comments,
I
can
anticipate
it
potentially
resolving
to
where
this
all
looks.
Different
I
may
not
be
super
inclined
to
take
a
look
at
the
specs,
but
I
may
look
at
the
specs
to
make
sure
like.
Are
we
doing
some
sort
of
wildly
off
pattern
that
we
need
to
fix?
A
It
was
really
just
that
one
component
too,
but
yeah
there's
often
times
where
the
timer's
going
off
and
I
spend
all
my
time
on
one
part
of
the
Mr.
So
usually,
if
I'm
leaving
comments
where
I'm
not
sure
if
the
Mr
is
going
to
look
the
same
afterwards,
I'll
leave
I'll
I'll
go
ahead
and
submit
that
review
and
I'll
say:
hey
I
didn't
really
take
a
look
at
the
specs,
because
I
think.
Let's
resolve
this
thing.
First
and
I've
done
that
before.