►
From YouTube: YUI Open Roundtable August 22, 2013
Description
YUI Open Roundtable August 22, 2013
A
Welcome
everyone
to
hey,
welcome
everyone
to
our
latest.
Why
do
I
open
around
table
today's
Thursday,
August,
22nd,
2013
and
we're
all
still
alive
and
I
want
to
welcome
everybody?
We've
got
a
number
of
items
on
our
plate
today,
I'd
like
to
thank
tilo,
read
and
dario
for
coming
today.
You
guys
are
awesome.
A
B
Sure
so
we
can
kick
it
off
with
button
so
as
we've
discussed
recently
and
other
Ron
Davis,
as
well
as
on
Yui
contrib,
there's
going
to
be
some
efforts
going
forward
to
make
button
awesome
because
right
now
there's
a
number
of
bugs
and
architectural
issues
that
kind
of
prevent
it
from
seeing
or
the
removal
of
the
data
tag.
So
yeah
I'd
really
like
to
get
that
stuff
taken
care
of,
and
one
of
those
things
is
the
architectural
issue.
B
Css
platen.
So
one
of
the
things
that
I
did
want
to
talk
about
was
kind
of
resolving
the
differences
and
similarities
and
kind
of
where
CSS
button
and
pure
button
bit
together.
So
I
guess
just
to
quickly
recap:
CSS
button
is
a
CSS
module
in
the
library
that
currently
yeah
it
was
designed
to
drop
it
one
CSS
file
onto
the
page,
and
then
you
yeah,
then
you
add
on
whatever,
whichever
skin
you
choose
into
your
India
body
class
and
from
there
you
can
it'll
have
both
skins
in
one
CSS
file.
B
Well,
oh
probably
that
does
present
some
issues
in
terms
of
additional
skinning
and
having
to
overwrite
some
of
the
some
of
the
options.
If
you
want
to
yeah,
if
you
want
to
customize
it
a
little
bit
more
so
but
take
a
deeper
dive
into
that
and
probably
make
that
more
along
the
lines
of
the
traditional
Yui
skinny
model,
which
is
separating
the
modules,
the
skin
mod,
the
skin
modules.
B
D
Yeah,
so
when
I,
when
I
wrote
a
pure
button,
I
did
take
a
bunch
of
styles
from
CSS
button.
So
there's
definitely
some
similarities.
I
think
the
right
way
to
go.
I
was
talking
Eric
really
briefly
for
a
while.
D
Just
now
is
like
you
were
saying,
you
want
to
split
out
CSS
button
into
maybe
a
core
and
ask
ennis
skin
and
I
think
the
skin
part
can
be
the
stuff
that
applies
kind
of
the
gradients
and
that
3d
ish
effect
that
CSS
button
has
in
fact,
looking
at
this
year's
button
source
code,
you
already
gonna,
have
that
done,
like
I
can
see
like
inside.
Why?
What
everybody?
D
You
have
structure
in
your
presentation,
so
I
think
it
would
be
useful
to
just
pull
in
pure
button
for
the
core
and
I
think
you
should
be
able
to
apply.
I.
Think
all
of
this
stuff
that
you
see
here
inside
CSS
button
should
also
be
inside.
Anything
structural
is
also
within
pure
button,
and
then
you
can
have
your
separate
skin
file,
which
is
separated
into
Sam
and
night
and
just
apply
the
stuff
that
you
see
under
presentation.
Yeah.
B
C
You
remove
some
of
them
for
size
and
concerns.
I
gotta,
look
at
see
which
ones
you're
talking
about
off
the
top
of
my
dog.
A
D
Of
moving
more
toward
the
pure
frame
right
mark,
Oh,
Eric
and
I
are
waiting
for
Bauer
to
finish
them
with
their
registry
work,
after
which
we
can
install
pure
via
our.
So
that
will
make
it
really
easy
for
the
why
you
watch
Q
integration,
because
you
could
Bower
install
a
pure
modules
and
use
those
as
a
base.
Until
then,
what
you
could
do
is
yeah.
So
if
you
were,
let's
say,
I
actually
have
a
little
CSS
parser.
D
That
I
wrote
that
you
can
use
to
just
pull
in
pure,
via
you
know,
via
the
CDN,
and
then
just
alter
the
class
names
to
change
pr2
Yui
through
YY
three,
and
then
just
use
that
as
your
base.
You
can
add
that
to
like
a
grunt
file
or
something
and
then
and
then
write
your
skins
on
top
yeah.
There
might
be
some
differences,
but
I,
don't
think
there's
much
in
terms
of
structure
if
I
didn't
remove
stuff,
it
might
be
because,
like
like
I
couldn't
run
it
without
it.
B
B
C
But
oh
yeah,
okay,
so
yeah
just
go
to
a
pulp.
D
Or
not,
you
want
to
go
to
Yui
dot
yahoo,
api's,
calm,
/,
pure
/,
02,
01,
/,
weapons,
dot,
CSS
and
battle.
Have
everything
that'll
have
core,
along
with
this
CSS
like
the
minimal
skin?
That
Pierre
has
and
you
should
use
this
problem.
I,
don't
know.
Let
me
check.
Let
me
look
into
core
to
see
what
core
I
was
saying.
You
could
use
buttons
dot,
CSS,
the
one
you
see
there
and
then
apply
use
that
as
the
core
and
apply
your
skin's
to
that
yeah
cool
yeah.
Oh.
B
D
I
also
have
a
pull
request
to
bring
button
CSS
button
groups
in
two
buttons
that
might
lend
like
for
the
next
release
of
pure
musher.
But
then
you
can
have
CSS
button
groups.
I,
don't
know
if
I
think
that
we
useful
to
have
the
CSS
only
button
as
well,
because
you
can.
B
Certainly
so
then,
basically
pures
buttons
are
ya.
Here's
buttons
are
the
the
Yui
buttons
explore
the
yeah,
the
Yui
CSS
buttons,
except
for
just
the
classes,
are
named
differently
and
we
can
easily
just
have
that
compiled
different
run
that,
through
a
compilation
step
to
to
rename
the
classes
yeah.
But
your
initiative,
the.
A
B
D
D
I,
don't
know
button
has
so
right
now,
I
don't
know,
but
has
it
have
you
been?
Have
you
worked
on
button
over
the
last?
Maybe
two
releases,
two
or
three
releases,
because
I
don't
think,
there's
been
much
that
has
changed
and
what
I
initially
made
but
I
kind
of
made
it
so
that
it
had
most
so
right
now
you
can
have
a
button
which
has
you
know
like
a
glyph
icon
or
something
like
an
icon
font
or
something
inside
it
and
stuff
like
that.
D
So
knowing
that
you
can
do
is
I
think
with
the
or
not
that
you
can't
do
but
I
think
that's
the
Yui
3
when
I
saw
an
example
which
has
like
a
picture
within
a
button
like
the
Y
way
three
logo
or
something
within
a
button.
I
haven't
touched
it
around
with
anything
like
that,
but
most
anything
you
should
be
able
to
do.
I
mean
you
can
wrap
anything
inside
a
button
element
right.
B
Yeah
so
the
more
complex
types
of
buttons
that
I
think
people
are
wanting,
which
would
give
Yui
button
why
you
had
threes
buttons,
more
feature,
parity
with
Yui
too
yeah
that
might
come
at
some
point.
I,
don't
know
it
hasn't
been
much
of
a
priority.
D
Arm
you
know
and
then
to
go
from
there.
The
other
thing
is:
there's
a
I've
been
working.
I
haven't
actually
started
work
on
this,
but
one
of
the
things
in
the
pull
request
people
want
with
in
pure
which
might
be
related
to
button
is
button
like
they
want
drop
downs,
so
drop
downs
that
aren't
purely
really
like
purely
not
isn't
related
to
your
butt,
just
like,
as
the
word
well
drop
downs
that
aren't
related
to
buttons,
but
you
know
an
element
which,
on
click,
can
reveal
a
drop-down
menu.
D
That
element
could
be
a
button.
That
element
could
be
like
part
of
a
wideout
menu
could
be
anything,
but
it's
something
that
opens
up
a
drop
down
menu.
Essentially,
so
then
you
can
put
soda
bootstrap
doesn't
really
in
some
ways,
so
I
don't
know
once
that
lands
in
pure.
D
D
Eric
is
the
one
who's
looking
into
that,
primarily,
but
from
my
understanding
and
my
discussions
with
them,
we're
basically
waiting
for
the
bower
stuff
to
do
a
full,
fledged
integration,
but
for
something
like
botton,
which
is
pretty
small
and
self-contained
I,
don't
know
it.
D
D
B
Mean
really
it's
just
grabbing
that
renaming
the
classes,
there's
probably
seven
or
eight
classes
on
there
renaming
those
and
then
putting
that
into
a
core
CSS
skin
and
then
creating
the
the
addition
and
splitting
up
the
skin
code
into
two
separate
files.
So
for
at
least
button
it
should
be
pretty
simple
to
to
do
but
yeah
any
other
kind
of
more
complex
ones.
If
we
wanted
to
start
looking
at
leveraging
here
a
little
bit
more,
then
it
could
get
a
little
bit
more
tricky,
but
in
button
that's
not
that
big
of
a
deal
so.
A
B
Yeah
I'll
have
to
think
a
little
bit
about
a
little
bit
more
about
how.
How
exactly
will
yeah
kind
of
do
do
the
development
for
it,
but
it
shouldn't
be
too
anything
too
complex
at
all
so
yeah,
so
that
was
kind
of
about
it
with
button.
Unless
anybody
else
had
questions
or
no
I'm
nothing.
A
A
G
F
It
might
look
obvious
kind
of
right
away
to
some
people
who
work
with
Yui
already
even
folks
are
in
public,
but
yeah
we're
not
inside
of
yahoo
right
now,
but
we're
working
on
basically
the
last
really
couple
of
years
we've
been
really
trying
to
get
great
test
coverage
or
everything
that
we
do
on
the
YY
project
and
especially
Yui
itself,
and
especially,
you
know,
showing
our
commitment
to
testing
on
all
of
our
target
environments
by
creating
that
infrastructure
or
it
doesn't
exist
or
working.
F
It's
also
hard
for
us
internally,
because
the
the
tools
that
we
have
at
our
disposal-
don't
really
necessarily
I,
like
the
kind
of
information
that
our
engineers
and
contributors
need
to
see,
and
so
we're
working
on
is
something
that
will
answer
that
question
not
just
for
people
at
Yahoo
but
any
contributor
to
the
Yui
open
source
projects
that
use
tests
right.
And
so
what
this
is
is
a
view
into
like
kind.
B
H
F
The
code
on
various
branches
of
the
white
light
project
and
it's
all
organized
by
commit,
which
is
I,
think
pretty
unusual
for
most
CI
systems
like
Travis's,
that
they
all
Travis
Jenkins
you.
They
all
go
by
my
build
numbers
right.
They
go
by
like
kind
of
overtime
and
instead,
what's
more
useful
for
us
to
look
at
things
more
from
a
commit
standpoint.
F
So
you
can
see
the
help
of
commits
as
bills
if
you
take
a
long
time
complete
and
give
a
full
picture
of
what
the
status
of
a
particular
commit
is,
and
so
this
kind
of
this
thing
right
here-
we
can
actually
see
just
by
looking
at
colors
that
are
inverted
on
the
stable
failures
column.
I
can
see
that
right
now,
def,
3x
and
death
master
had
for
failures
that
doesn't
exist
on
the
release
branch
and
that's
something
that
it's
actually
today
we're
working
to
find
the
right
way
to
fix
that
right
away.
F
Basically,
and
it's
given
us
a
lot
of
visibility,
just
in
the
last
few
days
since
this
tool
is
available
internally
and
as
we
get
more
people
push
anything
externally,
more
external
committers,
it's
going
to
be
important
that
this
tool,
this
view
is
available
to
people
in
the
public
as
well.
So
while
this
isn't
yet
available
to
the
public,
it's
definitely
a
part
of
its
design
that
this
is
going
to
be
something
that
will
kind
of
show
all
the
hard
work.
That's
gone
internally
into
testing,
I,
gotta
hope
the
phone.
F
Yeah,
so
so
this
is
an
overview.
I
can
go
and
look
at
I
won't
take
too
long
here,
but
basically,
you
can
go
in
to
look
at
the
whole
status
of
branch
and
I
can
further
go
into
clicking
on
to
a
commit
to
see
where
what
tests
have
been
failing
right
and
then
showing
metrics
for
the
entire
run.
So
some
things
that
are
missing
here
is
like
we
classify.
Actually,
this
is
something
that
may
not
be
actually
shouldn't
be
familiar
with.
F
Anyone
have
who
hasn't
been
watching
our
internal
built
system
so
for
people
who
are
looking
at
this
for
the
first
time
they
actually
classify
as
we
improve
our
testing
the
class
by
some
tests
as
known
stable
and
some
tests
as
known
flaky
as
an
they
can
fail
for
reasons
that
are
caused
by
something
other
than
busted
in
the
last
minute
right.
And
so,
as
we
find
these
things,
we
are
kind
of
like
working
to
bring
the
stuff.
F
That
it
was
caused
by
a
change
last
commit,
so
you
know
we
could
improve
this
by
showing
like,
as
we
bring
up
these
unstable
tests
and
encrypted
stable
and
actually
take.
The
city
is
safe
for
our
functional
tests,
arcelik
tests,
which
are
also
being
ran
and
reflected
here
like
in
the
totals
like
we
could
be
showing
those
as
well
so
that,
as
you
know,
so
that
we're
also
paying
attention
to
getting
our
to
our
goal
of
zero
flaky
component
too.
F
So
yeah
like
this
is
where
these
kind
of
numbers
confirm
that
we
want
to
ideally
run
a
hundred
thousand
tests
for
commit,
and
it's
true
once
we
have
once
we
get
to
a
completed
run
here,
we
run
over
a
100,000
and
sometimes
more
than
that,
if
we
run
by
saying
bill
twice
that
kind
of
thing.
So
so
this
is
something
yeah
it's
coming
soon
to
the
public
as
soon
as
I
can
possibly
get
it
out.
I'm
working
with
a
lot
of
different
teams
internally
that
have.
F
Sauce
labs,
if
sternly,
has
also
helped
us
because
they
provide
some
of
these
environments.
Some
of
these
are
internal.
These
all
running
on
on
inside
yahoo
easing
and
it
wouldn't
be
possible
without
all
this
infrastructure
support
that
are
tiny,
teen
gets
so
so
yeah.
We
want
to
show
this
so
that
anyone
who's
committing
to
why
you
I
can
watch
their
push
and
make
sure
that
their
changes
don't
introduce
new
regressions,
so
is.
A
F
Possibly
yeah
so
we're
really
focused
on
getting
this
working
well
for
for
us
and
then
what
we're?
What
I'm
planning
on
doing
personally
is
working
with
a
few
other
teams
at
Yahoo,
especially
open
source
projects.
Let's
say
when
he
does
a
logical
one
and
then
hopefully
another
open
source
project
at
Yahoo
and
working
with
them
right,
so
that
we,
instead
of
putting
this
kind
of
throwing
this
out
there
right
away.
I
want
to
make
sure
that
this
works
pretty
well
before
yeah,
making
this
available
or
nothing
crazy.
So
there's
nothing!
F
F
Was
the
most
see
I
systems
will
yeah,
they
run.
You
know
they.
If
you
look
at,
you
know,
say
Jenkins
right
see.
A
list
of
you
know
builds
at
our
brand
and
stupidly
useful.
You
have
one
CI
project
for
a
1c,
I
kind
of
project
or
your
project
right
for
your
repository.
It
works
pretty
well
and
you
probably
don't
need
something
this
complicated,
but
for
us
we
have
a
12,
see
I
like
12
Jenkins
jobs.
F
If
you
will
like
12
Jenkins,
you
know
separate
things
that
are
can
all
run
at
the
same
time.
That
in
turn
has
many
different
browsers
right
and
what
our
engineers
care
about.
What
our
committers
care
about
and
contributors
care
about
is
seeing.
F
Then
my
tests
fail
and
a
particular
environment
did
all
of
the
environments
that
ran
on
succeed
and
since
you
know
it
could
be
that,
even
though,
if
you're
looking
at
Jenkins
and
you
looking
at
12,
build
looking
at
a
going
to
quote
dashboard
mate
for
those
you
can
see,
everything
is
blue,
but
you
don't
know
if
there's
a
new
commit
that
just
came
in.
That
means
that
really
the
not
everything
is.
It's
all
now
and
validated
because
there's
a
new
commit
on
a
branch,
and
so
that's
that's.
F
Those
are
all
shortcomings
that
were
hoping
that
this
tool
will
highlight
right
like
as
soon
as
committed
as
me.
We
can
see
an
entry
in
it
here.
It
gets
its
own
link
that
can
be
shared.
That
will
show
love,
for
it
will
just
show
that
there's
not
all
environments
are
completed
yet,
and
no
environments
are
testing.
Is
that
builds
up?
As
you
know,
things
are
in
progress
that
also
can
get
highlighted
here
as
well.
F
Right
as
so
at
you
externally
can
have
visibility
into
seeing
that
right
now,
these
three
browsers
are
being
tested
simultaneously
and
those
things
will
be
coming
in
soon
right,
so
you
can
actually
sit
here
and
watch
your
commit
get
tested
as
it
goes,
is.
F
Based
on
amount
pending
bill
as
well
as
we
did
build
yeah
yeah,
it
doesn't
do
that
right
now
and
there's
no
example
of
that.
But
that's
that
is
what's
coming
up
next
for
this
thing,
so
so
yeah
we
run
iOS
5,
6
ie6
through
10
m
34
and
OS
10
Safari,
so
yeah,
it's
a
in
Chrome,
yeah,
firefox,
so
node
node
is
doesn't
show
up
here,
but
that's
also
tested
yet
so
getting
node,
0
da
taken
note,
00
10
on
here
is
another
thing
to
do
list
for
me,
so
yeah
yeah.
F
So
this
is
all
like
really
really
new,
and
you
know
it's
kind
of
solves
the
problem
that
we've
had.
You
know
we
have
really
ambitious
goals
for
testing
and
it's
not
always
really
clear.
Hopefully
this
is
this
is
the
first
step
to
getting
something
that
solves
these
goals,
and
so
I'm
excited
to
share
this
with
with
anyone
who's
watching
the
hangout
on
air
and
feel
free
to
get
ahold
of
me
on
IRC
I'm,
always
sitting
in
the
channel
has
green
so
feel
free
to
ping
me
there.
F
A
You
I
think
it's
really
amazing
that
you're
you're
taking
something
that's
very
complicated
and
can
be
very
involved
in
making
it
simple
and
that's
a
very
challenging
task.
I
mean
you
can
look
at
this
and
go
oil
isn't
much
going
on
there,
but
behind
the
scenes,
there's
a
lot
that
you're
choosing
on
the
show
as
much
as
you're
choosing
to
show
yeah.
F
Yeah
so
right
now,
I
mean
we
have
already
we've
been
using
this
tool.
This
tool
been
collecting
dated
for
about
two
and
a
half
weeks,
and
we
already
have
almost
almost
three
million
data
points
that
have
to
fit
all
come
together
in
our
summarized
here.
In
pages
that
you
know
have
you
know
puppet
show
quickly
and
summarize
all
this
data
into
something.
That's
really
easy
to
see.
Sad
sup,
so
yeah,
so
stay
tuned
more
to
come
as
this.
This
is
only
the
beginning
for
this
stuff.
B
Now
it's
super
awesome.
I
just
like
the
like
the
idea
that
all
of
the
our
testing
infrastructure
and
reporting
isn't
all
in
the
yahoo
side.
That's
out
there
for
all
contributors
and
committers
to
be
able
to
hmart
stuff
in
there
as
well
access.
What's
up,
maybe
something
want
a
benchmark.
Stefan
there's,
no,
maybe
maybe
so.
I
F
F
Not
just
when
I
build
fails
because
really
I
care
in
a
bill
fails
design
mutating
the
CI
system,
but
the
engineers
don't
they
should
care
when
the
test
things
right,
I
kind
of
thing
like
the
people
contributing
they
can't
help
with
a
infrastructure
problem,
and
so
that's
isn't
shown
here
because
it
doesn't
matter
but
so
yeah
like
having
notifications.
You
know
we
kind
of
missed
having
those
from
we
used
to
have
an
eyeless
to
head
this.
But
that's
back
when
we
had.
You
know
one
CIO,
one
Jenkins
humble
little
G
kids
project.
F
Now
we
have
way
more
than
that.
So
it's
a
that's.
This
tool
helped
considerably
with
not
only
just
bringing
that
you
know
no
almost
noisy
kind
of
notification
back,
but
actually
making
it
really
useful.
So
when
you
get
an
email,
it's
actually
something
that
you
need
to
fix,
and
so
that's
kind
of
wraps
it
up.
So
I
give
this
to
Gary
topics
a
good
week.
B
F
B
Hallway
here,
starting
to
fill
up
with
some
people,
but
let
me
know
if
it
gets
too
noisy
and
I
can
try
to
pop
into
another
room
anyways.
So
one
of
the
things
now
that
we
are
doing
full
performance
benchmarking
and
cracking
that
inside
a
CI
so
every
commit
we
do
a
full
run
of
there's
about
70
or
80
performance
tests
that
were
executing
against
five
different
branches
and
library.
So
we
get
a
pretty
good
idea
of
how
things
are
getting
faster.
B
The
problem
is
we're
just
looking
at
table
data,
so
it's
like
how
do
we,
the
the
challenge?
I,
guess
that
I'm
that
I'm
running
into
that
I
mean
but
I'd
like
to
kind
of
figure
out
is
how
do
we
actually
treat
performance
results?
Is
it
doesn't
make
sense
to
treat
those
as
regressions
and
if
so,
what
triggers
those
progressions?
Is
it
a
slowdown
of
five
percent?
Oh
you
can
give
variations
of
up
to
five
percent.
It
may
be
even
more.
Is
it
a
slowdown
of
twenty
percent?
B
Well,
then,
that
won't
catch
some
of
the
gradual
slowdowns
if
it's
getting
five
percent
slower
each
time.
So
basically,
one
of
the
things
that
I'm
yeah
I'm
going
to
try
to
figure
out
is
how
do
we
actually
treat
performance
benchmarks
as
regressions
and
so
yeah?
If
you
guys
have
any
ideas,
I
was
curious,
I'll,
be
curious
to
be
rational
right.
A
B
A
Good
look
one
thing:
I
think
it's
really
useful
about
the
weather
on
it's
a
slow
down
or
not
I
mean
you
know
acceptable.
One
thing
would
be
really
interesting
to
see
is
if
there's
a
way
you
can
see
that
you
know
this
check-in
introduced
to
slow
down
by
five
percent
and
and
find
out.
You
know,
go
to
that
check
in
itself
and
find
out
what
was
it
that
made
that
worse,
you
know
and
then
decide
from
there.
A
B
A
So
so
one
thing
you
do
is
I
know:
you're
checking
a
performance
like
on
sorted
as
a
test,
but
is
there
anything
that
can
check
performance,
as
you
know,
sort
of
on
the
client
side
as
a
whole?
Like
say
you
have
calendar
and
some
other
widget
and
the
app
framework
and
they
all
come
together
and
somehow
there's
something
that
only
occurs
because
they're,
you
know
interfering
each
other
River,
you
know
are
there
test,
for
that?
Is
it
mostly
like
Enoch
text-based,
know.
B
F
A
I
A
So
for
you
a
while
back
and
where
I'm
not
performance
benchmarks,
I
mean
you
talk
about.
There's,
you
have
tabular
data
now,
but
is
there
any
good
way
to
visualize
these
two?
That
would
be
another?
Should
you
know?
Is
it
sparklines?
Is
it
you
know
you?
How
many
do
you
overlay
I
mean
if
you
looked
at
how
to
visualize
that
they
then
yeah.
B
B
Gotcha
yeah,
so
let
me
look
here:
where's
the
bank.
B
Sorry,
it
was
super
quiet
here
for
a
while
and
then
the
hallway
filled
up,
anyways
I
think
we're
having
a
shootout
behind
you.
Yeah.
A
E
H
E
B
To
calm
down
a
little
over
here
so
do
is
yes,
this
is
what
we
currently
have.
Basically,
we're
able
to
look
at
three
data
points
per
commit
for
each
test,
so
yeah
you
get
the
number
of
iterations
the
variants
as
well
as
how
much
faster
or
slower.
In
some
cases
it
has
been
since
the
previous
commit.
So
basically
here's
here's
a
test
on
model
that
will
look
at
the
last
three
releases,
along
with
our
to
development
branches,
so
yeah
we
can
do.
B
We
can
do
a
graph
here
to
to
show
how
much
faster
we
are
getting
I.
While
that
would
be
kind
of
useful,
I'm
not
sure
that
that's
exactly
what
we
need
out
of
it,
though
it
it
it'll
kind
of
get
get
us
a
little
bit
there,
but
we
need
we
need.
We
need
some
more
triggers,
or
at
least
something
something
a
little
bit
more
I,
don't
get
it
into
the
roots
getting
to
the
root
issue
of
what
people
actually
want
to
see,
instead
of
just
seeing
graphs,
get
little
pretty
graphs
get
faster.
B
H
A
B
A
Feature
Nina
sure
about
Leah.
This
is
it.
B
B
What's
the
word
I'm
looking
for
yeah
and
it
was
to
paralyze
the
all
of
our
testing
and
to
form
a
test,
suite
op
two
clusters
with
phantom
j/s,
and
for
that
we
could
actually
even
just
use
pretty
low
powered
hardware.
It
doesn't
need
to
be
anything
fast,
because
is
this
number
in
terms
of
how
many
iterations
we
go
per?
Second,
that's
not
really
that
useful.
The
really
useful
bit
of
data
is
how
much
graphs
are
we
kidding,
not
necessarily
how
fast
are
we
going?
B
A
A
E
B
So
here's
the
thing
across
take
this
my
base
core
with
20
buried
attributes.
This
entire
test
suite
right
here
does
need
to
be
run
in
identical
circumstances.
It
needs
to
preferably
be
run
on
the
same
exact
Hardware
in
very
similar,
the
starting
state,
and
have
very
little
fluctuation
in
resources
available
on
the
on
the
CPU
and
memory
during
the
testing
process,
but
the
mybay
scored.
B
Take
all
of
these
tests,
which
are
there's
roughly
I,
don't
know
a
70
or
80
of
them
in
the
library
right
now
and
ship
them
off
to
70
or
80
different
phantom
Jas
instances
and
get
test
results
back
within
a
minute
as
opposed
to
right
now.
It
takes
us
about
90
minutes
to
run
across
the
entire
library,
so
I.
D
B
It
depends
on
how
light
on
how
high
powered
the
hardware
it's
is
it
something
that
can
support
three
VMs
running
it
it?
Some
of
these
tests
might
spike
the
CPU
to
one
hundred
percent.
So
is
it
three
beams
that
can
support
three
virtual
processors
running
that
fast
without
them
clobbering
each
other
I,
don't
know
it
all
depends
on
the
actual
hardware.
What's
their
rights.
B
B
B
Yeah
and
this
this
test
could
be
run
on
this
test.
Suite
can
be
run
on
sorry,
not
this
isn't
a
sweet.
This
is
an
actual
test,
but
so
this
test
can
be
run
on
really
crappy
hardware.
This
can
be
run
on
a
supercomputer
and
the
variance
between
those
two
tests
doesn't
matter
as
long
as
the
the
results
that
you
get
or
the
environment
that
they're
being
tested
in
a
relative
to
each
other
is
is
accurate.
B
All
right
all
right
now
bad,
so
one
of
the
things
is
always
even
know
how
I
got
on
this
class.
I
know,
but
I
was
looking
at
our
linting
results
and
it
actually
is
a
little
bit
sad.
We
have
and
pull
up
some
of
them.
We
have
2663
Linton
airs
in
the
library
that
that
seems
like
a
lot
to
me.
We.
A
B
Don't
know
I
have
been
haven't
done,
any
analysis
to
see
which
modules
or
which
components
have
more
errors,
which
ones
have
less
but
I
I
know
just
from
looking
at
some
of
them,
the
ones
that
haven't
been
touched
for
a
while
certain
they
do
have
more
than
10
years
and
then
there's
some
like
button
and
scroll
view
that
I
do
quite
a
bit
of
work
on
that
I,
pretty
sure
down
to
zero
so
and
I'm
sure
there's
a
lot
of
others
that
are
kind
of
like
that.
So.
A
B
Really
slower,
no
I,
don't
think
many
of
these
are
a
would
affect
performance
at
all.
What
they
do
affect,
though,
is
stability,
and
there's
I
mean
that's.
The
point
of
linting
is
to
reduce
any
is
to
reduce
any
bugs
that
might
creep
into
your
code.
Just
do
to
kind
of
silly
mistakes.
B
E
A
E
A
This
is
this
a
product
of
people
who
are
checking
in
new
code.
That's
not
blended,
or
is
this
a
product
of
old
code?
That's
there
for
a
long
time
that
no
one's
you
know,
there's
no
one.
No
one's
gonna
sit
down
and
go.
I'm
gonna
go
with
the
library
and
chicken
thinking.
You
know.
Thousand
files
yeah.
B
Well,
it
does
of
kind
of
what
I
did
last
night,
actually
with
the
with
the
poll
request.
I
submitted
that
fixed
all
the
trailing
whitespace
he's
in
the
library
and
there
was
like
a
hundred
sixty
or
so
files
and
it
resolved
what
about
1200
of
those
2600
errors.
So
you
just
thought
that
was
it
easy
because
ones
or
ones
that
that
you
can't
automate
on
just
a
simple
little
script,
so
mm-hmm
so
I
guess
my
question
is
yeah
okay.
B
We
can
go
ahead
and
take
care
of
some
of
these
really
easy
ones
and
write
a
script
to
go
through
and
just
do
it
all
those
only
one
line.
It
took
10
minutes
to
do
the
entire
process
and
it
got
rid
of
about
half
the
heirs,
but
the
remaining
half
are
going
to
be
much
more
on
are
going
to
be
a
little
bit
more
intensive
to
to
comb
through.
So
every
sympathy,
automated
I
mean
I.
B
B
That
they
work
right
so
one
example
is
I,
went
through
and
looked
at
the
top
15
linting
errors
that
we
have
so
far
and
away
the
most
popular
one
with
1370
errors
was
trailing
whitespace.
Well,
if
we
pull
in,
I
think
it's
polar
crossed
1120,
whichever
one
I
submitted
last
night
that'll
take
care
of
that
one.
So
then
the
then
the
new
number
one
is
167
hours
is
too
many
of
our
statements,
so
that
is
people
including
bar
statements,
more
than
once
within
a
function.
Some
of
those
might
be
intentional.
B
B
E
I,
what
happened
for
me
like
on
an
older
team
that
I
worked
on
was
that
we
decided
it
wasn't
worth
the
effort
to
fix
it.
All
in
one
go
well
said:
if
somebody
touched
that
file
in
the
Sol
linting
errors
and
they
would
fix
it
and
then
it
would
just
have
a
threshold
which
we
would
be
going
down.
You
know
Wes
time
progressed,
no
yeah.
B
So
I
certainly
wouldn't
advocate
that
we
drop
everything
we
do
and
spend
a
month.
Fixing
other
them
take
that
one,
but
they
spend
a
sizable
chunk
of
time.
Fixing
all
of
these
linting
errors
all
at
once.
It's
something
that
would
be
kind
of
more
gradual,
but
I
guess.
My
question
is
is:
is
that
something
that
we
think
is
a
worthwhile
endeavor
actually
do
I
would
vote
yes,
because.
A
A
Sort
of
like
a
cost,
benefit
thing
right.
If
you
spend
that
time
fixing
these
lenders-
and
you
could
have
been
you
know-
I
make
a
date
table
better
or
something.
This
is,
for
example,
which
one
do
you
think
would
happen.
That's
probably
why
they're
still,
there
is
because
that
cost
benefit
analysis
that
engineers
are
probably
subconsciously
doing
is
happening
where
there
you
know,
I'm
going
to
go
victims
better
yeah.
B
I
B
A
B
B
B
B
B
A
B
The
only
other
one
that
I
think
of
that
we
could
automate
is
we
have
86
warnings
for
mixed
tabs
and
spaces.
So
that's
all
the
same
lines
before
in
the
characters.
Again,
you
have
some
tabs
and
tons
some
spaces
indenting
the
codeine,
so
I
guess
enough
to
figure
out
some
way
to.
That
seems
like
something
that
could
easily
be
scripted
to
essentially
convert
those
tabs
into
spaces
right.
I
guess:
I'd
have
to
figure
out
the
proper,
regular
expression
to
look
for
n
number
of
tabs
and
spaces
mix
before
yeah
and
you're
off.
A
B
A
A
I
B
No
that's
about
it
for
the
for
the
linting
discussion
that
I
wanted
to
bring
up
by
yeah
I.
Think
some,
some
good
stuff
came
out.
So
all
yeah,
when
I
find
small
town
hall,
write
a
post
or
something
for
itself.
A
The
agenda,
so
we
have
just
our
every
meeting.
We
talked
about
that
for
grabs
bugs
which
encourage
anyone
who
is
wanting
to
get
their
feet
wet
in
the
project
to
go
check
those
out
and
if
you
are
a
committer
or
a
reviewer
that
have
bugs
that
you
want
it
fixed.
But
you
don't
have
the
time
for.
Please
include
in
the
net
programs.
I
B
B
Tell
you
when
I
the
first
time
I
took
one
of
my
large
projects
and
did
a
bunch
of
lending
them
and
I
learned
just
a
lot
about
yep
javascript
is
the
language.
Is
the
language
it
really
kind
of
helps
you
learn
and
kind
of
gets
those
best
practices
embedded
in
your
head,
so
yeah
any
movies
out
there
with
JavaScript
that
are
wanting
to
learn
a
little
bit
more
about
the
language.
Yeah
just
take
a
big
project
like
why
do
I
just
go
through
and
do
some
lifting
so.
B
Lent
your
github
com
/,
while
you
live
/
while
you
live
bash
the
vent,
and
that
is
a
package
that
includes
our
default
config
for
jsf
and
that's
when
yogi
lint
and
some
other
I
believe
in
shifter
uses
it
pulls
them
at
config
and
then
we're
going
to
actually
get
a
command-line
tool.
Just
called
Yui
Brent
that
uses
that
automatically
applies
the
jsm
to
convey
to
Jas
hint
and
runs
it
against
one
of
the
files
you
want.
B
B
Yeah,
so
that's
certainly
yeah.
It's
just
a
little
annoying
about
to
do
good,
big,
stop
out
so
I'm
cut
out
flat,
then
be
nice.
A
So,
in
terms
of
the
staffer
request
is
this
morning:
I
went
through
all
the
old
board
press
and
I
start
notifying
people
who
need
to
look
at
bugs
and
issues
that
have
been
sticking
around
for
a
long
time.
I
think
when
Tony
gets
back
from
vacation
I'm
going
to
bug
him
about
there's
a
lot
of
them
for
classroom,
saw
theme
that
I've
been
sticking
around
for
a
while
when
I
helped
him
get
those
pulled
in
and
basically
I'm
just
going
through
those
issues.
A
If
you
do
have
a
bug,
that's
been
sticking
around
our
quest.
It's
been
sitting
around
and
you
haven't
gotten
our
viewer
or
feedback
that
you
need.
Please
update
your
request
asking
for
that
and
that
will
show
up
the
usually.
What
happens
is
the
old
of
the
poor
quest
is
the
less
likely
it
is
to
get
attention
so
any
time
you
bump
a
comment
that
helps
sort
of
put
it
back
into
people's
view,
especially
when
they're
they're
checking
upon
request
at
once.
So
that
would
be
helpful.
A
A
Anything
things
are
going
really
good
on
our
release
branch,
as
you
can
see
from
Leeds
dashboard,
and
it
looks
like
we're
on
track
for
next
week
for
that
release,
so
things
we're
still
figuring
out
how
how,
when
there's
a
fixed
nating
the
release
branch,
how
that
manifests
itself
back
into
debt,
free
x
or
net
Nestor?
So
that's
someone
still
you
should
you
should
start
seeing
for
quests
or
hook
check-ins
that
just
sort
of
come
out
of
nowhere
and
sir,
because
their
favorite
I
came
from
the
lease
branch.