►
From YouTube: 2023-01-04 Code Review Performance Round Table
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).
A
And
we'll
live
welcome
everybody
to
the
first
of
the
year
performance,
Roundtable
of
code
review
and
we'll
get
right
on
it.
So
before
we
get
into
the
topics
that
are
open,
there's
a
few
housekeeping
discussions
that
we
need
to
cover.
The
first
is
that,
since
we
moved
to
having
the
issue
board
and
discussions
moved
to
separate
issues,
the
participation
on
the
matter
issues
I've
been
creating
weekly
has
declined.
A
Obviously,
because
discussions
can
move
to
the
issues
themselves
and
I
see
that
as
a
bonus
as
a
positive
thing,
but
he
does
question
the
act
of
creating
the
issue
every
week.
A
So
I
was
wondering
if
we
should
just
discuss
this
a
little
bit
better
here.
What
should
we
do
about
those
issues?
Should
we
keep
keep
one
issue
and
don't
close
it
every
week?
Just
keep
that
for
like
catch
all
questions
and
stuff
like
that,
or
should
we
move
the
monthly
issues
or
just
stop
creating
the
issues
and
use
the
board
to
visualize
the
topics
being
proposed
by
other
people
or
any
other
options?
What
do
you
all
think.
B
Yeah
I
think
I
think
questions
those
kind
of
general
questions
that
were
going
in
there.
Those
could
be
in
the
agenda
too
I
don't
want
to
lose
track
of
them
like
for
transparency
purposes,
but
it
seems
like
the
things
that
we
were
discussing
that
aren't
in
those
issues
those
might
fit
in
like
an
agenda
or
this
recording.
A
And
that's
where
the
what's
been
happening
anyway,
yeah
I,
guess
right
any
other
thoughts.
A
So
I'm
guessing
you're
go
you're,
suggesting
Matt,
stop
creating
these
shoes
and
use
the
important
agenda.
C
I
think
only
the
only
thing
I
could
say
against
issues
in
favor
keeping
like
the
weekly
Roundup
is
just
the
there's,
a
perceived
amount
of
overhead
to
create
an
issue
yeah.
So
if
I
have
like
a
random
like
little
tiny
idea,
that's
really
only
a
sentence
to
too
long
like
is
that
worth
creating
an
issue?
C
There
really
isn't
any
any
real
work
to
create
an
issue.
Obviously,
but
but
you
know
psychologically,
there
might
be
a
hurdle
or
some
things.
A
That's
fair
I,
don't
opposed
to
keeping
one
running
issue
open
to
for
those
for
those
things
and
I
could
I
could
craft
one
finally
shoot
to
kind
of
like
cover
the
next
couple
of
months,
where
you
would
direct
people
to
open
issues
with
those
labels,
but
if
you
have
any
any
questions
for
us,
just
just
drop
it
on
this
issue
and
then
we
keep
it
running
and
we
stop
doing
weeklies
I
think
it's
the
most
like
it
covers
all
grounds.
It
doesn't
have
a
necessary
work
every
week.
A
He
doesn't
have
this
like
moving
Target
for
discussions,
because
every
week,
like
you
put
a
comment
there,
it
disappears.
You
put
a
comment.
There
disappears.
So
we
for
transparency,
reasons
having
one
ongoing,
probably
beats
best.
It
doesn't
have
any
overhead
people
just
decide
to
go
to
the
issues
themselves,
so
it
feels
like
the
option.
One
might
be
more
inclusive
for
everybody.
A
Okay,
so
I'll
go
forward
with
option.
One
I'll
make
that
my
bias
for
action
and
just
to
keep
this
moving
but
yeah
we
can.
Can
you
read
on
it?
Can
I
move
on?
Anybody
has
any
thoughts
before
we
move
on.
A
I
just
had
to
cough
okay
cool
the
next
one.
We
had
set
December
15
as
the
reassessment
date
to
see
whether
these
calls
were
being
useful
or
not.
A
A
A
Okay,
thanks
Matt,
okay,
objections;
they're
all
good.
B
A
All
right
receding:
that's
your
post
cool!
So
we
go
into
topics
then
so
the
first
one
is
something
I
believe
I
created
and
I.
A
We
briefly
touched
it
on
the
previous
call,
but
again
it's
also
more
front-end
topic,
but
I'll
just
cover
it
very
quickly
for
your
edification
So,
the
plan
team
is
playing
around
with
the
Persistence
of
Apollo
cache
queries,
meaning
that
if
you
do
a
query
for
labels,
for
example,
when
you're
in
an
issue
list
the
the
next
time
you
visit
that
page,
if
he
was
very
shortly
after
you,
don't
have
to
go
to
the
server
again
to
get
the
listed
labels,
the
initial
list
of
labels,
so
the
experience
is
much
snappier
and
this
is
in
the
context
of
a
bigger
effort
from
from
Tim
and
the
snappy
GL
kind
of
thing,
but
that's
been
covered
before,
but
in
this
particular
case,
they're
they're
improving
one
specific
use
case
that
one
of
the
labels-
and
this
issue
is
more
for
us
to
kind
of,
like
think,
is
there
anything
similar
that
we
could
benefit
from?
A
You
know
in
our
in
the
world
of
good
review.
Of
course,
the
most
obvious
one
would
be
labels
in
merge,
request
lists.
I
do
I'm
not
fully
aware
where
whether
their
work
would
benefit
merge
requests
as
well,
or
we
would
have
to
report
that
as
well,
but
I
think
about
the
merge
requests.
The
only
things
that
could
directly
benefit
would
be
stuff
about
that
are
reusing
graphql
for
reviewers
and
stuff,
like
that.
A
So
beyond
that
there
might
be
other
opportunities
for
us
to
benefit
from
this
kind
of
strategy
to
persist.
Data
I
think
they
were
starting
with
local
storage
but
then
potentially
iterating
later
to
indexeddb.
So
there's
a
couple
of
links.
There
I
see
that
the
emergency
requests,
the
draft
merge
request-
is
emerged
yet
I'll
catch
up
with
with
Natalia
about
this
soon,
so
that
we
can
see
where
they
are.
But
if
anything
happens,
that
I
feel
like
are
more
yeah.
A
There's
movement
there
if
I
see
anything
worthwhile
for
us
I'll
bring
it
up
again
here
in
these
calls.
But
at
this
point
no
further
specific
things
on
my
end.
Does
anybody
have
any
questions
on
it?
Foreign.
D
This
sounds
like
a
weird
question,
but
like
no
go
ahead
that
doesn't
feel
like
a
great
in
state
to
have
multiple
caches.
We
already
do
like
a
lot
of
back-end
caching
and
then
like
front-end
caching
and
then
like.
We
already
have
enough
challenges
and
validating
cash
in
one
spot
like
this
feels
like
it's
gonna,
create
more
pain
and
places,
and
we
like
we
do
have
a
lot
with
all
the
real-time
work
like
we're,
moving.
Basically
everything
into
graphql,
let's
interacted,
which
is
so
it
just
means
like
it
just
I.
D
A
Multiple
problems,
it's
mounting,
yeah,
it's
a
fair
question.
The
way
I
see
it
is
it's
used
sparingly.
We're
not
going
to
be
like
it's
going
to
be
opt-in,
so
you're
going
to
have
to
move
them.
You
have
to
migrate
the
queries.
You
want
to
be
persisted
yourself,
and
this
is
more
like
between
page
loads
and
some
other
times.
The
crews
are
very
user
specific.
So
even
maybe
maybe
the
back
end
isn't
even
caching
that
much
because
it's
very
user
specific,
but
for
us
in
the
browser
we
can
make
some
decisions.
A
The
labels,
I
think
was
the
most
obvious
one
where
we
open
many
users
up
in
the
the
merge
aggressive
issues
list
very
often,
and
that
initial
load
only
changes
when
there
are
label
changes
on
the
project
or
on
the
groups
that
you're
in
which
is
a
far
more
far
less
frequent
event
and
then
again,
if
you're
searching
for
one
specific
label,
you're
going
to
type
your
query
and
that
query
will
probably
not
be
cached
across
pages.
A
So
you
still
go
through
to
the
server
so
I
think
it's
just
being
careful
about
what
we
choose
to
persist
will
be
key
to
making
this
useful,
but
for
the
things
that
are
so
in
terms
of
performance,
seeing
the
labels
pop
up
immediately
when
you
revisit
the
page,
it
will
be
impactful.
So
that's
the
kind
of
questions
we
need
to
ask
ourselves
for
code
review.
A
What's
those
kind
of
actions,
the
other
thing
that
came
up
recently
was
the
the
at
mention
lookup,
having
a
way
to
reuse
participants
in
the
issue
or
emerge
requests.
A
A
Should
we
move
on
next
one,
okay,
so
the
next
one
is
split
and
paginate
the
file
tree
from
this
metadata,
and
this
has
been
covered.
A
couple
calls
here:
I'll
just
open.
It
again
share
the
screen
so
that
we
can
keep
it
in
the
recording.
A
A
We
have
a
spike
scheduled
for
this
Milestone
that
standards
live
we'll
be
looking
into
when
it
gets
back
from
PTO
about
validating
the
performance
benefits
of
having
a
separate
endpoint
and
what
are
the
limits
where
this
makes
sense
and
when
it
stops
being
beneficial,
beneficial,
we'll
be
looking
into
validating
that
POC
and
then
eventually,
we'll
follow
up
with
an
implementation
that
might
in
the
online
in
production.
A
But
apart
from
that,
I'm,
not
sure
if
there's
any
outstanding
topics
to
be
discussed,
does
anybody
know
anybody,
because
I
feel
like
we
can
remove
the
refinement
level
on
this
issue
and
stop
revisiting
it?
Every
week,
what
do
you
all
think.
A
Okay,
so
I'll
do
for
the
sake
of
efficiency,
I'm,
removing
the
requirement
there.
We
go
label
here,
as
it
seems
like
the
topic
is
being
pushed
forward
in
its
own
Empires
I
have
scheduled
work
and
then
Patrick
LTC
Patrick,
just
to
make
sure
that
you
can
understand
his
life.
Make
sure
that
we
in
case
they
have
something
outstanding,
yeah
that
actually
unlabel
the
refinement
there.
We
go
all
right.
A
One
less
removed,
refinement
label,
all
right,
cool
virtual
scroller
I
believe
this
would
require
it's
kind
of
stalled,
so
I'm
gonna
go
the
next
thing.
This
is
covered
if
you're
the
board.
A
So
it
feels
like
we
also
might
be
able
to
drop
it
off
from
this,
so
I'm
going
to
tag,
Spanish
live
and
fill,
and
we
what
was
this
issue
yet
since
we
have
issues
for
service,
does
anybody
have
any
questions
on
this
topic.
A
And
then
we'll
leave
it
at
that.
There
might
be
some
things
that
we
want
to
still
investigate.
A
A
Thank
you
all
right,
so
we
added
they
will
do
this
cool
so
we'll
we
can
skip
that
discussion.
I
guess,
there's
I
think
the
missing
parties
are
sorry.
The
relevant
parties
are
missing
from
this
call,
so
I'll
just
leave
it
up
to
them
to
kind
of
like
close
it
off,
which
brings
us
to
this
one:
suppressing
files
and
divs
through
this
overall
payload
I,
haven't
followed
much
on
this,
but
I
think
we
had
movement
recently
right.
A
Yeah
I
remember
that
that
your
point
was
whether
we
should
just
instead
of
suppressing
them
completely
from
the
back,
and
everything
will
be
just
like
collapsing
them
and
having
a
link
to
load.
The
full
thing
right.
A
Is
that
issue
created,
should
we
create
an
issue
for
it
and
potentially
schedule
it?
What's
your
thoughts
there.
D
It's
based
in
the
so
there's
a
it's
based
in
the
it
may
be
based
on
like
the
larger
issue,
but
it's
like
I
think
Carrie
and
I
were
talking
about
like
currently
you
can.
D
Currently,
you
can
use
git
attributes
and
label
something
as
a
binary
file,
and
then
we
will
not
render
a
diff
for
you
so
we're
debating
about
whether
or
not
like
or
not
to
be
we're
talking
about
whether
or
not
the
that
actually
reduces
any
payload
stuff.
I
hadn't
followed
back
up
yet
to
see
if
I
had
made
sense
of
it,
but
I
think
like.
D
Yeah
Carrie
had
suggested
like
well,
we
could
have
a
front-end
option
to
show
the
diff,
which
I
think
would
be
a
good
Improvement.
If
people
were
manually
suppressing
disks,
but
I,
don't
know
how
many
people
manually
suppress
diffs
like
I.
Don't
think
that's
a
common
thing,
because
I
don't
think
that's
actually
I,
don't
I,
don't
think
people
want
to
suppress
deaths,
they
just
don't
want
them
unless
they
want
them.
I
think
is.
D
Maybe
the
better
way
to
think
about
that
like
in
the
suppression,
is
like
the
good
attributes
when
it's
super
heavy-handed
like
if
you
say
text
files
are
binary
like
that's
it
we're
never
showing
you
that
file
as
a
diff
ever
and
there's
no
way
to
get
that
back.
Unless
you
go
check
in
like
a
new
get
attributes
file
that
doesn't
say
that
anymore.
C
Yeah
I
mean
binary
is
the
easy
one
right
because,
like
humans
can't
read
it
so
why?
Why
even
bother,
but
it
gets
trickier
when
it's
like
any
any
file
type,
could
be
considered
useless
and
I.
Think
I
think
part
of
this
issue
touched
on,
like
the
auto-generated
diffs,
like
very,
very
large
config
files
that,
like
yeah,
nobody
really
needs
to
review
that
I.
Just
worry
that,
like
we're
we're
one.
C
Everybody's
going
to
be
one
right,
one,
like
wow
I,
really
wish.
We
had
reviewed
that
issue
from
from
complaining
about
it.
D
I
like
how
others
handle
this,
and
that
like
it
is
suppressed
and
I
think
there's
I.
Think
from
my
perspective,
like
the
idea
of
making
this
a
performance
thing
is
like
if
we
could
suppress
in
the
sense
that,
like
we
don't
ever
go,
get
anything
from
the
back
end
unless
someone
explicitly
asks
for
it.
That
feels
like
a
savings
in
terms
of
like
not
asking
Italy
for
it.
D
Although
we've
got
to
ask
for
like
the
file
I
guess
to
know
that
it
exists,
but
like
we
don't,
if
we
don't
get
the
diff,
we
don't
highlight
it.
We
don't
do
all
those
things
like
it
feels
like.
There's
performance
savings
there
and
it
feels
like
there's
a
standard
set
of
files
that
people
generally
agree
don't
need
to
be
shown
all
the
time.
Maybe
if
they
want
to.
D
I
think
the
library
that
I
linked
is
like
sort
of
where
that
comes
from,
but
like
to
me
that
feels
like
we
gain
it's,
not
it's
performance,
because
we're
going
to
reduce
payloads
on
like
certain
types
of
merge,
requests
that
contain
certain
files
right.
Is
it
going
to
be
like
everything,
I,
don't
know,
I
doubt
it?
D
Is
it
going
to
help
with
a
lot
of
our
customers
who
have
emerged,
requests
that
changed
250
files
and
20
000
lines,
probably
they're,
probably
the
kinds
of
customers
that
benefit
from
something
like
this,
and
so
it
feels
like
we
get
something
there,
but
I
agree
like
it
has
to
come
with
the
front
end
piece
of
like
you,
click
load
diff
anyway,
and
we
go
and
we
do
all
that
rendering
and
then
there's
a
spinner
and
eventually
it
comes
back
and
loads
like
that
has
to
that
has
to
be
there
like
it
can't
be
pure
suppression.
D
It's
got
to
be
suppression
with
a
way
to
get
to
it,
but
yeah
I,
don't
I,
don't
know
that
I
like
investing
in
like
the
get
attributes
hack
because
I
think
that's
sort
of
like
a
weird
thing,
but
maybe
if
we
need
the
front
end
viewer
or
that
anyways
it
like
gets
us
that
part
covered
now,
I,
don't
know
I
think
that's
the
thought.
The.
C
We
can
point
them
Point
them
to
it,
so
not
to
say
that
it's
a
great
solution,
but
it's
a
you
know.
It's.
C
We
already
sort
of
have
the
structure
in
the
code
to
go
in
and
say:
yes,
the
don't
don't
show,
don't
don't
highlight,
don't
cash,
don't
show
don't
send
these
manually
specified
files,
we
can
just
add
we
can
add
on
to
those
Pathways.
Also
don't
do
these
these
ones
that
we've
determined
algorithmically
aren't
correct.
C
D
So
I
assume
it's
using
its
own
viewer,
then
right
like
if
it's
it,
because
those
those
files,
if
you
suppress
them,
say
this
diff
was
suppressed
by
an
entry
and
get
attributes
like
if
you
were
to
add
like
like,
if
you
could
change
that
and
like
put
a
line
below
if
it
says
load
anyway,
like
I
assume
we
already,
we
already
know
this
is
happening,
and
then
it's
just
a
question
of
what
is
load
anyway.
D
C
D
C
A
Yes,
I
have
something
non-differable
not
difficult,
and
then
this
is
the
message
that
we
show
so
yeah.
We
do
have
a
viewer
for
this.
We
might
need
some
another
flag
to
distinguish
this
from
whether
it's
a
filing
coding
or
not,
but
it's
actually
yeah.
We
can
also
add
a
copy
to
like
try
to
show
it
anyway
and
then
it
just
tries
to
do
it
anyway.
When
the
user
clicks
yeah,
we
do
have
a
viewer.
That's
answering.
D
A
We
have
the
infrastructure
yeah,
so
yeah,
okay,
not
necessarily
new
stuff,
just
the
different
Behavior
yeah.
So
that
seems
schedulable
for
the
upcoming
milestone.
D
A
Get
attributes
right
so
we
have
one
one
part,
which
is
the
flag
that
tells
us
to
render
this
viewer
and
the
other
is
getting
the
actual
content.
We
already
have
that
on
some
files.
I
can't
remember
what
it
was.
The
moved
file
that
we
added
an
option
to
render
the
file
anyway
or
something
I
can't
remember
exactly
what
it
was
but
yeah
the
end
point
that
you
talk
about
it's
it's
like
the
Unicorn.
We
know
we
think
it
exists,
but
we
haven't
seen
him
in
a
while
in
a
while.
So
there's
also
yes,.
C
But
also
I
think
we
would
need
something
in
the
back
end
too
and
I'm
just
saying
this
out.
So
it's
real
to
to
kind
of
bust
through
that
special
casing,
because
otherwise
we're
going
down
the
same
Pathways,
it
would
say
it
would
fetch
the
fetch
it
from
Italy.
We
would
process
it
to
do
caching
and
highlighting
and
say:
oh
no,
it's
get
attributes
prohibited
and
then
we
would
never
right
back.
So
we
would
need
a
direct
like
a
direct
line
to
get
a
link.
Basically,
at
that
point,.
A
So
what
I,
what
I'm
reading
and
from
what
I
see,
I
think
we're
missing
an
issue
for
this
issue.
D
D
No,
we
probably
would
need
a
different
issue,
because
I
think
that
one
is
that
one's
more
broad
because
it
wants
to
like
Auto
collapse
things.
It's
certainly
related
to
that
concept
like,
but
this
isn't
going
to
be
Auto
collapsing.
This
is
going
to
be
like
user
initiated
suppression.
D
D
In
theory
like
and
we
do
something
that's
smarter
and
like
is
auto
suppressed
based
on
standards
like
people
need
to
go
and
unwind
all
of
their
git
attributes,
sort
of
like
weird
suppression
and
I,
don't
know
what
the
it'd
be
it'd
be
good
to
know
like
what
the
downstream
effects
are
of
labeling
telling
git
that
a
bunch
of
stuff
is
binary
like
what
happens
in
other
diff
tools
and
other
editors
like
what
happens
when
you
try
and
open
these
things
like
across
the
board?
How
how
much
does
other
stuff
complain
or
freak
out.
C
A
C
A
B
C
C
Should
always
look
to
ourselves
to
be
that
person
right
first,
but
you
know
if
we
had
I'm
sure
our
customer
in
line
that
we
could
say
hey.
Would
you
mind
doing
this
I,
don't
think
it
would
I,
don't
think
it
would
mess
up
any
other
tools.
C
Okay,
especially
if,
if
we
think
that
we're
going
to
do
an
iterative
solution
later
on,
that
is
not
get
attributes.
That's
maybe
you
know,
or
it's
its
own.
It
could
be
its
own
namespace
in
the
attributes.
File.
I,
don't
know
if
we,
if
it's,
that
extensible
I've
never
really
played
around
with
attributes,
but
if
it
is
maybe
like
a
git
lab,
Dash
suppressed
diff
or
something
like
that
or
if
we
just
need
to
build
a
UI
for
it.
At
some
point.
A
I
I
got
to
the
issue
we
you
were
discussing
yesterday
so
either
put
that
forward
or
or
create
a
separate
issue.
Kai,
do
you
want
to
take
that
action.
D
D
Maybe
we
mention
it
in
that
issue,
because
there
are
people
in
that
issue
that
are
potentially
doing
it
see
what
kind
of
feedback
comes
from
it
after
we
roll
it
out
and
if
it
works
or
doesn't
work
like
that
may
be
a
way
to
test
it,
because
I
think
I
would
love
to
dog
food
this
in
some
way,
but
I
don't
know
that
we
have
any
projects.
D
I,
don't
know
if,
like
yeah,
only
people
that
might
could
be
like
the
in
for
our
delivery
team,
like
it'd,
be
worth
checking
with
them,
potentially
because
they
they
would
end
up
in,
like
the
the
generated
code,
realm,
I,
think
of
cookbooks
or
whatever
they
call
them.
I,
don't
know
what
they
use
now
when
it
was
Chef,
it
was
cookbooks
I,
don't
know
what
it
is
now
so
like
they
might
end
up
in
that
in
that
Realm.
A
Okay,
so
separate
issue
seems
to
be
the
way
to
go.
Do
you
want
to
take
care
of
that
yeah
I'll
create.
D
A
Okay,
putting
you.
A
Right
cool
right,
so
there's
no
other
open
topics,
there's
still
things
ongoing
that
are
scheduled
or
or
in
the
pipeline
to
be
scheduled
something
so
we
don't
really
have
to
discuss
them
here,
but
does
anybody
have
any
other
thing?
That
is
in
the
back
of
your
mind
that
you
want
to
talk
about
before
I
move
on
to
some
leftovers.
A
I
know
there's
some
things
flying
about
HTTP,
caching,
that
I
think
Patrick
is
assigned
or
working
on,
but
those
seem
to
be
oh
by
the
way
coping
me
yesterday.
A
Okay,
are
about
the
okr
for
this
POC
things
we
need,
so
we
set
ourselves
to
create
two
pocs
and
validate
three
ideas
feel
free
to
take
a
look
after
the
call.
So
if
you
have
any
thoughts
or
suggestions
to
add
there
and
I
might
have
missed,
feel
free
to
add
them.
This
is
for
us
on
the
call
and
for
everybody
watching
the
recording,
because
I
know
you'll
be
watching.
A
All
right,
so
this
is
something
that
has
been
the
back
of
my
mind
and
it's
very
selfishly
something
that
I've
pushed
in
the
past
Kai.
You
know
this
has
popped
up
frequently
in
the
past
and
I.
Just
it
just
reminded
me
new
year.
I
just
wanted
to
make
sure
that
we
revisit
this
and-
and
we
know,
what's
the
blocker
at
this
point
for
rolling
this
out-
is
it
not
enough
representation
of
users
caught
being
caught
in
this
flow?
Is
it
a
ux
concerns?
A
Can
we
just
go
over
quickly?
These
two
things
to
see?
What's
the
hold
up
in
case,
they're
fully
blocked
their
block,
but
in
case
they're
unblockable
we
could
probably
take
some
time
to
to
unblock
them.
So
the
first
one
is:
don't
load
all
the
diffs
ahead
of
time
when
using
show
one
file
at
a
time
option
in
a
mark.
A
So
this
is
something
that
is
a
leftover
from
our
primary
implementation
of
the
show
one
file
at
a
time
we're
still
batch
diffing
as
usual,
but
this
basically
incurs
in
the
memory
costs
that
we
have
for
the
full
for
the
full
diff
which
stops
US
stops
the
show
one
file
at
a
time
from
being
an
actual
alternative
for
really
large
Mrs.
It's
still
going
to
use
the
memory
of
the
browser,
despite
not
rendering
the
components,
there's
still
a
lot
of
data
being
stored.
A
We
have
other
things
working
towards
decreasing
that
memory
usage,
but
wondering
do
you
remember,
do
we
know
exactly
what's
blocking
this
from
moving
forward,
it's
being
labeled
as
a
stretch
and
it's
kind
of
like
blowing
over
from
Mouse
into
milestone?
A
It's
not
that
heavy
either.
The
only
thing
that
this
would
change
is
when
you're
navigating
to
a
different
files.
You
probably
have
a
little
weight.
Oh,
we
might
have
that
pending
API
thing
to
get
the
file
actually,
so
we
can
do
some
investigation
around
that
I
guess
any
other
things
that
we
know
about
for
this
issue.
C
A
Yeah
see
this
is
the
the
Unicorn
we
were
talking
about
earlier,
oops
Yeah,
so
we
were
talking
about
the
last
thing
that
this
issue
has.
Is
this
comment
from
me
for
trying
this
URL
exactly
with
the
path?
I
did
some
reverse
engineering
on
the
Ruby
code
and
try
to
identify
exactly
how
this
worked
passing
path,
passing
paths,
passing
path
with
the
parenthesis
and
stuff
with
brackets
or
whatever
you
call
it,
and
none
of
it
worked.
A
So
if
we
can
figure
that
out
whether
we
do
have
that
option
and
we
just
just
a
problem
of
front
end.
Looking
at
Ruby
code,
the.
A
Right
exactly
so,
this
returns
empty
and
then
even
if
I,
you
know,
if
I
remove
the
thing.
Oh
so
not,
but
that's
plural
should
be
an
array
yeah.
It
doesn't
even
accept
it,
so
it
does
come
up
with
empty.
So
there
might
be
something
I'm
doing
wrong
for
sure.
So
I'm
going
to
do
this
I'm
going
to
leave
this
comment
here
and
then.
A
A
D
Yeah
I
was
looking
into
the
numbers.
The
numbers
and
I
need
to.
D
There's
two
metrics
and
I
I:
don't
know
how
to
distinguish
them.
One
is
really
low,
like
in
the
like
tenths
of
a
percent
of
users,
use
that
mode
sort
of
range.
D
So
if
that
is
the
case,
then,
like
you
know,
I
think
that's
where
this
is
like,
potentially
not
a
worthwhile
thing
to
spend
time
on
right.
If
it's,
if
that
is
the
number
that
I
think
it
is,
then
okay,
but
I
can
confirm
that,
for
you,
I
think
I
think
the
idea.
A
Then
that
brings
us
to
the
question:
do
they
not
use
the
mode
because
it's
still
slow,
Jeremy
and
and
yeah?
We
should
probably
check
the
numbers,
because
there's
one
thing
which
is
enabling
the
mode
and
the
other
is
loading
the
diffs
in
that
mode,
I'm
sure
we're
tracking
it
properly,
but
the
two
different
things.
A
Okay,
but
again
it's
a
two-point
way:
I
I.
We
can
still
take
care
of
it
outside
of
deliverable
work
in
terms
of
like
performance
tidying
up
again.
This
isn't
like
world
changing
improvements,
but
it
could
be
something
worth
like
closing
the
loop
one.
So
then,
when
we
have
capacity
again
we're
still
looking
for
insights
from
backend
once
we
have
that
unblocked
it,
this
would
be
a
fairly
walking
apart
if
I
expect
I'm
moving
on
to
the
next
issue.
A
For
the
sake
of
time,
then,
which
is
similar
but
not
exactly
to
be
supplies
to
all
modes,
but
so
this
is
similar
to
something
I
think
I
described
in
the
issue,
which
is
to
selectively
load
the
Mr
diff.
A
When,
following
the
comment
link-
and
this
actually
applies
to
any
link,
if
you
fall
into
a
link
to
a
file
or
something
we
don't
have
to
load
all
the
other
files
to
render
that
particular
file,
these
strategy
is
used
in
other
websites
like
insure
and
stuff
to
to
make
it
get
you
to
the
content
you're
pointing
at
as
quickly
as
possible
and
depending
on
the
Mr.
A
These
can
have
immense
user
experience
impacts
because,
if
you're
pointing
to
the
last
file
on
the
batch
divs
you're
going
to
have
to
wait
for
all
the
bad
shifts
to
finish
before
you
jump
through
that
file.
If
we
only
select
that,
if
you
only
load
that
file,
then
it's
immediate
right,
I
think
it
might
be
hung
up
on
the
back
end.
Endpoint
paths
thing
that
we're
just
talking
about
as
well,
but
there's
also
here
the
ux
concern
I
believe
right.
Kai.
D
A
But
because
we
don't
have
the
endpoint
yeah,
oh
okay,
if
the
input
is
sorted
out,
then
these
files,
then
these
issues
getting
blocked.
I'm
more
concerned
about
the
ux
of
it
like
you,
jump
in
you're,
landed
on
an
MR
that
has
50
files
and
we're
only
giving
them
one
making
sure
that
we
tell
them
and
make
it
absolutely
clear
that
you're
only
seeing
a
selection,
a
subset
of
the
DMR
is
key.
So
that's
where
I
think
the
UI
there's
a
figma
file,
I.
C
D
Design
stuff
is
like
could
probably
be
Revisited
and
just
make
sure
that,
like
everyone's
content
with
it,
but
I
think
that
that
I
don't
know
a
year
ago,
Phil
basically
had
the
same
concerns
that,
like
we
don't
know
how
to
get
you
the
right
file.
So
we're
not
actually
going
to
save
you
anything
yet
because
we
can't
right.
We
can't
do
that.
D
If
there's
an
endpoint
here,
I'd
also
say
like
looking
at
that
one
it'd
be
worth
revisiting
the
that
that
issue
in
the
context
of
the
restructuring
effort,
because
the
restructuring
effort,
sort
of
changes,
how
we're
thinking
about
where
comments
and
diffs
live
together,
anyways,
and
so
this
might,
it
might
not
exist
in
this
way
and
like
so,
it
could
be
like
not
a
not
a
good
effort
thing
if
we're
gonna
blow
it
up.
Anyways.
A
Yeah,
it
does
make
sense
so,
but
yeah
I
think
I
feel
like
if
anything
that
endpoint
will
be
useful
anyway,
because
even
so
the
restructuring
I
don't
expect
this
to
redo
the
entire
way.
Divs
are
loaded,
but
we
might
iterate
on
the
ones
we
have,
which
means
that
we're
probably
going
to
need
that
anyway.
So
I
feel
like
that's.
The
one
thing
that
I
take
out
of
these
two
discussions
is
yeah.
Let's
try
to
find
what's
happening
with
that.
A
Endpoint
and
probably
there's
a
bug,
probably
try
to
fix
it,
because
it's
going
to
be
useful
for
us
in
the
future
anyway.
Matt
do
you
want
me
to
create
an
issue
for
him?
Looking
into
the
specific
paths
thing
like
yeah,.
A
All
right,
potentially
fixing
it
sweet,
there's
no
other
points
in
the
agenda.
Anybody
has
any
other
points
to
discuss
before
we
go
off.
A
All
righty,
then,
thank
you
so
much
for
making
the
call
and
I'll
see
you
around
hi.
Everyone
thanks.